Stereo Camera Disparity Calculation With Uneven Deviation Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for calculating disparity from stereo camera images face challenges in regions with low texture, long-distance measurement resolution, and systematic errors in sub-pixel estimation, particularly in regions near image contours.
Innovation Solution
The proposed method calculates disparity using dynamic programming with unevenly spaced deviation amounts and fitting functions to handle low-texture regions and contours, and estimates sub-pixel disparities by aggregating cost values and adjusting for image displacements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If block matching method is used for disparity calculation, then calculation speed is improved, but measurement precision deteriorates in regions with low texture
Solution Approach 1:
The patent divides the image into multiple search areas and performs disparity calculation separately for each area. This segmentation allows the system to apply different search ranges and strategies to different regions, improving precision in low-texture areas without sacrificing overall calculation efficiency.
Solution Approach 2:
The patent implements local quality by adjusting search range and calculation parameters based on local image characteristics. In regions with low texture, the search range is expanded and more thorough searching is performed, while in high-texture regions, standard parameters are used, thus optimizing both precision and speed locally.
2Length of stationary object
If disparity calculation is performed at long distances, then measurement range is improved, but measurement precision deteriorates
Solution Approach 1:
The patent dynamically adjusts the search range based on the estimated distance to the subject. For long-distance objects, the search range is automatically expanded to accommodate larger disparity values, while for close objects, a tighter search range is used. This dynamic adaptation maintains measurement precision across varying distances.
Solution Approach 2:
The patent changes calculation parameters such as search range, block size, and aggregation window based on distance estimates. By adapting these parameters to the specific measurement scenario, the system maintains high precision whether measuring close or distant objects.
3Measurement precision
If sub-pixel estimation is used for disparity calculation, then measurement precision is improved, but systematic errors increase due to pixel locking
Solution Approach 1:
The patent introduces asymmetry by applying different processing to odd and even pixel positions. By calculating disparities for odd-positioned pixels and even-positioned pixels separately and then combining results, the system breaks the periodic pixel locking pattern while maintaining sub-pixel estimation accuracy.
Solution Approach 2:
The patent performs preliminary disparity calculations at integer pixel positions first, then uses these results as a basis for sub-pixel refinement. This two-stage approach allows the system to establish a reliable baseline before applying more complex sub-pixel estimation techniques.
4Measurement precision
If dynamic programming with multiple deviation amounts is used, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent performs aggregation of cost values for multiple deviation amounts even though only the minimum is ultimately needed. This excessive action ensures that all possible disparity candidates are thoroughly evaluated, improving precision by preventing premature convergence on incorrect minima.
Solution Approach 2:
The patent maintains continuous aggregation of cost values across multiple deviation amounts rather than evaluating each independently. This continuous approach allows the dynamic programming to leverage information from neighboring deviation amounts, improving precision while organizing complexity in a systematic manner.
Data Source
AI summary
An image processing device 20 includes a communication unit 21 and a controller 22 (processor). The communication unit 21 acquires a standard image and a reference image captured by a stereo camera 10. The controller 22 calculates a disparity based on the standard image and the reference image. The controller 22 calculates, at each pixel position, a cost value representing a degree of difference for each pixel position between the standard image and each of multiple images obtained by displacing the reference image by multiple different deviation amounts in a direction corresponding to a baseline length direction of the stereo camera 10. The processor aggregates the cost values using dynamic programming and calculates a disparity for each pixel position based on the aggregated cost values for each different deviation amount. The multiple different deviation amounts are discrete values at uneven intervals.