WO2022259373A1 - 認識装置、認識方法、及び認識プログラム - Google Patents
認識装置、認識方法、及び認識プログラム Download PDFInfo
- Publication number
- WO2022259373A1 WO2022259373A1 PCT/JP2021/021778 JP2021021778W WO2022259373A1 WO 2022259373 A1 WO2022259373 A1 WO 2022259373A1 JP 2021021778 W JP2021021778 W JP 2021021778W WO 2022259373 A1 WO2022259373 A1 WO 2022259373A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- roadway
- change
- area
- boundary
- time
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims description 23
- 238000001514 detection method Methods 0.000 claims abstract description 25
- 230000008859 change Effects 0.000 claims description 61
- 238000012545 processing Methods 0.000 claims description 17
- 230000008569 process Effects 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 24
- 238000012937 correction Methods 0.000 description 23
- 238000005516 engineering process Methods 0.000 description 15
- 238000007493 shaping process Methods 0.000 description 13
- 238000004364 calculation method Methods 0.000 description 9
- 238000011156 evaluation Methods 0.000 description 9
- 239000003086 colorant Substances 0.000 description 7
- 230000000694 effects Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000005192 partition Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 241000283070 Equus zebra Species 0.000 description 3
- 230000006866 deterioration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000003973 paint Substances 0.000 description 1
- 238000000059 patterning Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/02—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to ambient conditions
- B60W40/06—Road conditions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
Definitions
- the technology disclosed relates to a recognition device, a recognition method, and a recognition program.
- Such a function may use an image captured by a camera mounted on the vehicle for recognizing the surrounding environment of the vehicle.
- the surrounding environment that cars should recognize is the boundary with the roadway or sidewalk, but there are various patterns such as straight lines, dotted lines, and combinations of these. However, there are cases where patterning is not possible. For example, various lane markings whose colors or shapes cannot be patterned, cases where lane markings do not exist, effects of shadows such as features, deterioration of lane markings over time, and effects of other vehicles.
- the boundary of the roadway includes the boundary between running lanes and the boundary between the roadway and a non-roadway (sidewalk).
- the disclosed technology has been made in view of the above points, and aims to provide a recognition device, a recognition method, and a recognition program capable of appropriately recognizing the surrounding environment of a running vehicle.
- An aspect of the present disclosure is a recognition device that detects a boundary of a roadway using time-series images captured by a camera mounted on a vehicle, and an acquisition unit that acquires the time-series images; classifying the regions in the acquired time-series images into the roadway, which is a region with a small amount of change, and the region other than the roadway, which is a region with a large amount of change, according to the magnitude of the amount of change; and a detection unit that detects a boundary of the area with the large amount of change as a boundary of the roadway.
- FIG. 4 is a diagram schematically showing means for recognizing the surrounding environment reflected in the drive recorder of the vehicle to be observed.
- FIG. 10 is a diagram showing an example of a case where a boundary cannot be patterned due to coloration or the like;
- FIG. 10 illustrates an example of a case where a boundary cannot be patterned due to shadow effects and lack of partition lines;
- FIG. 10 is a diagram showing an example of a case where even if there is a partition line, the boundary cannot be patterned;
- It which shows an example of the time series image image
- It is a block diagram showing a hardware configuration of a recognition device of the present disclosure. 2 is a block diagram showing the functional configuration of the recognition device of the present disclosure;
- FIG. 10 is a diagram showing an example of a case where a boundary cannot be patterned due to coloration or the like;
- FIG. 10 illustrates an example of a case where a boundary cannot be patterned due to shadow effects and lack
- FIG. 2 is an image diagram of a finite difference method; It is a figure which shows an example of calculation of a difference image.
- FIG. 10 is a diagram showing an example of a difference image calculated by emphasizing a boundary line; It is a difference image obtained by superimposing the obtained frame difference images within a certain range of time width.
- FIG. 10 is a diagram showing an example of correction for truncating pixels with low frequency in consideration of appearance frequency;
- FIG. 10 is a diagram showing an example of correction for truncating pixels with low frequency in consideration of appearance frequency;
- FIG. 10 is a diagram showing an example of correction for truncating pixels with low frequency in consideration of appearance frequency;
- FIG. 10 is a diagram showing an example of correction for truncating pixels with low frequency in consideration of appearance frequency;
- FIG. 10 is a diagram showing an example of correction for truncating pixels with low frequency in consideration of appearance frequency;
- FIG. 10 is a diagram showing an example of correction for truncating pixels
- FIG. 10 is a diagram showing an example of correction so as to divide the area of the difference image from the vanishing point; It is a figure which removes the area
- FIG. 4 is a diagram showing an example of segmentation of regions based on vanishing points; FIG.
- FIG. 10 is a diagram showing an example of region correction using a vanishing point; 4 is a flow chart showing the flow of recognition processing by the recognition device of the present disclosure; FIG. 10 is a diagram showing an example in which recognition of the surrounding environment of the present technology is utilized for alert notification to the driver or automatic driving; It is a figure showing an example at the time of utilizing recognition of surrounding environment of this art for object detection. It is a configuration example of a recognition device in a modified example of the present disclosure.
- the technology of the present disclosure can be applied to means for recognizing the surrounding environment reflected in the drive recorder of the vehicle to be observed, as shown in FIG.
- the surrounding environment can be recognized by classifying objects around the vehicle into objects that appear to move and objects that do not appear to move relatively, and by detecting lane markings and boundaries between roadways and sidewalks.
- FIG. 2 is a diagram showing an example of demarcation lines that cannot be patterned.
- L1 in FIG. 2 is a case in which characters are drawn on the road and a plurality of marking line patterns are included.
- L2 to L4 are cases in which a plurality of colored demarcation lines such as white lines, blue lines, and green lines are mixed.
- L4 is a bicycle lane area, and it is difficult to determine how far the roadway extends. In these cases, it is difficult to pattern and recognize boundaries.
- L5 in FIG. 3 is a diagram showing an example where there is no partition line and there is a shadow effect. Due to sunlight conditions and shadows, the "white" division lines do not always appear to be the same color. In addition, when a "white" partition line is used as a detection condition, a guardrail also falls under this condition, and may be detected as a boundary and lead to erroneous detection.
- L6 is a scene in which no lane marking exists at the boundary between the roadway and the sidewalk, and the boundary between the roadway and the sidewalk is a block, not a lane marking.
- FIG. 4 is a diagram showing an example of demarcation lines that cannot be patterned.
- L7 in FIG. 4 has a dashed line, it is difficult to identify the area because there is less than one lane between the dashed line and the block.
- L8 is a demarcation line that has faded due to deterioration over time.
- the left and right sides are usually dashed lines, but L9 is a case where, for example, even in the case of two or more lanes, the area before the intersection is a solid line.
- the technique of the present disclosure recognizes the boundary by capturing the changing surroundings. For example, if the color cannot be patterned or the shape cannot be patterned, the characteristics as a whole are similar even if the pattern cannot be patterned, so it is assumed that the change in the time-series image will be less than the background. be done.
- the present technology can be applied to a real space region having predetermined characteristics and other regions.
- a physical space region having a predetermined feature is a region in which a plurality of patterns are continuous or similar patterns are continuous, and the proportion of features that do not apply to the pattern is small. Therefore, the present technology may be applied to railroad tracks and the like.
- FIG. 5 is a diagram showing an example of time-series images captured by an in-vehicle camera. From the time-series images shown in FIG. 5, the following logics (1) to (6) can be explained. (1) The own vehicle moves. (2) With an in-vehicle camera of the own vehicle, the “surroundings" appear to be moving instead of the own vehicle. (3) Of the “surroundings", the roadway is made of the same material such as concrete, is constructed continuously over a certain section, and has similar colors that appear continuously, so that the color does not easily change. (4) Curb blocks, sidewalks, or clusters of buildings with respect to the roadway tend to change due to their various colors. (5) Therefore, it can be considered that the area with relatively large change is the surrounding area and not the roadway. (6) It is possible to recognize the surrounding environment from the magnitude of time-series change in pixels at the same coordinates or in the same region on the image.
- FIG. 6 is a block diagram showing the hardware configuration of the recognition device 100 of the present disclosure.
- the recognition device 100 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage 14, an input unit 15, a display unit 16, and a communication interface ( I/F) 17.
- a CPU Central Processing Unit
- ROM Read Only Memory
- RAM Random Access Memory
- storage 14 an input unit 15, a display unit 16, and a communication interface ( I/F) 17.
- I/F communication interface
- the CPU 11 is a central processing unit that executes various programs and controls each section. That is, the CPU 11 reads a program from the ROM 12 or the storage 14 and executes the program using the RAM 13 as a work area. The CPU 11 performs control of each configuration and various arithmetic processing according to programs stored in the ROM 12 or the storage 14 . In this embodiment, the ROM 12 or storage 14 stores a recognition program.
- the ROM 12 stores various programs and various data.
- the RAM 13 temporarily stores programs or data as a work area.
- the storage 14 is configured by a storage device such as a HDD (Hard Disk Drive) or SSD (Solid State Drive), and stores various programs including an operating system and various data.
- HDD Hard Disk Drive
- SSD Solid State Drive
- the input unit 15 includes a pointing device such as a mouse and a keyboard, and is used for various inputs.
- the display unit 16 is, for example, a liquid crystal display, and displays various information.
- the display unit 16 may employ a touch panel system and function as the input unit 15 .
- the communication interface 17 is an interface for communicating with other devices such as terminals.
- the communication uses, for example, a wired communication standard such as Ethernet (registered trademark) or FDDI, or a wireless communication standard such as 4G, 5G, or Wi-Fi (registered trademark).
- FIG. 7 is a block diagram showing the functional configuration of the recognition device 100 of the present disclosure.
- Each functional configuration is realized by the CPU 11 reading a recognition program stored in the ROM 12 or the storage 14, developing it in the RAM 13, and executing it.
- the recognition device 100 includes an acquisition unit 110, a detection unit 112, and a road database 114.
- the acquisition unit 110 of the recognition device 100 acquires time-series images captured by a camera mounted on a vehicle.
- the detection unit 112 is a processing unit that detects the boundary of the surrounding environment of the vehicle to be observed (hereinafter simply referred to as the vehicle) and outputs the detection result. It is divided into a portion 122 and an evaluation portion 124 .
- the road database 114 stores road information including the number of lanes on the road. As shown in FIG. 8, it is possible to specify which lane is the lane of the host vehicle by applying information on the number of lanes in the road information to the image.
- the detection unit 112 classifies the regions in the time-series images acquired by the acquisition unit 110 according to the magnitude of the change, and detects the boundary of the roadway. By classifying the area into the roadway, which is an area with a small amount of change, and the area other than the roadway, which is an area with a large amount of change, the boundary between the area with a small amount of change and the area with a large amount of change is detected as the boundary of the roadway. Processing of each unit in the detection unit 112 will be described.
- the difference calculation unit 120 calculates a difference image between time-series images.
- a frame difference method (hereinafter referred to as a difference method) is used to calculate the difference image.
- FIG. 9 shows an image diagram of the finite difference method.
- the difference method is a method of detecting a "changed pixel" by taking a logical product from the calculation results of absolute difference values between frames.
- the difference method can be applied in a method of detecting the surrounding environment with a camera fixedly installed on a moving body.
- FIG. 10 is a diagram showing an example of calculation of a difference image. Based on the image of time series t, the image of t+1, and the image of t+2, the difference image is binarized and calculated so as to emphasize the change at t+1. This method is based on the fact that the color of concrete roads basically looks the same except for construction marks, etc., and that the color of roads does not change.
- FIG. 11 is a diagram showing an example of a difference image calculated by emphasizing a boundary line.
- the upper part of FIG. 11 is a difference image obtained by binarizing the entire image with the same threshold according to the normal frame difference method. In this case, dark colors such as colored division lines disappear, and temporarily dark places such as “shade” are also detected.
- the lower part of FIG. 11 is a difference image calculated by emphasizing the boundary line. Since the method emphasizes the boundary line, only outline information is required for shade, and colored division lines can also be detected. If the marking line is truly evenly painted, only the outline information is detected, but in reality, there are dents or protrusions on the road surface, or paint rubbing, etc., so differences are also detected within the marking line.
- FIG. 12 is a difference image obtained by superimposing the obtained frame difference images within a certain time width range.
- the shaping section 122 corrects the difference image calculated by the difference calculation section 120 .
- a method suitable for each difference image is used as described below.
- FIG. 13 is an example of correction in which low-frequency pixels are cut off in consideration of appearance frequency.
- pixels with low frequency locally existing on roadways and sidewalks such as manholes and shades are removed. It also includes objects on the roadway that appear infrequently and are detected as large changes. In this case, pixels corresponding to lane markings, buildings, etc., which frequently change due to the influence of protrusions, dents, and slight lateral movement of the vehicle, remain without being removed.
- the boundary line between the roadway and the sidewalk can also be extracted from the changes in the color of the road shoulder and step blocks.
- the information of momentary shade and passing vehicles is reduced. Vehicles can be further reduced by applying processing to the rectangular areas detected by object detection.
- guardrails are monochromatic on their own, but because they are installed by pillars, the sidewalks and buildings in the background are naturally reflected between the pillars of the guardrails, and they change frequently.
- the sidewalk itself is colorful (for example, if it is composed of tiles of two or more colors and has a grid pattern or pattern), the pixel values change at high frequency, so the difference is emphasized and the division is further reduced. easier. In this way, black pixels are extracted from locations where the pixel value changes are small and the difference is unlikely to occur, and white pixels are extracted from locations where the pixel value changes are large and the difference is emphasized.
- the shaping unit 122 may perform correction by segmenting the area of the difference image from the vanishing point. With the vanishing point as the center, the area around the difference image is divided into, for example, triangular areas, and each area is divided into areas where the ratio of black pixels/white pixels has changed or has not changed. As shown in FIG. 16, if the installation position and angle of view information of the camera (the angle of view information is fixed for each model and for each internal setting) and the width of each lane on the road on which the vehicle is running are known, The position of the division line reflected in the image can be roughly specified.
- the shaping unit 122 may perform correction to remove the area in which the own vehicle is reflected in the difference image. As shown in FIG. 17, the lower area of the difference image is cut because the own vehicle may be reflected there. In this way, the shaping unit 122 performs processing so that the detection unit 112 does not detect the area in which the vehicle itself is photographed.
- the difference image is extracted with white in areas where the amount of change is large, and in black in areas with small amount of change.
- the evaluation unit 124 classifies the regions and detects boundaries by specifying the region of the driving lane of the host vehicle and the regions on the left and right thereof from the difference image corrected by the shaping unit 122 .
- the evaluation unit 124 determines a threshold value for distinguishing black and white in the area classification.
- the area of the difference image after correction is divided into a "black” area with a small amount of change and a "white” area with a large amount of change.
- the “white” area includes lines that are extracted as partition lines.
- the evaluation unit 124 In the classification of the area by the evaluation unit 124, for example, road information corresponding to the number of lanes of the road on which the vehicle is traveling is acquired from the road database 114, and based on the acquired road information and the extracted lane markings, the change Areas of low volume are classified as the driving lane of the host vehicle and the surrounding driving lanes. Then, the evaluation unit 124 classifies an area with a small amount of change as a roadway, and an area with a large amount of change as an area other than the roadway.
- FIG. 18 is a diagram for identifying and classifying regions from the corrected difference image.
- the region (1) specifies that the black region in the center of the image is the driving lane of the host vehicle. From the adjacent white demarcation line, it is specified that there are surrounding driving lanes in the area (2). Further, from the adjacent white division lines, it is specified that there is a driving lane in the area (3A) and that there is no driving lane in the area (3B). Furthermore, it is specified that there is no driving lane in the area (4) to the left of (3A).
- the area is classified into the roadway, which is an area where the amount of change is small, and the area other than the roadway, which is an area where the amount of change is large. is the boundary of . Also, referring to the number of driving lanes on the left and right and road information, it detects that the vehicle is driving on the third lane based on the information that this road is a "4-lane road on one side". can.
- Fig. 19 is an example of boundary correction when it is difficult to distinguish the boundary between the roadway and the sidewalk.
- a sidewalk may appear like a roadway if only similar colors appear continuously for a certain section, but elements such as road shoulders and curbs are “white” areas. Therefore, since these elements can be regarded as a fixed range of "white” area, the part of the fixed range of "white” area can be detected as a boundary.
- FIG. 20 is an example of correction for an area of a road that changes sharply, corresponding to the so-called zebra zone.
- a continuous area of the zebra zone can be corrected by the shaping unit 122 as an area with a large change.
- the narrow “black” area is regarded as “white” and filled in by the correction of the shaping section 122 .
- the shaping unit 122 corrects the width of the “black” area to a “white” area when the width is narrower than a predetermined length.
- FIG. 22 is an example of segmentation of areas based on vanishing points.
- the "black” area immediately below the vanishing point is regarded as the driving lane of the host vehicle. If the width of the "white” area on the left is narrow, and if there is a “black” area with a certain width or more on the left, it is determined that there is an adjacent driving lane. In FIG. 22 as well, due to the correction by the shaping unit 122, the narrow “black” area is regarded as “white” and painted over.
- FIG. 23 shows region correction using vanishing points. Assuming that the lane markings between the driving lanes cannot be recognized and that the wide "black” area is the driving lane of the vehicle, the left and right ratio of the "black” area is calculated, and if there is an extreme difference, it is divided into two areas. , and a “white” area is sandwiched between them.
- FIG. 24 is a flowchart showing the flow of recognition processing by the recognition device 100 of the present disclosure. Recognition processing is performed by the CPU 11 reading out the recognition program from the ROM 12 or the storage 14, developing it in the RAM 13, and executing it.
- step S100 the CPU 11, as the acquisition unit 110, acquires time-series images captured by a camera mounted on the vehicle.
- step S102 the CPU 11, as the difference calculation unit 120, calculates a difference image of the time-series images.
- step S104 the CPU 11, as the shaping unit 122, corrects the calculated difference image.
- step S106 the CPU 11, as the evaluation unit 124, classifies areas with a small amount of change into the driving lane of the host vehicle and the roadway of surrounding driving lanes, and classifies areas with a large amount of change into other than the roadway.
- the classification is performed by obtaining from the road database 114 road information corresponding to the number of lanes of the roadway on which the vehicle is traveling, and based on the road information and lane markings extracted from the corrected difference image.
- step S108 the CPU 11, as the evaluation unit 124, detects the boundary line between the lanes of the driving lane and the roadway boundary representing the boundary line between the roadway and the non-roadway from the classification result, and outputs the detection result.
- the recognition device 100 of the present embodiment it is possible to appropriately recognize the surrounding environment of the running vehicle.
- FIG. 25 is an example in which the recognition of the surrounding environment of this technology is utilized for alert notification to the driver or automatic driving.
- the autonomous vehicle misidentifies that "both left and right are dashed lines" and determines that it is traveling in the second lane, and changes lanes to the left, there is a risk of running onto the sidewalk. occur.
- the left side is actually a cycling zone, not a dashed line. If the recognition of the surrounding environment by this technology finds that the boundary of the roadway is just to the left of the driving lane of the host vehicle, such a problem will not occur. Also, if it is known that there is a cycle zone between the lane marking and the road boundary, it can issue an alert when the driver moves to park there.
- Fig. 26 is an example of using the recognition of the surrounding environment of this technology for object detection. For example, when a vehicle is detected as an object, the vehicle on the roadway and the vehicle on the site (parking lot, etc.) may be reflected at the same time. If the boundary line between the roadway and the sidewalk can be detected, even if the division line is not on the edge of the road or cannot be seen because it is rubbed, it is possible to determine whether or not the vehicle is on the roadway and classify the vehicle.
- a speed acquisition unit 130 that acquires the speed of the vehicle is provided, and the difference calculation unit 120 of the detection unit 112 calculates
- the differential image may be calculated using only the captured images.
- the reason why the images are selected according to the speed in this way is that the part of this technology that captures "the background is moving" due to the running of the vehicle to be observed is an important technology. For example, using vehicle speed information, a process of calculating a difference image using only images when the vehicle is traveling at a speed of 20 km/h or more is performed.
- the recognition processing executed by the CPU by reading the software (program) in the above embodiment may be executed by various processors other than the CPU.
- the processor is a PLD (Programmable Logic Device) whose circuit configuration can be changed after manufacturing, such as an FPGA (Field-Programmable Gate Array), and an ASIC (Application Specific Integrated Circuit) to execute specific processing.
- a dedicated electric circuit or the like which is a processor having a specially designed circuit configuration, is exemplified.
- the recognition processing may be performed by one of these various processors, or by a combination of two or more processors of the same or different type (for example, multiple FPGAs, a combination of a CPU and an FPGA, etc.). ) can be run.
- the hardware structure of these various processors is an electric circuit in which circuit elements such as semiconductor elements are combined.
- the recognition program is stored (installed) in advance in the storage 14 , but it is not limited to this.
- Programs are stored in non-transitory storage media such as CD-ROM (Compact Disk Read Only Memory), DVD-ROM (Digital Versatile Disk Read Only Memory), and USB (Universal Serial Bus) memory.
- CD-ROM Compact Disk Read Only Memory
- DVD-ROM Digital Versatile Disk Read Only Memory
- USB Universal Serial Bus
- the program may be downloaded from an external device via a network.
- a recognition device that detects the boundary of a roadway using time-series images captured by a camera mounted on a vehicle, obtaining the time-series images; Areas in the acquired time-series images are classified according to the magnitude of the amount of change into an area with a small amount of change, the roadway, and an area with a large amount of change, other than the roadway. Detecting the boundary of the area with a large amount of change as the boundary of the roadway;
- a recognizer configured to:
- a non-temporary storage medium storing a computer-executable program for executing recognition processing for detecting roadway boundaries using time-series images captured by a camera mounted on a vehicle, obtaining the time-series images; Areas in the acquired time-series images are classified according to the magnitude of the amount of change into an area with a small amount of change, the roadway, and an area with a large amount of change, other than the roadway. Detecting the boundary of the area with a large amount of change as the boundary of the roadway; Non-transitory storage media.
- recognition device 110 acquisition unit 112 detection unit 114 road database 120 difference calculation unit 122 shaping unit 124 evaluation unit 130 speed acquisition unit
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Automation & Control Theory (AREA)
- Mathematical Physics (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
Description
(1)自車両が動く。
(2)自車両の車載カメラでは、自車両ではなく「周囲」が動いているように見える。
(3)「周囲」のうち、車道はコンクリートのように同じ素材で作られ、一定区間連続して構築され、近しい色が連続して出現してくるため、色合いが変化しにくい。
(4)車道に対して縁石ブロック、歩道、又はビル群は色合いが様々であるため変化しやすい。
(5)よって相対的に変化が大きい領域は周囲であり、車道ではない、と捉えられる。
(6)同一座標の画素若しくは画像上で同一の領域における時系列での変化の大きさから、周辺環境を認識することが可能である。
上記実施形態の変形例として、図27に示すように、車両の速度を取得する速度取得部130を有するようにし、検出部112の差分算出部120は、速度が所定の閾値以上であるときに撮影された画像のみを用いるように差分画像を算出するようにしてもよい。このように速度によって画像を選別するのは、本技術が、観測対象の車両が走行することにより、「背景が動いている」ことを捉える部分が肝要な技術だからである。例えば、車両の速度の情報を用いて、時速20km/h以上で走行している場合の画像のみを用いて差分画像を算出する、といった処理を行う。
メモリと、
前記メモリに接続された少なくとも1つのプロセッサと、
を含み、
前記プロセッサは、
車両に搭載されたカメラから撮影された時系列画像を用いて車道の境界を検出する認識装置であって、
前記時系列画像を取得し、
取得された時系列画像内における領域を、変化量の大きさに応じて変化量が少ない領域である車道と変化量が大きい領域である車道以外とに分類し、前記変化量が少ない領域と前記変化量が大きい領域の境界を前記車道の境界として検出する、
ように構成されている認識装置。
車両に搭載されたカメラから撮影された時系列画像を用いて車道の境界を検出する認識処理を実行するようにコンピュータによって実行可能なプログラムを記憶した非一時的記憶媒体であって、
前記時系列画像を取得し、
取得された時系列画像内における領域を、変化量の大きさに応じて変化量が少ない領域である車道と変化量が大きい領域である車道以外とに分類し、前記変化量が少ない領域と前記変化量が大きい領域の境界を前記車道の境界として検出する、
非一時的記憶媒体。
110 取得部
112 検出部
114 道路データベース
120 差分算出部
122 整形部
124 評価部
130 速度取得部
Claims (8)
- 車両に搭載されたカメラから撮影された時系列画像を用いて車道の境界を検出する認識装置であって、
前記時系列画像を取得する取得部と、
取得された時系列画像内における領域を、変化量の大きさに応じて変化量が少ない領域である車道と変化量が大きい領域である車道以外とに分類し、前記変化量が少ない領域と前記変化量が大きい領域の境界を前記車道の境界として検出する検出部と、
を有する認識装置。 - 前記検出部は、さらに前記時系列画像から前記車両自体が撮影されている領域を前記検出部が検出する対象とならないよう処理を施す請求項1記載の認識装置。
- 前記車両の速度を取得する速度取得部を有し、
前記検出部は、前記速度が所定の閾値以上であるときに撮影された画像のみを用いる請求項1又は請求項2記載の認識装置。 - 前記検出部は、前記時系列画像の差分画像を算出し、前記差分画像の画素の変化により領域の前記変化量を算出する請求項1~請求項3の何れか1項に記載の認識装置。
- 前記差分画像の二値の一方の領域の幅が所定の長さより狭い場合に、当該領域を二値の他方の領域に補正する請求項4に記載の認識装置。
- 前記検出部は、前記車両が走行中の車道の車線数に対応する道路情報を取得し、前記道路情報と、前記時系列画像において抽出される区画線とに基づいて、前記変化量が少ない領域として前記車両の走行レーン、及び周囲の走行レーンを特定する請求項1~請求項5の何れか1項に記載の認識装置。
- 車両に搭載されたカメラから撮影された時系列画像を用いて車道の境界を検出するコンピュータにおける認識方法であって、
前記時系列画像を取得し、
取得された時系列画像内における領域を、変化量の大きさに応じて変化量が少ない領域である車道と変化量が大きい領域である車道以外とに分類し、前記変化量が少ない領域と前記変化量が大きい領域の境界を前記車道の境界として検出する、
処理をコンピュータに実行させる認識方法。 - 車両に搭載されたカメラから撮影された時系列画像を用いて車道の境界を検出するコンピュータにおける認識プログラムであって、
前記時系列画像を取得し、
取得された時系列画像内における領域を、変化量の大きさに応じて変化量が少ない領域である車道と変化量が大きい領域である車道以外とに分類し、前記変化量が少ない領域と前記変化量が大きい領域の境界を前記車道の境界として検出する、
処理をコンピュータに実行させる認識プログラム。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/021778 WO2022259373A1 (ja) | 2021-06-08 | 2021-06-08 | 認識装置、認識方法、及び認識プログラム |
JP2023527211A JPWO2022259373A1 (ja) | 2021-06-08 | 2021-06-08 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/021778 WO2022259373A1 (ja) | 2021-06-08 | 2021-06-08 | 認識装置、認識方法、及び認識プログラム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022259373A1 true WO2022259373A1 (ja) | 2022-12-15 |
Family
ID=84425833
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/021778 WO2022259373A1 (ja) | 2021-06-08 | 2021-06-08 | 認識装置、認識方法、及び認識プログラム |
Country Status (2)
Country | Link |
---|---|
JP (1) | JPWO2022259373A1 (ja) |
WO (1) | WO2022259373A1 (ja) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008262333A (ja) * | 2007-04-11 | 2008-10-30 | Nissan Motor Co Ltd | 路面判別装置および路面判別方法 |
JP2013051987A (ja) * | 2011-08-31 | 2013-03-21 | Olympus Corp | 画像処理装置、画像処理方法、及び画像処理プログラム |
JP2015149028A (ja) * | 2014-02-07 | 2015-08-20 | トヨタ自動車株式会社 | 区画線検出システム及び区画線検出方法 |
JP2020113207A (ja) * | 2019-01-16 | 2020-07-27 | アルパイン株式会社 | 車載装置、レーン関連処理方法 |
-
2021
- 2021-06-08 JP JP2023527211A patent/JPWO2022259373A1/ja active Pending
- 2021-06-08 WO PCT/JP2021/021778 patent/WO2022259373A1/ja active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2008262333A (ja) * | 2007-04-11 | 2008-10-30 | Nissan Motor Co Ltd | 路面判別装置および路面判別方法 |
JP2013051987A (ja) * | 2011-08-31 | 2013-03-21 | Olympus Corp | 画像処理装置、画像処理方法、及び画像処理プログラム |
JP2015149028A (ja) * | 2014-02-07 | 2015-08-20 | トヨタ自動車株式会社 | 区画線検出システム及び区画線検出方法 |
JP2020113207A (ja) * | 2019-01-16 | 2020-07-27 | アルパイン株式会社 | 車載装置、レーン関連処理方法 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2022259373A1 (ja) | 2022-12-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111212772B (zh) | 用于确定车辆的驾驶策略的方法和设备 | |
Wen et al. | A deep learning framework for road marking extraction, classification and completion from mobile laser scanning point clouds | |
CN107609491B (zh) | 一种基于卷积神经网络的车辆违章停车检测方法 | |
WO2020000251A1 (zh) | 基于摄像机协同接力的路口违章视频识别方法 | |
US10552706B2 (en) | Attachable matter detection apparatus and attachable matter detection method | |
CN103077617B (zh) | 基于计算机视觉的人行横道智能交通灯监管***及方法 | |
KR101605514B1 (ko) | 차선 인식 장치 및 방법 | |
CN105678285B (zh) | 一种自适应的道路鸟瞰图变换方法和道路车道检测方法 | |
CN103593649B (zh) | 一种用于车道偏离预警的车道线检测方法 | |
CN101702263B (zh) | 一种人行横道信号灯绿波自适应控制***及方法 | |
US20160176358A1 (en) | Early detection of turning condition identification using perception technology | |
CN103996053A (zh) | 基于机器视觉的车道偏离报警方法 | |
US20100098290A1 (en) | Method for detecting a clear path through topographical variation analysis | |
WO2014017517A1 (ja) | 立体物検出装置及び立体物検出方法 | |
CN103778786A (zh) | 一种基于显著车辆部件模型的交通违章检测方法 | |
US8718329B2 (en) | Top-down view classification in clear path detection | |
CN105005771A (zh) | 一种基于光流点轨迹统计的车道线实线的检测方法 | |
CN103714538A (zh) | 道路边缘检测方法、装置及车辆 | |
CN102592454A (zh) | 一种基于车辆侧面与路面交线检测的交叉口车辆运动参数测量方法 | |
CN108520635B (zh) | 一种混合现实道路显示*** | |
CN106778548A (zh) | 用于检测障碍物的方法和装置 | |
CN105654073A (zh) | 一种基于视觉检测的速度自动控制方法 | |
CN112349087B (zh) | 一种基于路口信息全息感知的可视化数据输入方法 | |
CN106127177A (zh) | 一种无人驾驶压路机 | |
JP6003986B2 (ja) | 立体物検出装置、立体物検出方法及び異物検出装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21945051 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023527211 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18566154 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 21945051 Country of ref document: EP Kind code of ref document: A1 |