CN108184062B - High-speed tracking system and method based on multi-level heterogeneous parallel processing - Google Patents
High-speed tracking system and method based on multi-level heterogeneous parallel processing Download PDFInfo
- Publication number
- CN108184062B CN108184062B CN201711498701.2A CN201711498701A CN108184062B CN 108184062 B CN108184062 B CN 108184062B CN 201711498701 A CN201711498701 A CN 201711498701A CN 108184062 B CN108184062 B CN 108184062B
- Authority
- CN
- China
- Prior art keywords
- module
- speed
- image
- level
- image data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/67—Focus control based on electronic image sensor signals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/80—Camera processing pipelines; Components thereof
- H04N23/81—Camera processing pipelines; Components thereof for suppressing or minimising disturbance in the image signal generation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/14—Picture signal circuitry for video frequency region
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Studio Devices (AREA)
- Image Processing (AREA)
Abstract
A high-speed tracking system and method based on multilevel heterogeneous parallel processing. The high-speed tracking system comprises a high-speed camera with an autonomous focusing function, a two-dimensional servo execution control response module, a two-dimensional servo module, an image data acquisition and transmission module and an image processing module. The main working flow is as follows: the image data acquisition and transmission module acquires image data through the high-speed camera, packages the image data and transmits the image data to the image processing module, and the image processing module calculates the image data according to different algorithms to obtain position information of a target and rotation angle information of the servo module; the servo module tracks the target according to the rotation angle transmitted from the image processing module, and the high-speed camera with the automatic focusing function performs automatic focusing according to the target position information. The system and the method can be applied to high-speed target identification and detection scenes in production and life, and can realize stable tracking of high-speed moving targets in complex scenes.
Description
Technical Field
The invention relates to the field of programmable vision chips and high-speed object detection and tracking, in particular to a high-speed tracking system and a high-speed tracking method based on multi-level heterogeneous parallel processing, which have the advantages of high speed, high integration, low power consumption and low cost, and can be applied to the tracking of high-speed targets in the fields of safety monitoring, industrial production and life.
Background
With the development of computer application and machine vision, a high-speed visual target tracking technology for acquiring images through a camera and detecting and tracking high-speed targets is greatly improved and widely applied to industrial and agricultural production and life. According to the type of the camera, the camera of the current visual target tracking system adopts a traditional image sensor to acquire image data, so that the frame rate of image acquisition is low, and the image acquisition resolution is low; according to the type of processors, the current tracking system of the visual target based on a computer/general processor is low in processing speed due to the limitation of transmission speed and lack of targeted support of a tracking method, and cannot be applied to the field of high-speed visual tracking with high real-time requirement.
Disclosure of Invention
It is therefore an objective of the claimed invention to provide a high speed tracking system and method, which at least partially solve at least one of the above problems.
In order to achieve the above object, as an aspect of the present invention, a high speed tracking system is provided, wherein the high speed tracking system includes a high speed image acquisition device, a two-dimensional servo execution control response module, a two-dimensional servo module, an image data acquisition and transmission module, and an image processing module; wherein:
the high-speed image acquisition device has an autonomous focusing function;
image signals acquired by the high-speed image acquisition device are input to the image processing module through the image data acquisition and transmission module;
the two-dimensional servo execution control response module generates a control instruction of the two-dimensional servo module based on the control information output by the image processing module and sends the control instruction to the two-dimensional servo module;
the image processing module integrates pixel level, block level and distributed multi-level heterogeneous parallel processing units and has the processing functions of image feature detection, feature extraction, feature classification and deep learning convolution algorithm; the image processing module completes a pixel level parallel algorithm through a pixel processing unit PE array of a pixel level processor, then accesses an image block in the pixel processing unit PE array of the pixel level processor through a block processing unit PPU in the block level parallel processor to complete a local feature extraction algorithm, and finally utilizes distributed computation to realize rapid classification of extracted local features to generate the control information, thereby realizing parallel processing based on multilevel isomerism.
As another aspect of the present invention, there is provided a high-speed tracing method based on multi-level heterogeneous parallel processing, which is implemented by using the high-speed tracing system described above, wherein the high-speed tracing method includes the following steps:
when a tracking target appears in the visual field of the high-speed image acquisition device with the autonomous focusing function, the image information acquisition and transmission module receives image data from the high-speed image acquisition device, and transmits the image data to the image processing module after image denoising processing;
the image processing module performs pixel-level, block-level and distributed parallel processing on the image data to obtain position information of the tracking target, transmits the position information to the two-dimensional servo execution response control module, judges the position information at the same time to obtain a focusing coefficient and transmits the focusing coefficient to the high-speed image acquisition device with the autonomous focusing function, and completes autonomous focusing of the high-speed image acquisition device;
the two-dimensional servo execution response control module calculates the position information of the tracking target to obtain direction information and rotation angle information; the two-dimensional servo module receives the direction information and the rotation angle information from the two-dimensional servo execution response control module and then correspondingly rotates;
meanwhile, the image processing module displays the tracking target information on a display module.
Compared with the prior art, the invention has the beneficial effects that:
(1) the high-speed camera with the automatic focusing function and integrated with the high-speed image sensor can acquire picture data with the frame rate of 1000 fps; the method can be applied to scenes of high-speed target identification and detection in production and life; the high-speed camera has an autonomous focusing function, and can automatically acquire a target to the central position of a view field through the focusing function after image processing is carried out by a multi-level heterogeneous parallel processing image processing module;
(2) the image acquisition and transmission module acquires the image data and performs noise reduction on the image data at the same time, so that the identification of the target based on the multi-level heterogeneous parallel processing image processing module is accelerated;
(3) the image processing module based on multilevel heterogeneous parallel processing has pixel level, block level and distributed parallel processing capability, can flexibly select image blocks and can work in a multi-instruction multi-data mode, and the speed of target identification in image data processing is increased;
(4) the high-speed tracking system based on multilevel heterogeneous parallel processing has the advantages of high speed, high integration, low power consumption and low cost, and can be applied to the fields of production and life, scientific research, sports and the like;
(5) by designing a target tracking algorithm corresponding to the target tracking algorithm, the robust tracking of the high-speed moving target under a complex scene can be realized.
Drawings
FIG. 1 is a schematic diagram of a high-speed tracking system according to the present invention;
FIG. 2 is a flowchart illustrating the operation of the high-speed tracking system during tracking;
FIG. 3 is a flowchart of the image processing module of the present invention performing multi-level heterogeneous parallel processing on image data.
Detailed Description
The invention discloses a high-speed tracking system, which comprises a high-speed image acquisition device, a two-dimensional servo execution control response module, a two-dimensional servo module, an image data acquisition and transmission module and an image processing module, wherein the two-dimensional servo execution control response module comprises a first servo module, a second servo module, a third servo module, a fourth servo module and a fourth servo module, wherein the first servo:
the high-speed image acquisition device has an autonomous focusing function, for example, a high-speed camera, a high-speed CMOS image sensor is adopted as a pixel device, and the picture reading rate is 1000 fps. The optical lens adopts an optical lens with an autonomous focusing function, and after image data is processed by an image processing module, autonomous and real-time focusing is realized according to environment and a target algorithm, so that the imaging quality of the high-speed image acquisition device is ensured.
And the two-dimensional servo execution control response module is used for connecting the image processing module with the two-dimensional servo module, preferably by a bus mode.
The two-dimensional servo module comprises an upper motor and a lower motor which have stepping functions, wherein the high-speed image acquisition device with automatic focusing is fixed on the upper motor in a mechanical mode, and the upper motor is arranged on the lower motor. The upper motor realizes the up-and-down rotation of the high-speed image acquisition device, and the lower motor realizes the three-dimensional rotation of the whole two-dimensional servo module through the left-and-right rotation.
The image processing module integrates a pixel-level heterogeneous parallel processor, a block-level heterogeneous parallel processor and a distributed multi-level heterogeneous parallel processor in the framework, has the processing functions of image feature detection, feature extraction, feature classification and deep learning algorithm convolution, and can efficiently and flexibly complete various typical and currently developed visual processing algorithms; the instruction set is a special visual processing instruction set based on a multi-level heterogeneous parallel processing architecture, programming of pixel level, block level and distributed parallel operation is supported, and a block processing unit circuit, a lane circuit and a cache control circuit are designed to realize the function of the instruction set.
The image processing module comprises a pixel-level parallel processor, a block-level parallel processor and a distributed parallel processor, and the pixel-level parallel processor and the distributed parallel processor are completed by using the same hardware through time division multiplexing. The pixel processing unit PE (processing element) array of the pixel level processor completes pixel level parallel algorithms such as filtering, morphology, background subtraction, feature point detection, feature generation and the like, then a block processing unit (PPU) in the block level parallel processor accesses image blocks in a pixel processing unit PE subarray to complete a local feature extraction algorithm, and the extracted features are rapidly classified by utilizing distributed computation, so that parallel processing based on multilevel isomerism is realized.
And the display module is connected with the image processing module through an HDMI (DP) interface. The method comprises the steps of displaying a target and background information according to the operation result of an image processing module on image data, and marking the target in real time.
And the power distribution module is used for distributing power for all modules of the whole system and ensuring the stable work of the system.
In the scheme, the high-speed image sensor in the high-speed camera machine with the automatic focusing function and integrated with the high-speed image sensor is the high-speed image sensor with the image acquisition speed of more than or equal to 1000 fps; the automatic focusing function of the system automatically focuses according to the environment and the target requirement after the image processor module based on multi-level heterogeneous parallel processing processes the image data and obtains the target position information.
In the scheme, the image data acquisition and transmission module is used for fully connecting the image sensor in the high-speed camera with the automatic focusing function and integrated with the high-speed image sensor, receiving the image data acquired by the high-speed camera with the automatic focusing function and integrated with the high-speed image sensor in real time and carrying out simple noise reduction processing on the data, and transmitting the processed image data to the image processor module based on multi-layer heterogeneous parallel processing through optical fiber communication.
In a preferred embodiment, the image processor module based on multi-level heterogeneous parallel processing comprises a system bus, a sensor control module, a sensor cache, a program cache, a network interface, a computing core for massive parallel processing and an MPU microprocessor. The sensor cache is used to store data from the image sensor for multiple accesses by the processor. The program buffer stores therein an algorithm program for visual processing. The network interface is used for communicating with the PC. The computing core provides main computing performance for a multi-level heterogeneous parallel processing architecture and can realize multi-instruction multi-data work. The same or completely different image data operations may be performed between the compute cores, depending on the control instructions of the MPU.
In the scheme, the image processing module comprises a processor, an instruction set of the processor is a special visual processing instruction set based on a multi-level heterogeneous parallel processing architecture, programming of pixel-level, block-level and distributed parallel operation is supported, and a block processing unit circuit, a processing unit circuit and a cache control circuit are designed to realize the function of the instruction set;
in the scheme, the pixel-level parallel processing and the distributed parallel processing are completed by time-sharing multiplexing and using the same hardware; preferably, the pixel-level parallel processing includes operations such as filtering, morphology, background subtraction, feature point detection, and feature generation.
In the scheme, the two-dimensional servo execution response control module is connected with the image processor module for multi-layer heterogeneous parallel processing, and after the two-dimensional servo execution response control is used for processing image data to obtain target position information, the rotation angle required by the servo module is calculated, and corresponding direction information and rotation pulses are sent.
In the scheme, the two-dimensional servo module can realize 360-degree rotation in the whole environment and is connected with the two-dimensional servo execution response control module, and after the direction information reflecting the rotation angle and the rotation pulse are sent to the servo module, the servo module rotates in the corresponding direction, so that the high-speed camera carried on the two-dimensional servo module is ensured to display the target in the camera view field in real time.
In the scheme, the display module is connected with the image processor module based on multi-layer heterogeneous parallel processing through an HDMI (DP) interface, and after the image data processing is finished, the target is displayed on the display module by the multi-layer heterogeneous parallel processing image processor module.
In the scheme, the high-speed camera with the independent focusing function and integrated with the high-speed image sensor in the system can be in real-time communication with the upper computer through the USB3.0, and the processor based on multi-layer heterogeneous parallel processing can be in real-time communication with the upper computer through the PCIe protocol and the internet access protocol.
The invention also discloses a high-speed tracking method based on multi-level heterogeneous parallel processing, which is realized by adopting the high-speed tracking system, and the high-speed tracking method comprises the following steps:
when a tracking target appears in the visual field of the high-speed image acquisition device with the autonomous focusing function, the image information acquisition and transmission module receives image data from the high-speed image acquisition device, and transmits the image data to the image processing module after image denoising processing;
the image processing module performs pixel-level, block-level and distributed parallel processing on the image data to obtain position information of the tracking target, transmits the position information to the two-dimensional servo execution response control module, judges the position information at the same time to obtain a focusing coefficient and transmits the focusing coefficient to the high-speed image acquisition device with the autonomous focusing function, and completes autonomous focusing of the high-speed image acquisition device;
the two-dimensional servo execution response control module calculates the position information of the tracking target to obtain direction information and rotation angle information; the two-dimensional servo module receives the direction information and the rotation angle information from the two-dimensional servo execution response control module and then correspondingly rotates;
meanwhile, the image processing module displays the tracking target information on a display module.
Preferably, the high-speed tracking method further includes the steps of:
when the tracking target is displayed on the display module, judging whether the tracking target is lost or not at regular intervals, if so, matching the two-dimensional servo execution control response module with the two-dimensional servo module to execute tracking action to search the track of the tracking target; if not, the two-dimensional servo execution control response module and the two-dimensional servo module are matched to execute locking action, and the tracking target is locked in the visual field of the high-speed image acquisition device with the self-focusing function.
In a preferred embodiment, the invention adopts a high-speed camera with an autonomous focusing function and a field-programmable high-speed visual target tracking processor with a multi-level heterogeneous parallel processing architecture; in order to better support the work between the high-speed camera with the autonomous focusing function and the field-programmable high-speed visual target tracking processor with a multi-level heterogeneous parallel processing framework, the high-speed camera with the autonomous focusing function can process high-frame-rate images by the field-programmable high-speed visual target tracking processor under the condition of high-frame-rate image acquisition, and a high-speed target tracking system based on multi-level heterogeneous parallel processing is designed. The system has the characteristics of high integration level, high performance, low cost and low power consumption, and realizes the stable tracking of the high-speed moving target in a complex scene by designing a target tracking algorithm corresponding to the system.
In order that the objects, technical solutions and advantages of the present invention will become more apparent, the present invention will be further described in detail with reference to the accompanying drawings in conjunction with the following specific embodiments.
The overall structure of the high-speed target tracking system based on multi-level heterogeneous parallel processing is shown in fig. 1. The method for tracking the high-speed target by adopting the high-speed tracking system is that the tracking target appears in the field of view of a high-speed camera with an autonomous focusing function 1 in figure 1; the image information acquisition and transmission module integrated in the 1 in the figure 1 receives image data from a high-speed CMOS image sensor of a camera, and transmits the image data to the 2 in the figure 1 after simple image noise reduction processing, and the image processing module is based on multi-level heterogeneous parallel processing; the image processing module based on multi-level heterogeneous parallel processing performs pixel level, block level and distributed parallel processing on an image to obtain position information of a high-speed target, the position information is respectively transmitted to 4 in the figure 1, the two-dimensional servo execution response control module judges the position information at the same time to obtain a focusing coefficient and transmits the focusing coefficient to 1 in the figure 1, and the camera with the function of autonomous focusing completes the autonomous focusing of the high-speed camera; in 4 of fig. 1, the two-dimensional servo execution response control module calculates the position information of the target to obtain direction information and rotation angle information; in fig. 1, 5, after receiving the direction information and the rotation angle pulse information from the two-dimensional servo execution response control module, the two-dimensional servo module performs corresponding rotation; meanwhile, 2 in fig. 1, the image processing module based on multi-level heterogeneous parallel processing displays the high-speed target information on 3 in fig. 1, the display module.
1. High-speed camera with autonomous focusing function
The high-speed camera with the autonomous focusing function consists of a high-speed image sensor, an optical lens with the autonomous focusing function and a high-speed image acquisition and transmission module, can finish 1000fps high-speed image acquisition and transmission, and can finish online autonomous focusing; and the image information acquisition and transmission module behind the high-speed camera acquires and transmits data of the image sensor, and can simply reduce noise of the image data.
2. Image processing module based on multilevel heterogeneous parallel processing
The image processing module based on multilevel heterogeneous parallel processing is composed of a pixel-level processor, a block-level processor, distributed parallel processing and an MPU microprocessor.
The image processing module of multilevel heterogeneous parallel processing can perform original pixel level image parallel processing, image parallel processing among local image blocks and distributed parallel processing. The image processing module is composed of an image sensor cache, a program cache, a system bus, a calculation core, an MPU microprocessor and the like. The image sensor caches and stores images acquired from the image sensor, the program cache stores a program for processing the images, and the system bus is used as a link between the image processing program and the computing core and controls the computing core to process the images. The computing core comprises a pixel-level processor capable of performing pixel-level parallel processing function and a block-level processor capable of performing block-level local feature extraction, the pixel-level processor and the block-level processor access the same hardware to complete the operation through time division multiplexing, and the MPU is mainly used for system management. The image processing module based on multilevel heterogeneous parallel processing comprises the following main processes: after the image sensor collects and transmits an image to a sensor cache, pixel-level parallel algorithms such as filtering and feature generation are completed through a pixel-level processor, and then a block-level processor accesses image blocks stored in a pixel-level processor sub-array and provides a program control instruction to execute a local feature extraction algorithm to obtain local statistical features; after the processing of the computation block, the features extracted by the block-level processor can be classified by adopting a distributed parallel computation mode. Thereby realizing the operation of pixel-level image processor, block-level processor and distributed parallel processing multi-level heterogeneous parallel image processing.
3. Display module
The display module is connected with the image processing module for multilevel heterogeneous parallel processing and is responsible for displaying the image and the position information of the high-speed target determined by the display module.
4. Two-dimensional servo execution response control module
The image processing module for multilevel heterogeneous parallel processing sends the position information and the angle information to the two-dimensional servo execution response control module, and the two-dimensional servo execution response control module obtains the direction and the pulse of the two-dimensional servo module which need to rotate through calculation and controls the two-dimensional servo module.
5. Two-dimensional servo module
The two-dimensional servo module is composed of an upper motor and a lower motor, the space is divided into (x, y, z) three-dimensional planes, the lower motor moves on a bottom plane, namely the (x, y) plane, the upper motor can move in the (x, z) plane and the (y, z) plane respectively according to the rotating position of the lower motor, and finally the two-dimensional servo module can freely rotate in the 360-degree area of the full range of the three-dimensional space.
6. Power distribution module
Each module needs a stable and continuous voltage, and in order to make the system work stably, the power distribution module supplies stable power to other modules of the system.
As an application of the present invention, for example, for automatically recognizing and tracking a moving target even a high-speed moving target, which is automatically classified and recognized by a neural network classifier in an image processing module of the present invention, after the high-speed target tracking system of the present invention is started, the system automatically recognizes the moving target from an image captured by a high-speed camera with an autonomous focusing function, and performs subsequent tracking and locking work.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention and are not intended to limit the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A high-speed tracking system is characterized by comprising a high-speed image acquisition device, a two-dimensional servo execution control response module, a two-dimensional servo module, an image data acquisition and transmission module and an image processing module; wherein:
the high-speed image acquisition device adopts an optical lens with an autonomous focusing function, and after image data is processed by an image processing module, the optical lens can realize autonomous and real-time focusing according to environment and a target algorithm, so that the imaging quality of the high-speed image acquisition device is ensured;
the image data acquisition and transmission module is fully connected with an image sensor in the high-speed image acquisition device, receives image data acquired by the image sensor in real time, performs noise reduction processing on the image data, and transmits the image data to the image processing module through optical fiber communication;
the two-dimensional servo execution control response module generates a control instruction of the two-dimensional servo module based on the control information output by the image processing module and sends the control instruction to the two-dimensional servo module; the two-dimensional servo module comprises an upper motor and a lower motor which have stepping functions, the high-speed image acquisition device is fixed on the upper motor, the upper motor is arranged on the lower motor, so that the upper motor realizes the up-and-down rotation of the high-speed image acquisition device, the lower motor realizes the left-and-right rotation, and the whole two-dimensional servo module can realize the three-dimensional rotation;
the image processing module integrates pixel level, block level and distributed multi-level heterogeneous parallel processing units and has the processing functions of image feature detection, feature extraction, feature classification and deep learning convolution algorithm; the image processing module completes a pixel level parallel algorithm through a pixel processing unit PE array of a pixel level processor, then accesses an image block in the pixel processing unit PE array of the pixel level processor through a block processing unit PPU in the block level parallel processor to complete a local feature extraction algorithm, and finally utilizes distributed computation to realize rapid classification of extracted local features and generate the control information, thereby realizing parallel processing based on multilevel isomerism; the image processing module comprises a processor, an instruction set of the processor is a special visual processing instruction set based on a multi-level heterogeneous parallel processing architecture, programming of pixel level, block level and distributed parallel operation is supported, and a block processing unit circuit, a processing unit circuit and a cache control circuit are designed to realize the function of the instruction set; the pixel-level parallel processing and the distributed parallel processing are performed by time-division multiplexing using the same hardware.
2. The high-speed tracking system of claim 1, wherein the high-speed image capture device is a high-speed camera, the pixel devices of the high-speed camera are high-speed CMOS image sensors, and the picture readout rate is greater than or equal to 1000 fps.
3. The high-speed tracking system of claim 1, wherein the two-dimensional servo execution control response module connects the image data processing module and the two-dimensional servo module by a bus.
4. The high-speed tracking system of claim 1, wherein the image processing module further comprises a system bus, a sensor control module, a sensor cache, a program cache, and a network interface, wherein the sensor cache is configured to store data from a high-speed image capture device; the program cache stores an algorithm program for visual processing; the network interface is used for communicating with a PC.
5. The high-speed tracking system of claim 1, wherein the pixel-level parallel processing includes filtering, morphology, background subtraction, feature point detection, and feature generation.
6. The high-speed tracking system of claim 1, further comprising a display module connected to the image processing module via an HDMI/DP interface;
and the display module displays the target and the background information according to the information obtained after the image processing module operates the image, and marks the target in real time.
7. The high-speed tracking system of claim 1, further comprising a power distribution module that distributes power to the system to ensure stable operation of the system.
8. The high-speed tracking system according to claim 1, wherein the high-speed image capturing device is capable of real-time communication with the upper computer via USB3.0, and the processor in the image processing module is capable of real-time communication with the upper computer via PCIe protocol and/or internet protocol.
9. A high-speed tracing method based on multi-level heterogeneous parallel processing, which is implemented by using the high-speed tracing system of any one of claims 1 to 8, wherein the high-speed tracing method comprises the following steps:
when a tracking target appears in the visual field of the high-speed image acquisition device with the autonomous focusing function, the image information acquisition and transmission module receives image data from the high-speed image acquisition device, and transmits the image data to the image processing module after image denoising processing;
the image processing module performs pixel-level, block-level and distributed parallel processing on the image data to obtain position information of the tracking target, transmits the position information to the two-dimensional servo execution response control module, judges the position information at the same time to obtain a focusing coefficient and transmits the focusing coefficient to the high-speed image acquisition device with the autonomous focusing function, and completes autonomous focusing of the high-speed image acquisition device;
the two-dimensional servo execution response control module calculates the position information of the tracking target to obtain direction information and rotation angle information; the two-dimensional servo module receives the direction information and the rotation angle information from the two-dimensional servo execution response control module and then correspondingly rotates;
meanwhile, the image processing module displays the tracking target information on a display module.
10. The high-speed tracking method according to claim 9, further comprising the steps of:
when the tracking target is displayed on the display module, judging whether the tracking target is lost or not at regular intervals, if so, matching the two-dimensional servo execution control response module with the two-dimensional servo module to execute tracking action to search the track of the tracking target; if not, the two-dimensional servo execution control response module and the two-dimensional servo module are matched to execute locking action, and the tracking target is locked in the visual field of the high-speed image acquisition device with the self-focusing function.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711498701.2A CN108184062B (en) | 2017-12-29 | 2017-12-29 | High-speed tracking system and method based on multi-level heterogeneous parallel processing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711498701.2A CN108184062B (en) | 2017-12-29 | 2017-12-29 | High-speed tracking system and method based on multi-level heterogeneous parallel processing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108184062A CN108184062A (en) | 2018-06-19 |
CN108184062B true CN108184062B (en) | 2020-08-25 |
Family
ID=62549629
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711498701.2A Active CN108184062B (en) | 2017-12-29 | 2017-12-29 | High-speed tracking system and method based on multi-level heterogeneous parallel processing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108184062B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109741237B (en) * | 2018-12-28 | 2020-10-23 | 中国科学院半导体研究所 | Large-scale image data processing system and method |
CN111932589A (en) * | 2020-08-11 | 2020-11-13 | 沈阳帝信人工智能产业研究院有限公司 | Data processing method, system and equipment |
CN114841848A (en) * | 2022-04-19 | 2022-08-02 | 珠海欧比特宇航科技股份有限公司 | High bandwidth signal processing system, apparatus, method and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101814137A (en) * | 2010-03-25 | 2010-08-25 | 浙江工业大学 | Driver fatigue monitor system based on infrared eye state identification |
CN103020890A (en) * | 2012-12-17 | 2013-04-03 | 中国科学院半导体研究所 | Visual processing device based on multi-layer parallel processing |
CN103544714A (en) * | 2013-10-22 | 2014-01-29 | 中国科学院半导体研究所 | Visual tracking system and method based on high-speed image sensor |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7054491B2 (en) * | 2001-11-16 | 2006-05-30 | Stmicroelectronics, Inc. | Scalable architecture for corresponding multiple video streams at frame rate |
-
2017
- 2017-12-29 CN CN201711498701.2A patent/CN108184062B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101814137A (en) * | 2010-03-25 | 2010-08-25 | 浙江工业大学 | Driver fatigue monitor system based on infrared eye state identification |
CN103020890A (en) * | 2012-12-17 | 2013-04-03 | 中国科学院半导体研究所 | Visual processing device based on multi-layer parallel processing |
CN103544714A (en) * | 2013-10-22 | 2014-01-29 | 中国科学院半导体研究所 | Visual tracking system and method based on high-speed image sensor |
Also Published As
Publication number | Publication date |
---|---|
CN108184062A (en) | 2018-06-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7106665B2 (en) | MONOCULAR DEPTH ESTIMATION METHOD AND DEVICE, DEVICE AND STORAGE MEDIUM THEREOF | |
US11862021B2 (en) | Method and apparatus for remanent imaging control | |
EP3631756B1 (en) | Block-matching optical flow and stereo vision for dynamic vision sensors | |
CN107659774B (en) | Video imaging system and video processing method based on multi-scale camera array | |
US8848035B2 (en) | Device for generating three dimensional surface models of moving objects | |
CN105592269B (en) | Moving target captures and tracking equipment and moving target capture and tracking | |
CN108184062B (en) | High-speed tracking system and method based on multi-level heterogeneous parallel processing | |
CN112785628B (en) | Track prediction method and system based on panoramic view angle detection tracking | |
CN109934108B (en) | Multi-target and multi-type vehicle detection and distance measurement system and implementation method | |
CN107613262B (en) | Visual information processing system and method | |
CN111104960B (en) | Sign language identification method based on millimeter wave radar and machine vision | |
CN105791686B (en) | Multiple mobile object captures and tracking equipment and multiple mobile object capture and tracking | |
CN109089048B (en) | Multi-lens panoramic linkage device and method | |
CN104184936B (en) | Image focusing processing method and system based on light field camera | |
CN113721631A (en) | Sensor data processing method, system and readable storage medium | |
CN116580169B (en) | Digital man driving method and device, electronic equipment and storage medium | |
CN109121194B (en) | Method and apparatus for state transition of electronic device | |
CN116012609A (en) | Multi-target tracking method, device, electronic equipment and medium for looking around fish eyes | |
CN115598744A (en) | High-dimensional light field event camera based on micro-lens array and extraction method | |
CN205490868U (en) | Target of doing more physical exercises is caught and tracking device | |
CN115984124A (en) | Method and device for de-noising and super-resolution of neuromorphic pulse signals | |
CN112529943A (en) | Object detection method, object detection device and intelligent equipment | |
CN111272271A (en) | Vibration measurement method, system, computer device and storage medium | |
TW202014665A (en) | Position inspection method and computer program product | |
Wu et al. | FlyTracker: Motion tracking and obstacle detection for drones using event cameras |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |