CN114445565A

CN114445565A - Data processing method and device, electronic equipment and computer readable medium

Info

Publication number: CN114445565A
Application number: CN202011232976.3A
Authority: CN
Inventors: 王飞; 王民康
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2020-11-06
Filing date: 2020-11-06
Publication date: 2022-05-06

Abstract

The embodiment of the disclosure provides a data processing method and device, electronic equipment and a computer readable medium. The method comprises the following steps: determining a plurality of candidate areas from the target area based on static obstacle information in a plurality of acquisition processes, wherein the plurality of acquisition processes are used for acquiring point cloud data of a point cloud map for constructing the target area. The method further comprises the following steps: and determining at least one region to be repaired from the plurality of candidate regions based on a plurality of parts corresponding to the plurality of candidate regions in the point cloud map. The method further comprises the following steps: and determining missing point cloud data in the at least one region to be repaired by using the point cloud data of the reference region associated with the at least one region to be repaired in the point cloud map. In the embodiment of the disclosure, the static obstacle information can be used for determining the area to be repaired in the point cloud map, and then the point cloud data of the area to be repaired can be repaired, so that the efficiency, the accuracy and the automation level of point cloud map repairing are improved.

Description

Data processing method and device, electronic equipment and computer readable medium

Technical Field

Embodiments of the present disclosure relate generally to data processing and high-precision mapping, and more particularly, to a data processing method, a data processing apparatus, an electronic device, and a computer-readable medium.

Background

The high-precision map is a machine-oriented digital map, and can be used in application scenes such as automatic driving, robot navigation and positioning. The high-precision map plays an important role in an automatic driving system, and the environment sensing unit, the path planning unit and the vehicle positioning unit all depend on the high-precision map to work in different degrees. The point cloud map is one of the manufacturing bases and important components of the high-precision map, and simulates a real environment by using dense point clouds, and special point clouds can be marked by a label to reflect special identification and the like in the real environment.

Generally, in order to obtain a point cloud map of a target area, point cloud data of the target area may be acquired using a laser scanning measurement. The acquired point cloud data may then be processed and processed to generate a point cloud map of the target area. During the acquisition process, part of the point cloud data of the target area may not be acquired for various reasons, which results in the point cloud map having an area to be repaired in which the point cloud data needs to be repaired. However, in the conventional scheme, various defects still exist in the manner of determining and processing the region to be repaired of the point cloud map to be optimized.

Disclosure of Invention

The embodiment of the disclosure provides a technical scheme for determining and repairing a to-be-repaired area in a point cloud map, and particularly provides a data processing method, a data processing device, an electronic device and a computer readable medium.

In a first aspect of the disclosure, a data processing method is provided. The method comprises the following steps: determining a plurality of candidate areas from the target area based on static obstacle information in a plurality of acquisition processes, wherein the plurality of acquisition processes are used for acquiring point cloud data of a point cloud map for constructing the target area. The method further comprises the following steps: and determining at least one region to be repaired from the plurality of candidate regions based on a plurality of parts corresponding to the plurality of candidate regions in the point cloud map. The method further comprises the following steps: and determining missing point cloud data in the at least one region to be repaired by using the point cloud data of the reference region associated with the at least one region to be repaired in the point cloud map.

In a second aspect of the present disclosure, a data processing apparatus is provided. The device includes: a candidate region determination module configured to determine a plurality of candidate regions from the target region based on static obstacle information in a plurality of acquisition processes for acquiring point cloud data for constructing a point cloud map of the target region. The device also includes: a candidate region determination module configured to determine a plurality of candidate regions from the target region based on static obstacle information in a plurality of acquisition processes for acquiring point cloud data for constructing a point cloud map of the target region. The apparatus further comprises: the missing point cloud data determining module is configured to determine missing point cloud data in at least one to-be-repaired area by using point cloud data of a reference area in the point cloud map, wherein the reference area is associated with the at least one to-be-repaired area.

In a third aspect of the disclosure, an electronic device is provided. The electronic device includes one or more processors and memory. The memory is for storing computer-executable instructions. The computer-executable instructions are executed by one or more processors to implement a method according to the first aspect of the present disclosure.

In a fourth aspect of the disclosure, a computer-readable medium having stored thereon computer-executable instructions is provided. The computer executable instructions, when executed by the processor, implement the method according to the first aspect of the present disclosure.

It should be understood that the statements herein reciting aspects are not intended to limit the critical or essential features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description.

Drawings

The above and other objects, features and advantages of the embodiments of the present disclosure will become readily apparent from the following detailed description read in conjunction with the accompanying drawings. In the drawings, several embodiments of the present disclosure are shown by way of example and not limitation.

FIG. 1 illustrates a schematic diagram of an example environment in which embodiments of the present disclosure may be implemented.

FIG. 2 shows a flow diagram of an example data processing method according to an embodiment of the present disclosure.

Fig. 3A-3D illustrate schematic diagrams of multiple sets of static obstacle regions associated with a multiple acquisition process, according to an embodiment of the present disclosure.

Fig. 3E illustrates a schematic diagram of an intersection of multiple sets of static obstacle regions, in accordance with an embodiment of the present disclosure.

Fig. 4 illustrates an example process for determining one or more regions to be repaired in accordance with an embodiment of the disclosure.

Fig. 5 shows a schematic diagram of a plurality of corresponding portions of a plurality of candidate regions in a point cloud map, according to an embodiment of the disclosure.

Fig. 6A shows a schematic diagram of a candidate region intersecting a road boundary line according to an embodiment of the present disclosure.

Fig. 6B shows a schematic diagram of intersection of a corresponding portion of a candidate region in a point cloud map with a road boundary point cloud, in accordance with an embodiment of the present disclosure.

Fig. 7 shows a schematic diagram of taking an intermediate region between an extension region and a region to be repaired as a reference region according to an embodiment of the disclosure.

FIG. 8 shows a block diagram of an example data processing apparatus, according to an embodiment of the present disclosure.

FIG. 9 illustrates a block diagram of an example device that can be used to implement embodiments of the present disclosure.

Throughout the drawings, the same or similar reference numerals are used to designate the same or similar components.

Detailed Description

The principles and spirit of the present disclosure will be described below with reference to a number of exemplary embodiments shown in the drawings. It is understood that these specific embodiments are described merely to enable those skilled in the art to better understand and implement the present disclosure, and are not intended to limit the scope of the present disclosure in any way.

As indicated above, in the conventional scheme, there are also various drawbacks to be optimized in the way of determining and processing the region to be repaired of the point cloud map. In particular, during the acquisition of point cloud data of a target area, static obstacles may exist in the target area. Such static obstructions may occlude the area where the point cloud data needs to be acquired, thereby causing the desired point cloud data to be unavailable for acquisition. For example, in point cloud data acquisition for a road surface, facilities that are stationary with respect to the road and vehicles parked on the road surface may obstruct the road surface, resulting in failure to obtain point cloud data and real road surface information of the road surface. In the case of a static obstacle obstruction, the point cloud data acquired by laser scanning is actually point cloud data of the static obstacle, not point cloud data of the desired area. Therefore, in the subsequent processing and processing of the point cloud data, the point cloud data of the static obstacle may need to be removed from the acquired point cloud data, which may result in "holes" in the subsequently generated point cloud map, that is, the point cloud data of the desired area is missing from the point cloud map.

Incomplete point cloud maps may reduce the performance of various applications that use the point cloud maps, and in some cases may also render the objectives of these applications unfeasible. Therefore, in the process of manufacturing the point cloud map, the point cloud data of the area to be repaired in the point cloud map may need to be repaired. For example, for a point cloud map of a road environment, due to the characteristics of the road condition, under the condition of static obstacle shielding, the point cloud data obtained by scanning with a laser radar cannot record the height information of a real road surface. When applied to automatic driving, the road surface height information in the point cloud map is crucial to detecting road obstacles affecting the vehicle driving. Thus, in point cloud mapping of a road environment, it may be desirable to address situations where static obstacles (e.g., roadside parked vehicles) block the road surface.

In conventional solutions, since the collected point cloud data relates to a large geographic range, the edges of the region to be repaired in the point cloud map may need to be manually marked by an operator. Then, for the marked region to be repaired, the computing device may generate patch point cloud data for addition to the point cloud map. However, such conventional solutions have deficiencies in processing efficiency, accuracy, and automation level.

In view of the above-mentioned problems and other potential problems in the conventional solutions, embodiments of the present disclosure provide a technical solution for determining and repairing a region to be repaired in a point cloud map. In this solution, the computing device may determine a plurality of candidate regions of the region to be repaired from the target region based on the static obstacle information in the multiple acquisition processes. Then, based on a plurality of portions of the point cloud map corresponding to the plurality of candidate regions, the computing device may determine one or more regions to patch from the candidate regions. The computing device may then utilize point cloud data of a reference region in the point cloud map associated with the region to be repaired to determine missing point cloud data in the region to be repaired. In the technical scheme of the disclosure, the computing device can automatically determine the area to be repaired in the point cloud map by using the static obstacle information, and then can supplement the point cloud data of the area to be repaired based on the point cloud data of the reference area, so that the efficiency, the accuracy and the automation level of point cloud map repair are improved.

Fig. 1 illustrates a schematic diagram of an example environment 100 in which embodiments of the present disclosure may be implemented. As shown in fig. 1, the example environment 100 may include an acquisition device 110 that may acquire point cloud data 121 of the target area 105 to produce a point cloud map 123 of the target area 105. For example, the acquisition device 110 may use a lidar measurement device to emit laser light 115 to various objects in the target area 105, and then perform correlation measurements and calculations based on the received laser light reflected back from the objects, thereby obtaining point cloud data 121 associated with the target area 105. Although in the example of fig. 1, acquisition device 110 is depicted as an acquisition cart with a lidar measurement device mounted thereto, this depiction is merely illustrative and is not intended to limit the scope of the present disclosure in any way. In other embodiments, the acquisition device 110 may have any other possible form. For example, the acquisition device 110 may be another vehicle (such as a civilian automobile, electric vehicle, aircraft, etc.) carrying a lidar measurement device or a lidar measurement device carried by a person. Furthermore, the lidar measurement device may also be replaced by any other type of point cloud data acquisition device. In other words, the acquisition device 110 of embodiments of the present disclosure may be any device having point cloud data acquisition capabilities.

As used herein, data acquired by the acquisition device 110 via a lidar measurement device may be referred to as "point cloud data. In general, "point cloud data" may refer to data information of respective points of an object or an object surface returned when a beam of laser light is irradiated on the object or the object surface, which may include coordinate information of each point (for example, coordinate values in a three-dimensional coordinate system), and may also include laser reflection intensity (also referred to as "reflection value"). Thus, in the context of the present disclosure, "point cloud data" may generally refer to data associated with a point in space, e.g., point coordinate data in space. In some cases, "point cloud data" may also include other aspects of information associated with points in space, such as color information or reflection intensity information, etc. In embodiments of the present disclosure, "point cloud data" may also refer to any spatial point data that can be used to make a point cloud map or high precision map.

In the context of the present disclosure, a high-precision map generally refers to an electronic map with high-precision data. For example, the high precision here means that the absolute coordinate precision of the high-precision electronic map is higher on the one hand. Absolute coordinate accuracy refers to the accuracy between an object on a map and a real world-outside thing. For another example, the absolute accuracy of a high-accuracy map is generally on the sub-meter level, i.e., within 1 meter, and the relative accuracy in the transverse direction (e.g., the relative position accuracy between a lane and a lane, and between a lane and a lane line) is often higher. On the other hand, the high-precision map contains more abundant and detailed road traffic information elements than the conventional map. Furthermore, in some embodiments, the high-accuracy map not only has high-accuracy coordinates, but also has an accurate road shape, and data of the gradient, curvature, heading, elevation, and roll of each lane are also included. In some embodiments, the high-precision map may not only depict roads, but also how many lanes there are on a road to truly reflect the actual style of the road.

In the example of fig. 1, the target area 105 where the acquisition device 110 needs to acquire point cloud data is an area including the road 101. As shown in fig. 1, a road 101 may be defined by

road boundary lines

102 and 104, and may be divided into three lanes by

lane lines

106 and 108. The acquisition device 110 is depicted as driving in a center lane and emitting a laser 115 to acquire point cloud data of the target area 105. However, it should be understood that such depiction is merely exemplary, and embodiments of the present disclosure are not limited to the acquisition device 110 being in a particular location in the target area 105, but apply equally to the case where the acquisition device 110 is in any possible location. Furthermore, it should be noted that the depiction of the road 101 in fig. 1 as having three lanes is merely illustrative and is not intended to limit the scope of the present disclosure in any way. In other embodiments, the road 101 may have any number of lanes and road patterns.

In the example of FIG. 1, vegetation, such as trees 112-1 to 112-6, etc., is disposed beyond the

roadway boundary lines

102 and 104. Further, traffic aids, such as traffic lights 114, are provided outside the road boundary line 102. In addition, during the acquisition of the point cloud data of the target area 105 by the acquisition device 110, the vehicle 116 and the vehicle 118 are parked on the road 101. That is, the vehicle 116 and the vehicle 118 are not in a traveling state, but remain stationary with respect to the road 101. It will be appreciated that the trees 112-1 through 112-6, traffic lights 114, and

vehicles

116 and 118 depicted in FIG. 1 are merely illustrative and are not intended to limit the scope of the present disclosure in any way, and that the target area 105 of embodiments of the present disclosure may be a roadway environment including any other objects or facilities, as well as other environments other than a roadway environment. Moreover, the particular number of various objects or elements depicted in fig. 1 is merely illustrative, and embodiments of the present disclosure are equally applicable to a target region 105 having any number of similar objects or elements.

As shown in fig. 1, to obtain a point cloud map 123 of the target area 105, the acquisition device 110 may provide acquired point cloud data 121 of the target area 105 to the computing device 120. The computing device 120 may then process and process the point cloud data 121 to generate a point cloud map 123 of the target area 105. For example, the computing device 120 may derive a three-dimensional reconstructed point cloud map 123 from the point cloud data 121 via a simultaneous localization and mapping (SLAM) method. Of course, embodiments of the present disclosure are not so limited, and computing device 120 may also generate point cloud map 123 from point cloud data 121 using any other available method.

As mentioned previously, when the acquisition device 110 acquires the point cloud data 121 of the target area 105, a static obstacle in the target area 105 may block an area that the acquisition device 110 desires to acquire (e.g., point cloud data of the road surface of the road 101), resulting in the acquisition device 110 failing to acquire the point cloud data of the desired area. As used herein, a "static obstacle" may refer to an obstacle that remains stationary relative to the target area 105 while the acquisition device 110 acquires point cloud data of the target area 105. This means that static obstacles may remain occluded from the same area at all times. In this regard, static obstacles may be distinguished from "dynamic obstacles" that are in motion relative to the target region 105 because as a dynamic obstacle moves relative to the target region 105, an area previously occluded by the dynamic obstacle may become no longer occluded by the dynamic obstacle so that point cloud data for that area may be acquired by the acquisition device 110 (such as through multiple scan measurements). As an illustrative example, in FIG. 1, traffic signal 114, trees 112-1 to 112-6, and vehicle 116 and vehicle 118 may be considered static obstacles with respect to road 101. In contrast, if the vehicle 116 and the vehicle 118 are in a driving state on the road 101, they may be considered as dynamic obstacles.

In some cases, to fully acquire the point cloud data 121 to the target area 105, or to reduce the impact of dynamic obstacles on a single acquisition process, the acquisition device 110 may perform multiple acquisition processes of point cloud data to the target area 105. During each acquisition, the acquisition device 110 may detect static obstacles in the target area 105, such as by using a context-aware component or module to detect the static obstacles, to generate static obstacle information 125. For example, the static obstacle information 125 may include size information and location information of the static obstacle. For another example, the static obstacle information 125 may also include shape information and type information of the static obstacle, and the like. More generally, the static obstacle information 125 may include any information related to static obstacles. In some embodiments, the static obstacle information 125 may be represented in the form of a detection box surrounding the static obstacle, for example, the detection box may have a cubic shape. Of course, in other embodiments, the static obstacle information 125 may be represented in any other suitable form.

As shown in fig. 1, in addition to the point cloud data 121, the acquisition device 110 may provide the detected static obstacle information 125 to the computing device 120, so that the computing device 120 may perform a fix to the point cloud map 123 based on the static obstacle information 125 and the point cloud map 123 to obtain a fixed point cloud map 127. It should be noted that although the static obstacle information 125 is provided to the computing device 120 by the acquisition device 110 in the example of fig. 1, in other embodiments, the computing device 120 may obtain the static obstacle information 125 from other static obstacle detection devices other than the acquisition device 110.

In some embodiments, in the mapping process in which the computing device 120 generates the point cloud map 123 based on the point cloud data 121, the static obstacle information 125 may be used to remove the point cloud data of static obstacles in the point cloud data 121 frame-by-frame. In the context of the present disclosure, a "frame" may refer to point cloud data or a point cloud frame formed by acquisition device 110 performing a point cloud data acquisition process on target area 105 once, which may correspond to, for example, a point cloud data or a point cloud frame formed by a laser radar measurement device completing a 360 degree scan. As explained before, since the point cloud data of the static obstacle is not the point cloud data desired to be acquired, the computing device 120 may remove the point cloud data of the static obstacle from frame to frame in the point cloud data 121 based on the static obstacle information 125. However, as a result, the point cloud data of the area blocked by the static obstacle will be absent from the point cloud data 121, so that the portion of the generated point cloud map 123 corresponding to the static obstacle may not have the point cloud data but form a point cloud "hole". Therefore, in an embodiment of the present disclosure, the computing device 120 may also complement the point cloud data missing in the point cloud map 123 based on the static obstacle information 125 and the point cloud map 123, thereby obtaining a repaired point cloud map 127.

In some embodiments, computing device 120 may comprise any device capable of performing computing and/or control functions, which may be any type of fixed, mobile, or portable computing device, including but not limited to a special purpose computer, general purpose computer, desktop computer, laptop computer, notebook computer, netbook computer, tablet computer, multimedia computer, mobile phone, general purpose processor, microprocessor, microcontroller, or state machine. Computing device 120 may be implemented as an individual computing device or combination of computing devices, e.g., a combination of a Digital Signal Processor (DSP) and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.

It is noted that in some embodiments, computing device 120 may be a computing device in acquisition device 110. In other words, the generation and patching of the point cloud map 123 may be done by the acquisition device 110. In other embodiments, the computing device 120 may also be a computing device for producing a point cloud map or high-precision map, which may be located remotely from the acquisition device 110, e.g., the computing device 120 may be a cloud-side or server-side computing device. In this case, the acquisition device 110 may transmit the point cloud data 121 and the static obstacle information 125, etc. to the remote computing device 120. Further, it should be noted that the various processes or operations described herein as being performed by computing device 120 may also be performed by a plurality of computing devices, respectively, i.e., each computing device may implement a portion of the processes or operations, respectively, and the computing devices may be located in different geographic locations or belong to different entities. Additionally, in the context of the present disclosure, computing device 120 may also be referred to as electronic device 120, and these two terms may be used interchangeably herein.

Furthermore, it should be understood that fig. 1 schematically illustrates only objects, units, elements, or components of an example environment 100 that are relevant to embodiments of the present disclosure. In practice, the example environment 100 may also include other possible objects, units, elements, or components, and so forth. In addition, the particular number of objects, units, elements, or components shown in fig. 1 is merely illustrative and is not intended to limit the scope of the present disclosure in any way. In other embodiments, the example environment 100 may include any suitable number of objects, units, elements, or components, among others. Thus, embodiments of the present disclosure are not limited to the specific scenario depicted in fig. 1, but are generally applicable to any technical environment for patching point cloud maps. An example data processing method of an embodiment of the present disclosure is described below with reference to fig. 2.

FIG. 2 shows a flow diagram of an example data processing method 200 according to an embodiment of the present disclosure. In some embodiments, the example method 200 may be implemented by the computing device 120 of fig. 1, e.g., may be implemented by a processor or processing unit of the computing device 120. In other embodiments, all or part of the example method 200 may also be implemented by other electronic devices independent of the example environment 100, or may be implemented by other devices or units in the example environment 100. For ease of illustration, the example method 200 will be described with reference to fig. 1, taking as an example the computing device 120 performing the example method 200.

In the example of fig. 1, it is assumed that the acquisition device 110 performs a plurality of acquisition processes for acquiring point cloud data 121 that build a point cloud map 123 of the target area 105. In this case, at block 210 of fig. 2, the computing device 120 may determine a plurality of candidate regions from the target region 105 based on the static obstacle information 125 of the acquisition device 110 over a plurality of acquisitions. As used herein, a "candidate region" may refer to a region in the point cloud map 123 that may require patching of point cloud data, i.e., a candidate for a region to be patched in the point cloud map 123. Accordingly, the computing device 120 may further determine the region to be repaired in the point cloud map 123 among the plurality of candidate regions in subsequent processing. It should be noted that various areas mentioned in the context of the present disclosure refer to areas in the target area 105 in a strict sense, but since an object, a position, or a range in the point cloud map 123 of the target area 105 has a corresponding relationship with a real object, an object, a position, or a range in the target area 105, various areas herein are sometimes used to refer to areas in the point cloud map 123 without causing confusion.

It will be appreciated that as the acquisition device 110 performs multiple acquisition processes of the point cloud data for the target area 105, the static obstructions may be different in different acquisition processes. For example, in the example of fig. 1, one or both of vehicle 116 or vehicle 118 may be a static obstacle during a certain acquisition. However, during another different acquisition process, one or both of vehicle 116 or vehicle 118 may be traveling in target area 105 or have been driven away from target area 105 and are no longer static obstacles. Thus, in some embodiments, the computing device 120 may determine the location and size of the static obstacle during each acquisition of the acquisition device 110, i.e., the area corresponding to the static obstacle, from the static obstacle information 125. Computing device 120 may then synthesize the static obstacle regions over the course of multiple acquisitions to determine candidate regions. Examples of determining candidate regions for a region to be repaired based on static obstacle regions are further described below with reference to fig. 3A-3D.

Fig. 3A-3D illustrate schematic diagrams of multiple sets of static obstacle regions 310-340 associated with a multiple acquisition process, in accordance with an embodiment of the present disclosure. In particular, fig. 3A shows a schematic diagram of a first set of static obstacle regions 310, the first set of static obstacle regions 310 being associated with a first acquisition process of the acquisition device 110 for the target region 105. Similarly, fig. 3B shows a schematic diagram of a second set of static obstacle regions 320, the second set of static obstacle regions 320 being associated with a second acquisition process of the acquisition device 110 for the target region 105. Fig. 3C shows a schematic diagram of a third set of static obstacle regions 330, the third set of static obstacle regions 330 being associated with a third acquisition process of the acquisition device 110 for the target area 105. Fig. 3D shows a schematic diagram of a fourth set of static obstacle regions 340, the fourth set of static obstacle regions 340 being associated with a fourth acquisition process of the acquisition device 110 for the target area 105.

It should be noted that although the static obstacle area is depicted as a rectangular shape in fig. 3A to 3D, this depiction is merely illustrative and is not intended to limit the scope of the present disclosure in any way. In other embodiments, the static obstacle region of the target region 105 corresponding to the static obstacle may also be any other shape, such as a circular shape, an elliptical shape, a shape defined by the outline of the static obstacle, and so on. It should be noted that, in actual processing, a static obstacle region corresponding to a static obstacle in the target region 105 may be a three-dimensional space region, because the static obstacle itself is generally an object or an object occupying a three-dimensional space. Thus, in some embodiments, the various static obstacle regions in fig. 3A-3D may be considered projections of three-dimensional static obstacle regions relative to the ground. In other embodiments, the various two-dimensional static obstacle regions of fig. 3A-3D may be considered a schematic representation. It should be further noted that the particular number (4) of acquisition processes performed by the acquisition device 110 described herein is merely illustrative, and that embodiments of the present disclosure are equally applicable to any number of acquisition processes by the acquisition device 110.

As shown in FIG. 3A, during a first acquisition of the target area 105 by the acquisition device 110, the traffic lights 114, trees 112-1 through 112-6, and

vehicles

116 and 118 parked along the roadside are static obstacles. Accordingly, based on the static obstacle information 125, the computing device 120 may determine a first set of static obstacle regions 310 associated with the first acquisition process. Specifically, first set of static obstacle regions 310 may include static obstacle region 301 corresponding to traffic light 114, static obstacle region 303, static obstacle region 305, static obstacle region 307, static obstacle region 311, static obstacle region 315, and static obstacle region 317 corresponding to trees 112-1 through 112-6, respectively, and static obstacle region 309 and static obstacle region 313 corresponding to vehicle 116 and vehicle 118, respectively. The computing device 120 may remove the point cloud data in these areas from the point cloud data 121 because they are actually point cloud data of static obstacles, while point cloud data of areas desired to be acquired.

As shown in fig. 3B, in the second acquisition process of the acquisition apparatus 110 for the target area 105 with respect to the first acquisition process, the vehicle 118 parked at the roadside has been driven away from the target area 105, so the traffic signal lights 114, the trees 112-1 to 112-6, and the vehicle 116 parked at the roadside are static obstacles. Accordingly, based on the static obstacle information 125, the computing device 120 may determine a second set of static obstacle regions 320 associated with the second acquisition process. Specifically, second set of static obstacle regions 320 may include static obstacle region 301 corresponding to traffic light 114, static obstacle region 303 corresponding to trees 112-1 through 112-6, static obstacle region 305, static obstacle region 307, static obstacle region 311, static obstacle region 315, and static obstacle region 317, respectively, and static obstacle region 309 corresponding to vehicle 116. The computing device 120 may remove the point cloud data in these areas from the point cloud data 121 because they are actually point cloud data of static obstacles, while point cloud data of areas desired to be acquired.

As shown in fig. 3C, in a third acquisition process of the acquisition device 110 for the target area 105, relative to the second acquisition process, a new static obstacle appears in the target area 105. For example, the new static obstacle may be a vehicle that is temporarily parked in the target area 105 due to a malfunction or waiting for a red light. In this case, the traffic signal 114, the trees 112-1 to 112-6, the vehicle 116 parked at the roadside, and the temporarily parked vehicle are static obstacles. Accordingly, based on the static obstacle information 125, the computing device 120 may determine a third set of static obstacle regions 330 associated with a third acquisition process. Specifically, the third set of static obstacle regions 330 may include static obstacle region 301 corresponding to traffic light 114, static obstacle region 303 corresponding to trees 112-1 through 112-6, static obstacle region 305, static obstacle region 307, static obstacle region 311, static obstacle region 315, and static obstacle region 317, static obstacle region 309 corresponding to vehicle 116, and static obstacle region 319 corresponding to temporarily parked vehicles. The computing device 120 may remove the point cloud data in these areas from the point cloud data 121 because they are actually point cloud data of static obstacles, while point cloud data of areas desired to be acquired.

As shown in fig. 3D, in a fourth acquisition process of the acquisition device 110 for the target area 105, relative to the third acquisition process, the temporarily parked vehicle has left the target area 105 and another new static obstacle appears in the target area 105. For example, the another new static obstacle may be a guardrail facility or the like temporarily set in the road 101 due to road maintenance. In this case, the traffic lights 114, trees 112-1 to 112-6, vehicles 116 parked at the roadside, and guardrail facilities are static obstacles. Accordingly, based on the static obstacle information 125, the computing device 120 may determine a set of static obstacle regions 340 associated with the fourth acquisition process. Specifically, fourth set of static obstacle regions 340 may include static obstacle region 301 corresponding to traffic light 114, static obstacle region 303, static obstacle region 305, static obstacle region 307, static obstacle region 311, static obstacle region 315, and static obstacle region 317 corresponding to trees 112-1 through 112-6, respectively, static obstacle region 309 corresponding to vehicle 116, and static obstacle region 321 corresponding to a guardrail facility. The computing device 120 may remove the point cloud data in these areas from the point cloud data 121 because they are actually point cloud data of static obstacles, while point cloud data of areas desired to be acquired.

In some embodiments, in determining the candidate regions for the region to be repaired in the point cloud map 123, the computing device 120 may determine a union of the first set of static obstacle regions 310, the second set of static obstacle regions 320, the third set of static obstacle regions 330, and the fourth set of static obstacle regions 340 as the candidate regions for the region to be repaired. For example, the union of the first through fourth sets of static obstacle regions 310 through 340 may include static obstacle regions 301 through 321, which may be determined as candidate regions for a region to be repaired. In this way, the static obstacle region in each acquisition process of the acquisition device 110 for the target region 105 is taken as a candidate region, so that the computing device 120 can ensure that the missing of point cloud data due to the static obstacle is not missed. In addition, the computational overhead of the computing device 120 is low since this approach only involves the merging of multiple sets of static obstacle regions and not the comparison between multiple sets of static obstacle regions.

On the other hand, it may be noted that if a static obstacle is not present in the target area 105 during each acquisition, the acquisition device 110 may still acquire point cloud data that is not acquired during a certain acquisition process but temporarily occluded by the static obstacle during another acquisition process. For example, although vehicle 118 may cause acquisition device 110 to fail to acquire the desired point cloud data in static obstacle region 313 in the first acquisition process, computing device 120 may still acquire the desired point cloud data in static obstacle region 313 in the second through fourth acquisition processes. In other words, in the point cloud map 123 generated based on the point cloud data 121 of the multiple acquisition processes, desired point cloud data in the static obstacle region 313 may already exist. Based on such considerations, in some embodiments, in determining candidate regions for a region to be repaired in the point cloud map 123, the computing device 120 may obtain one or more candidate regions based on the intersection of the multiple sets of static obstacle regions 310-340. In this way, the computing device 120 may effectively exclude areas where point cloud data already exists among candidate areas for determining an area to be repaired, thereby simplifying subsequent processing by the computing device 120.

Fig. 3E shows a schematic diagram of an intersection 350 of multiple sets of static obstacle regions 310-340, in accordance with an embodiment of the present disclosure. As shown in fig. 3E, computing device 120 may determine an intersection 350 of first set of static obstacle regions 310, second set of static obstacle regions 320, third set of static obstacle regions 330, and fourth set of static obstacle regions 340. Specifically, intersection 350 may include static barrier region 301 corresponding to traffic light 114, static barrier regions 303 corresponding to trees 112-1 through 112-6, respectively, static barrier region 305, static barrier region 307, static barrier region 311, static barrier region 315, and static barrier region 317, and static barrier region 309 corresponding to vehicle 116. It will be noted that intersection 350 may not include static obstacle region 313 corresponding to vehicle 118, static obstacle region 319 corresponding to temporarily parked vehicles, and static obstacle region 321 corresponding to a guardrail facility, as the desired acquired point cloud data in these static obstacle regions may be obtained by integrating the point cloud data of multiple acquisition processes.

Referring back to fig. 2, at block 220, after determining a plurality of candidate regions for the region to be patched, the computing device 120 may determine one or more regions to patch from the plurality of candidate regions based on a plurality of portions of the point cloud map 123 corresponding to the plurality of candidate regions. That is, the computing device 120 may determine whether the candidate region is a region to be repaired requiring repair of the point cloud data according to the actual situation of the corresponding portion in the point cloud map 123. For example, in some embodiments, the computing device 120 may determine whether the candidate region is a region to be repaired based on whether point cloud data is present in the corresponding portion of the candidate region in the point cloud map 123. For example, the computing device 120 may count the point cloud data of the candidate region in the corresponding portion of the point cloud map 123, and if the point cloud data in the corresponding portion is not missing, the computing device 120 may filter out the candidate region, that is, determine that the candidate region is not the region to be repaired. Such an embodiment will be further described hereinafter with reference to fig. 4 and 5.

In other embodiments, the candidate region may have point cloud data in the corresponding portion of the point cloud map 123, but the computing device 120 may determine that the point cloud data in the corresponding portion is insufficient, possibly affecting subsequent applications of the point cloud map 123. For example, the point cloud data may be insufficient because the candidate region is acquired less often than other candidate regions that are not affected by the static obstacle. In this case, the computing device 120 may also determine a candidate region where the point cloud data is insufficient as the region to be patched. In further embodiments, the corresponding portion of the candidate region in the point cloud map 123 may have point cloud data present, but the computing device 120 may determine that the point cloud data in the corresponding portion is actually point cloud data of a static obstacle, not point cloud data that is intended to be acquired. This may be, for example, because static obstacles obstructing the area are not detected, so the point cloud data of the corresponding portion of the candidate area in the point cloud map 123 is not removed. In this case, the computing device 120 may also determine the candidate region as a region to be patched.

At block 230, upon determining one or more regions to be patched in the target region 105 that require patching point cloud data, the computing device 120 may utilize point cloud data of a reference region in the point cloud map 123 associated with the region to be patched to determine missing point cloud data in the region to be patched. As used herein, a "reference region" may generally refer to a region that has a similarity in some property to a region to be repaired, so point cloud data in the reference region may be used to determine missing point cloud data in the region to be repaired.

For example, the reference region may be a region near or around the region to be repaired, and thus there is a similarity in geographic location between the reference region and the region to be repaired. For another example, if the area to be repaired is a part of the road 101, the reference area and the area to be repaired may be substantially located in the road width direction, which means that the positions of both in the road length direction have similarity. Since the road 101 is generally kept substantially horizontal in its width direction without large inclination or height fluctuation, the point cloud data in the reference area can be used to determine the missing point cloud data in the area to be repaired if the reference area and the area to be repaired have similar positions in the road length direction. In other embodiments, the computing device 120 may determine the reference region of the region to be repaired by expanding the region to be repaired, such embodiments being further described below with reference to fig. 7.

It should be noted that the computing device 120 may derive the missing point cloud data in the region to be repaired from the point cloud data in the reference region in any suitable manner. For example, in a scenario in which the reference area and the area to be repaired are both road areas, the computing device 120 may directly take the height data (e.g., the height average of the point clouds) in the point cloud data in the reference area as the height data of the point cloud data in the area to be repaired. Then, the computing device 120 may linearly fit latitude and longitude data in the point cloud data in the region to be repaired according to the latitude and longitude coordinates in the point cloud data in the reference region. In other embodiments, the computing device 120 may further employ a surface reconstruction-based patching algorithm, a local dilation concept-based patching algorithm, and the like, to derive missing point cloud data in the region to be patched based on the point cloud data in the reference region.

In other embodiments, the computing device 120 may perform a plane fit using the point cloud data of the reference region to obtain missing point cloud data in the region to be repaired. As such, the computing device 120 may use a low complexity algorithm to reasonably fit out the missing point cloud data. Specifically, in a scenario in which the reference region and the region to be repaired are both road regions, the computing device 120 may perform plane fitting using the point cloud data of the reference region to derive an approximate plane equation description of the road pavement. The computing device 120 may then use the derived plane equation description to calculate missing point cloud data, such as height data of the point cloud, in the region to be repaired. In some embodiments, the plane fitting process needs to take robustness to outliers into account, so a random sample consensus (RANSAC) algorithm may be adopted for the plane fitting. Points that are beyond a certain threshold from the fitted plane may be considered outliers or outliers (e.g., possibly due to residual static obstacle traces) and deleted.

By way of example method 200, the computing device 120 may automatically utilize the static obstacle information 125 to determine the area to be repaired in the point cloud map 123, so the efficiency and accuracy of the determination of the area to be repaired may be improved while avoiding manual labeling of the area to be repaired in the point cloud map 123 by an operator. Further, since the computing device 120 patches missing point cloud data using the point cloud data of the reference region already in the point cloud map 123, the computing device 120 may quickly, accurately, and automatically determine the missing point cloud data in the region to be patched. In summary, the example method 200 may improve the efficiency, accuracy, and automation level of point cloud map patching.

As mentioned above in describing block 220 of example method 200, in some embodiments, computing device 120 may determine whether a candidate region is a region to be repaired based on whether point cloud data is present in the corresponding portion of the candidate region in point cloud map 123. Such an embodiment is further described below with reference to fig. 4 and 5.

Fig. 4 illustrates an example process 400 for determining one or more regions to repair in accordance with an embodiment of the disclosure. In some embodiments, the example process 400 may be implemented by the computing device 120 of fig. 1, e.g., may be implemented by a processor or processing unit of the computing device 120. In other embodiments, all or part of the example process 400 may also be implemented by other electronic devices separate from the example environment 100, or may be implemented by other devices or units in the example environment 100.

Fig. 5 shows a schematic diagram of a plurality of candidate regions 301-321 in a plurality of corresponding portions 501-521 in a point cloud map 123, according to an embodiment of the disclosure. As shown in fig. 5, the point cloud map 123 may be composed of a plurality of points including point 510. It is noted that, in the example of fig. 5, it is assumed that the computing device 120 has already taken the union of the first set of static obstacle regions 310, the second set of static obstacle regions 320, the third set of static obstacle regions 330, and the fourth set of static obstacle regions 340 as candidate regions of the region to be repaired, and therefore the static obstacle regions 301 to 321 depicted in fig. 3A to 3D are all candidate regions of the region to be repaired. In this case, in the point cloud map 123 generated based on the point cloud data 121, the computing device 120 may determine a plurality of portions 501 to 521 corresponding to the plurality of candidate areas 301 to 321.

Additionally, it should be noted that although the plurality of portions 501-521 are depicted in fig. 5 as rectangular in shape, such depiction is merely illustrative and is not intended to limit the scope of the present disclosure in any way. In other embodiments, the plurality of portions 501-521 in the point cloud map 123 may also be any other shape, such as a circular shape, an elliptical shape, a shape defined by the outline of a static obstacle, and so forth. It should be noted that, in actual processing, the plurality of portions 501 to 521 in the point cloud map 123 may be three-dimensional spatial regions, since the static obstacle itself is generally an object or an object occupying a three-dimensional space. Thus, in some embodiments, the plurality of portions 501-521 in fig. 5 may be considered a projection of the three-dimensional region with respect to the ground. In other embodiments, the various two-dimensional regions in FIG. 5 may be considered a schematic representation. It is further noted that the particular distribution of the point clouds depicted in fig. 5 is merely illustrative to depict whether a point cloud is present in a particular area, and is not intended to limit the scope of the present disclosure in any way. In an actual scene, the distribution of the point cloud should actually have a shape defined by the outer contour of the object or objects.

Referring to fig. 4 and 5 concurrently, at block 410 of fig. 4, the computing device 120 may determine one or more portions of the point cloud map 123 where point cloud data is missing from among the multiple portions 501-521. For example, the computing device 120 may determine whether the point cloud density in a certain portion is below a threshold density. If the point cloud density in the portion is below a threshold density, the point cloud data in the portion may be considered missing. Conversely, if the point cloud density in the portion is not below the threshold density, the point cloud data may be deemed to be present in the portion. In some embodiments, the threshold density may be determined according to the specific technical environment and accuracy requirements. As an example, in the scenario depicted in fig. 5, computing device 120 may determine that portion 501, portion 503, portion 505, portion 507, portion 509, portion 511, portion 515, and portion 517 are portions where point cloud data is missing. In contrast, computing device 120 may determine that portion 513, portion 519, and portion 521 are portions where point cloud data exists.

After determining one or more portions of the point cloud map 123 where point cloud data is missing, the computing device 120 may determine a candidate region corresponding to the one or more portions among the plurality of candidate regions 301-321 at block 420. For example, in the example depicted in fig. 5, since portion 501, portion 503, portion 505, portion 507, portion 509, portion 511, portion 515, and portion 517 are portions where point cloud data is missing, computing device 120 may determine candidate regions corresponding to these portions among candidate regions 301-321, that is, candidate region 301, candidate region 303, candidate region 305, candidate region 307, candidate region 309, candidate region 311, candidate region 315, and candidate region 317.

At block 430, after determining one or more candidate regions corresponding to the missing portions of the point cloud data, the computing device 120 may obtain one or more regions to patch based on the one or more candidate regions. In some embodiments, the computing device 120 may determine each of the one or more candidate regions as a region to be patched. For example, in the example of fig. 5, the computing device 120 may determine each of the

candidate regions

301, 303, 305, 307, 309, 311, 315, and 317 as regions to be patched. In other embodiments, the computing device 120 may perform further selection or filtering of the one or more candidate regions to determine a region to be repaired for which the point cloud data needs to be repaired. Such embodiments will be described in detail below.

By way of example process 400, based on whether point cloud data exists in a portion of the point cloud map 123 corresponding to the candidate area, the computing device 120 may exclude the candidate area in which point cloud data already exists from the point cloud map 123, thereby eliminating the possibility of performing point cloud data patching on the candidate area in which point cloud data has already been acquired, thereby simplifying subsequent processing of the computing device 120, saving processing resources of the computing device 120, and also avoiding a reduction in accuracy of the point cloud map 123 caused by replacing actually acquired point cloud data with fitted point cloud data.

As mentioned above, the computing device 120 may perform further selections of the one or more candidate regions obtained in block 430 of the example process 440 to determine the regions to be repaired in the point cloud map 123 that require patching of the point cloud data. In particular, in some embodiments, particularly when the target area 105 includes a road 101, the computing device 120 may be more focused on the road surface point cloud within the road 101 because the patched point cloud map 127 generated by the computing device 120 may subsequently be used for automated driving of vehicles that are typically traveling within the road 101. In this case, if the candidate region is determined to lack point cloud data in the corresponding portion in the point cloud map 123, the computing device 120 may further determine whether the candidate region is required as a region to be repaired based on the positional relationship of the candidate region with the road 101. In this way, the computing device 120 may select the region to be repaired more specifically, i.e., increase the effectiveness of the determined region to be repaired. Meanwhile, since a possibility that a part of the candidate region that does not necessarily need to be patched of the point cloud data is excluded as the region to be patched, the subsequent processing overhead of the computing device 120 can be reduced.

Specifically, in some embodiments, to select one or more regions to be patched from the one or more candidate regions, the computing device 120 may determine which of the one or more candidate regions is within the road 101 based on the road boundary point cloud corresponding to the

road boundary line

102 or 104 in the point cloud map 123. For example, in the example of fig. 5, based on the positional relationship of the road boundary point cloud 520 with the portion 501, the portion 503, the portion 505, the portion 507, and the portion 509, the computing device 120 may determine that the candidate region 309 depicted in fig. 3A-3D is inside the road 101, while the candidate region 301, the candidate region 303, the candidate region 305, and the candidate region 307 are outside the road 101. Similarly, based on the positional relationship of the road boundary point cloud 540 with the portion 511, the portion 515, and the portion 517, the computing device 120 may determine that the candidate region 311, the candidate region 315, and the candidate region 317 depicted in fig. 3A through 3D are outside the road 101. Hereinafter, for convenience of description, the candidate region 309 located inside the road 101 may also be referred to as a first candidate region. The computing device 120 may then determine a first candidate region 309 located within the road 101 as a region to be repaired requiring repair of the point cloud data.

In some cases, the candidate region may not be completely inside the road 101 nor completely outside the road 101, but may be partially inside the road 101 and partially outside the road 101. For example, the corresponding portion of the candidate region in the point cloud map 123 may intersect with the road

boundary point cloud

520 or 540. In such a case, the computing device 120 may determine a portion of the candidate region that is located within the road 101 as the region to be patched. In this way, the computing device 120 may selectively patch more important missing point cloud data located inside the road 101 in the point cloud map 123, and avoid patching point cloud data that is less significant outside the road 101 in the point cloud map 123, thereby simplifying the processing of the computing device 120 and saving the processing overhead of the computing device 120. Such an embodiment is described in detail below with reference to fig. 6A and 6B.

Fig. 6A shows a schematic diagram of the intersection of the candidate region 610 with the road boundary line 102 according to an embodiment of the present disclosure. In fig. 6A, only the road 101 and its related elements are shown for clarity, and other objects or elements are omitted. As shown in fig. 6A, it is assumed that the candidate region 610 corresponds to a certain static obstacle in the target region 105. For example, the static obstacle may be a vehicle parked at the roadside. However, the vehicle may not be parked regularly, but a part is located inside the road 101 and another part is located outside the road 101. Therefore, the candidate region 610 corresponding to the vehicle intersects the road boundary line 102 of the road 101, so that the candidate region 610 has a sub-region 612 located inside the road 101 and a sub-region 614 located outside the road 101. Hereinafter, for convenience of description, the candidate region 610 intersecting the road boundary line 102 of the road 101 may also be referred to as a second candidate region.

Fig. 6B shows a schematic diagram of a candidate region 610 intersecting a corresponding portion 620 in the point cloud map 123 with the road boundary point cloud 520, according to an embodiment of the disclosure. In fig. 6B, only the road

boundary point clouds

520 and 540 are shown for clarity, with point clouds of other objects or elements omitted. As shown in fig. 6B, the computing device 120 may determine that the candidate region 610 intersects the road boundary line 102 of the road 101 based on the road boundary point cloud 520, e.g., based on the positional relationship between the road boundary point cloud 520 and the portion 620. For example, in the point cloud map 123, the computing device 120 may determine that the road boundary point cloud 520 has just passed through the portion 610, such that the portion 610 has a sub-portion 622 located inside the road boundary point cloud 520 and a sub-portion 624 located outside the road boundary point cloud 520. Thus, referring back to fig. 6A, the computing device 120 may correspondingly determine a second candidate sub-region 612 of the second candidate region 610 within the road 101. The computing device 120 may then determine the second candidate sub-region 612 as a region to be repaired requiring repair of the point cloud data. In this way, the computing device 120 may selectively patch more important missing point cloud data located inside the road 101 in the point cloud map 123, and avoid patching point cloud data that is less significant outside the road 101 in the point cloud map 123, thereby simplifying the processing of the computing device 120 and saving the processing overhead of the computing device 120.

As mentioned above in describing block 230 of example method 200, in some embodiments, computing device 120 may determine a reference region of a region to be repaired by extending the region to be repaired. Such an embodiment will be further described below with reference to fig. 7.

Fig. 7 shows a schematic diagram of taking the middle region 715 between the extension region 720 and the region to be repaired 710 as a reference region according to an embodiment of the disclosure. In the example of fig. 7, a real point cloud map 700 is shown acquired by an acquisition device 110 (e.g., a drone carrying a lidar measurement device) for a certain real road. As shown in fig. 7, a region to be repaired 710 where point cloud data is missing exists in the vicinity of the road boundary point cloud 730. To determine a reference region for the region to be repaired 710, the computing device 120 may expand the region to be repaired 710 to obtain an expanded region 720. For example, the computing device 120 may select an appropriate extension distance, such as tens of centimeters, by which to move outward the respective edges of the region to be repaired 710.

In fig. 7, the computing device 120 may expand the other three edges of the area to be repaired 710, except for the edge coinciding with the road boundary point cloud 730, to properly cover the area around the area to be repaired 710 where the point cloud data exists. This is because even if the edge coinciding with the road boundary point cloud 730 expands outward, the expanded area 720 cannot cover a larger area where valid point cloud data exists. Of course, if the area to be repaired is located at the middle position of the road, the area to be repaired may also be expanded in all directions. After deriving the expanded region 720, the computing device 120 can determine an intermediate region 715 between the expanded region 720 and the region to be repaired 710 to serve as a reference region for determining missing point cloud data in the region to be repaired 710. In this way, the computing device 120 may use the point cloud data that is closest to the region to be repaired to patch the point cloud data missing in the region to be repaired, thereby improving the accuracy of the patched point cloud map 127.

Fig. 8 shows a block diagram of an example data processing apparatus 800 in accordance with an embodiment of the present disclosure. In some embodiments, the apparatus 800 may be included in the computing device 120 of fig. 1 or implemented as the computing device 120.

As shown in fig. 8, the example apparatus 800 may include a candidate region determination module 810, a region to be repaired determination module 820, and a missing point cloud data determination module 830. The candidate region determination module 810 may be configured to determine a plurality of candidate regions from the target region based on static obstacle information in a plurality of acquisition processes for acquiring point cloud data that construct a point cloud map of the target region. The to-be-repaired region determination module 820 may be configured to determine at least one region to be repaired from a plurality of candidate regions based on a plurality of portions of the point cloud map corresponding to the plurality of candidate regions. The missing point cloud data determining module 830 may be configured to determine missing point cloud data in at least one region to be repaired using point cloud data of a reference region in the point cloud map associated with the at least one region to be repaired.

In some embodiments, the to-be-repaired region determining module 820 may include: a point cloud data-free portion determination module configured to determine, among a plurality of portions, at least one portion in which point cloud data is missing; a point cloud-free data candidate region determination module configured to determine, among a plurality of candidate regions, at least one candidate region corresponding to at least one portion; and a region-to-be-repaired obtaining module configured to obtain at least one region-to-be-repaired based on the at least one candidate region.

In some embodiments, the region to be repaired obtaining module may include: a road boundary point cloud determination module configured to determine a road boundary point cloud corresponding to a boundary of a road in a point cloud map; and a to-be-repaired region selection module configured to select at least one to-be-repaired region from the at least one candidate region based on the road boundary point cloud.

In some embodiments, the region to be repaired selection module may include: a first in-road candidate region determination module configured to determine that a first candidate region of the at least one candidate region is within a road based on the road boundary point cloud; and a first in-road region-to-be-repaired determination module configured to determine the first candidate region as a region-to-be-repaired of the at least one region-to-be-repaired.

In some embodiments, the region to be repaired selection module may include: a second in-road candidate region determination module configured to determine a second candidate sub-region of the second candidate region within the road if it is determined that the second candidate region of the at least one candidate region intersects with a boundary of the road based on the road boundary point cloud; and a second in-road region-to-be-repaired determination module configured to determine the second candidate sub-region as a region-to-be-repaired of the at least one region-to-be-repaired.

In some embodiments, the candidate region determination module 810 may include: a static obstacle region set determination module configured to determine, based on the static obstacle information, a plurality of static obstacle region sets respectively corresponding to the plurality of times of acquisition processes; and a candidate region obtaining module configured to obtain a plurality of candidate regions based on an intersection of the plurality of sets of static obstacle regions.

In some embodiments, the example apparatus 800 may further include: a first extended region obtaining module configured to extend a first region to be repaired in at least one region to be repaired to obtain a first extended region; and a first reference region determination module configured to determine an intermediate region between the first extension region and the first region to be repaired as a first reference region corresponding to the first region to be repaired among the reference regions.

In some embodiments, the missing point cloud data determination module 830 may include: a plane fitting module configured to perform plane fitting using the point cloud data of the reference region to obtain missing point cloud data.

Fig. 9 illustrates a block diagram of an example device 900 that may be used to implement embodiments of the present disclosure. As shown in fig. 9, the example device 900 includes a Central Processing Unit (CPU)901 that can perform various appropriate actions and processes in accordance with computer program instructions stored in a read-only memory device (ROM)902 or loaded from a storage unit 908 into a random access memory device (RAM) 903. In the RAM 903, various programs and data required for operation of the example device 900 may also be stored. The CPU 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input/output (I/O) interface 905 is also connected to bus 904.

A number of components in the example device 900 are connected to the I/O interface 905, including: an input unit 906 such as a keyboard, a mouse, and the like; an output unit 907 such as various types of displays, speakers, and the like; a storage unit 908 such as a magnetic disk, optical disk, or the like; and a communication unit 909 such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 909 allows the example device 900 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunications networks.

The various processes and processes described above, e.g., the various example methods or example processes, may be performed by the processing unit 901. For example, in some embodiments, various example methods or example processes may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program may be loaded and/or installed onto example device 900 via ROM 902 and/or communications unit 909. When loaded into RAM 903 and executed by CPU 901, a computer program may perform one or more steps of the various example methods or example processes described above.

As used herein, the terms "comprises," comprising, "and the like are to be construed as open-ended inclusions, i.e.," including, but not limited to. The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first," "second," and the like may refer to different or the same object. Other explicit and implicit definitions may also be included herein.

As used herein, the term "determining" encompasses a wide variety of actions. For example, "determining" can include calculating, computing, processing, deriving, investigating, looking up (e.g., looking up in a table, a database or another data structure), ascertaining and the like. Further, "determining" can include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), and the like. Further, "determining" may include resolving, selecting, choosing, establishing, and the like.

It should be noted that the embodiments of the present disclosure can be realized by hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or specially designed hardware. Those skilled in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, for example such code provided in programmable memory or on a data carrier such as an optical or electronic signal carrier.

Further, while the operations of the methods of the present disclosure are depicted in the drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Rather, the steps depicted in the flowcharts may change the order of execution. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions. It should also be noted that the features and functions of two or more devices according to the present disclosure may be embodied in one device. Conversely, the features and functions of one apparatus described above may be further divided into embodiments by a plurality of apparatuses.

While the present disclosure has been described with reference to several particular embodiments, it is to be understood that the disclosure is not limited to the particular embodiments disclosed. The disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method of data processing, comprising:

determining a plurality of candidate areas from a target area based on static obstacle information in a plurality of acquisition processes, wherein the plurality of acquisition processes are used for acquiring point cloud data of a point cloud map for constructing the target area;

determining at least one region to be repaired from the plurality of candidate regions based on a plurality of portions of the point cloud map corresponding to the plurality of candidate regions; and

and determining missing point cloud data in the at least one region to be repaired by utilizing point cloud data of a reference region in the point cloud map, which is associated with the at least one region to be repaired.

2. The method of claim 1, wherein determining the at least one region to be repaired comprises:

determining, among the plurality of portions, at least one portion of the point cloud data that is missing;

determining at least one candidate region corresponding to the at least one portion among the plurality of candidate regions; and

obtaining the at least one region to be repaired based on the at least one candidate region.

3. The method of claim 2, wherein obtaining the at least one region to be repaired comprises:

determining a road boundary point cloud corresponding to the boundary of the road in the point cloud map; and

selecting the at least one region to be patched from the at least one candidate region based on the road boundary point cloud.

4. The method of claim 3, wherein selecting the at least one region to be repaired comprises:

determining, based on the road boundary point cloud, that a first candidate region of the at least one candidate region is within the road; and

and determining the first candidate region as a region to be repaired in the at least one region to be repaired.

5. The method of claim 3, wherein selecting the at least one region to be repaired comprises:

determining a second candidate sub-region of the second candidate region within the road if it is determined that a second candidate region of the at least one candidate region intersects the boundary of the road based on the road boundary point cloud; and

determining the second candidate sub-region as a region to be repaired in the at least one region to be repaired.

6. The method of claim 1, wherein determining the plurality of candidate regions from the target region comprises:

determining a plurality of static obstacle area sets respectively corresponding to the multiple acquisition processes based on the static obstacle information; and

obtaining the plurality of candidate regions based on an intersection of the plurality of sets of static obstacle regions.

7. The method of claim 1, further comprising:

expanding a first region to be repaired in the at least one region to be repaired to obtain a first expanded region; and

determining an intermediate region between the first extension region and the first region to be repaired as a first reference region corresponding to the first region to be repaired in the reference region.

8. The method of claim 1, wherein determining missing point cloud data in the at least one region to be repaired comprises:

performing a plane fitting using the point cloud data of the reference region to obtain the missing point cloud data.

9. A data processing apparatus comprising:

a candidate region determination module configured to determine a plurality of candidate regions from a target region based on static obstacle information in a plurality of acquisition processes for acquiring point cloud data constructing a point cloud map of the target region;

a region-to-be-repaired determining module configured to determine at least one region-to-be-repaired from the plurality of candidate regions based on a plurality of portions of the point cloud map corresponding to the plurality of candidate regions; and

a missing point cloud data determination module configured to determine missing point cloud data in the at least one region to be repaired by using point cloud data of a reference region in the point cloud map associated with the at least one region to be repaired.

10. The apparatus of claim 9, wherein the region to be repaired determination module comprises:

a point cloud data-free portion determination module configured to determine, among the plurality of portions, at least one portion where point cloud data is missing;

a point cloud-free data candidate region determination module configured to determine at least one candidate region corresponding to the at least one portion among the plurality of candidate regions; and

a region-to-be-repaired obtaining module configured to obtain the at least one region-to-be-repaired based on the at least one candidate region.

11. The apparatus of claim 10, wherein the region to be repaired obtaining module comprises:

a road boundary point cloud determination module configured to determine a road boundary point cloud corresponding to a boundary of a road in the point cloud map; and

a to-be-patched region selection module configured to select the at least one to-be-patched region from the at least one candidate region based on the road boundary point cloud.

12. The apparatus of claim 11, wherein the to-be-repaired region selection module comprises:

a first in-road candidate region determination module configured to determine that a first candidate region of the at least one candidate region is within the road based on the road boundary point cloud; and

a first in-road region-to-be-repaired determination module configured to determine the first candidate region as a region-to-be-repaired of the at least one region-to-be-repaired.

13. The apparatus of claim 11, wherein the to-be-repaired region selection module comprises:

a second in-road candidate region determination module configured to determine a second candidate sub-region of the second candidate region within the road if it is determined, based on the road boundary point cloud, that a second candidate region of the at least one candidate region intersects the boundary of the road; and

a second in-road region-to-be-repaired determination module configured to determine the second candidate sub-region as a region-to-be-repaired of the at least one region-to-be-repaired.

14. The apparatus of claim 9, wherein the candidate region determination module comprises:

a static obstacle region set determination module configured to determine, based on the static obstacle information, a plurality of static obstacle region sets respectively corresponding to the plurality of acquisition processes; and

a candidate region obtaining module configured to obtain the plurality of candidate regions based on an intersection of the plurality of sets of static obstacle regions.

15. The apparatus of claim 9, further comprising:

a first extended region obtaining module configured to extend a first region to be repaired in the at least one region to be repaired to obtain a first extended region; and

a first reference region determination module configured to determine an intermediate region between the first extension region and the first region to be repaired as a first reference region corresponding to the first region to be repaired among the reference regions.

16. The apparatus of claim 9, wherein the missing point cloud data determination module comprises:

a plane fitting module configured to perform a plane fitting using the point cloud data of the reference region to obtain the missing point cloud data.

17. An electronic device, comprising:

one or more processors; and

a memory storing computer-executable instructions that, when executed by the one or more processors, cause the electronic device to implement the method of any of claims 1-8.

18. A computer-readable medium having stored thereon computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the method of any one of claims 1 to 8.