A debugging device capable of accurately adjusting and correcting through infrared operation for a binocular camera
A binocular camera and infrared technology, which is applied in the field of ADAS, can solve the problems of wasting the time of operators and installation personnel, the external structure design is not ideal, and the installation and adjustment are not convenient enough.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
example 1
[0029] After the system is started, the initialization unit of the FPGA will initially set the camera, and the adjustment and correction parameter table will be transmitted from the external PC to the internal SRAM. Afterwards, the pictures captured simultaneously by the stereo camera will be input to the IF interface module through the stereo camera, and after initial processing and synchronization of the left and right pictures, it will be stored in the internal SRAM memory of the adjustment and correction module of the left and right cameras. The images read by the internal SRAM are used in the left and right camera adjustment module to perform tilt correction and camera position correction by referring to the initial parameter table stored in other internal SRAMs, and store them in the internal SRASM of the stereo matching module. Afterwards, the picture data can be read out from the internal stereo matching unit and the weak mode elimination unit at any time, and carry out...
example 2
[0035] Since the matching is carried out in two-dimensional mode, in order to be able to process correctly, feature regions with edge enhancement are required in both directions. First, use formula 1 to obtain the brightness gradient on the pixels connected to the difference filter The strength and direction of:
[0036]
[0037]
[0038] The same process is performed on the feature area of 4×4 pixels, and the histogram of each direction is obtained. Quantization is performed in 8 directions, and if the intensity does not exceed the threshold, no voting is performed. If there are no two directions of 180 degrees and its frequency exceeds the threshold, its characteristic area is regarded as a singularity;
[0039] The two-dimensional matching process adopts the regional reference method. The size of the search area is 512×8 pixels, and the window size for detecting singularities is 4×4 pixels. In addition, SADSum of Absolute Difference is used in the evaluation func...
example 3
[0044] The various processing methods described above realize small scale and high speed by parallelization and pipeline in units of 4 lines. And it can be realized under the premise of only relying on the internal memory SRAM of the FPGA without using an external memory. Table 1 shows the FPGA resources of the stereo camera system including this real-time auto-correction:
[0045]
[0046]
[0047] In order to investigate whether the processing is appropriate, the comparison image is rotated by 0.97mrad, so that it is offset by 1.5 pixels in the vertical direction for testing. The center of rotation is the image center. After that, automatic correction processing is performed. After 1 minute from the start of processing, the slope can be found to be 0.73mrad, and the vertical offset is 1.5 pixels, and then trimmed;
[0048] In addition, for one automatic correction process, when the threshold of the number of singular points detected in one large module is set to 50%...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com