Stereo Camera Disparity Correction via Region Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for stereo cameras face challenges in accurately calculating disparity, particularly at distant locations, due to reduced distance resolution and increased computation complexity, including the pixel locking phenomenon and high calculation amounts.
Innovation Solution
An image processing apparatus that calculates disparity using a calculation circuit, a determination circuit, and an interpolation circuit to generate an interpolated image, which corrects disparities based on distribution statistics, improving accuracy while reducing computation complexity by finely extracting pixels near the disparity standard and discretely extracting pixels far from it.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sub-pixel calculation is performed on the entire screen to increase distance resolution at distant locations, then distance resolution at distant locations is improved, but computation amount becomes considerably large
Solution Approach 1:
The patent segments the image into multiple regions based on disparity magnitude. Pixels are classified into a first region (small disparity values) and a second region (large disparity values), allowing different processing strategies for each region. This segmentation enables the system to apply computationally intensive sub-pixel calculation only where needed, resolving the contradiction between improving distance resolution and reducing computation amount.
Solution Approach 2:
The patent applies local quality by using different calculation methods for different regions. For the first region with small disparities, sub-pixel calculation is performed to achieve high distance resolution. For the second region with large disparities, integer-unit disparity calculation is used to reduce computation. This localized approach optimizes both measurement precision and productivity by matching the processing method to the specific needs of each pixel region.
2Measurement precision
If disparity is calculated in integer unit followed by sub-pixel calculation on pixels before and after disparity, then distance resolution is improved, but calculation amount becomes considerably large
Solution Approach 1:
The patent segments pixels into two distinct groups based on disparity magnitude thresholds. This segmentation allows the system to apply simple integer-unit calculation to the majority of pixels (second region) while reserving sub-pixel calculation only for the smaller subset of pixels with small disparities (first region), thereby reducing overall calculation complexity while maintaining improved distance resolution where needed.
Solution Approach 2:
The patent applies partial action by performing sub-pixel calculation only on the subset of pixels that require it (first region with small disparities), rather than applying it universally to all pixels. This partial application of the more complex calculation method achieves the necessary distance resolution improvement without the excessive computational cost that would result from applying it to the entire image.
3Measurement precision
If polygonal line fitting or parabola fitting is performed on cost value to estimate sub-pixel disparity, then distance resolution is improved, but pixel locking phenomenon occurs causing biased estimation
Solution Approach 1:
The patent segments pixels into regions based on disparity magnitude, which helps prevent pixel locking by ensuring that fitting operations are performed on appropriate pixel subsets. By separating pixels with small disparities (first region) from those with large disparities (second region), the system can apply fitting methods only where they are most effective, reducing the likelihood of biased estimation while maintaining distance resolution improvement.
4Reliability
If fitting function is improved to reduce pixel locking phenomenon, then estimation bias is reduced, but computation becomes more complex
Solution Approach 1:
The patent uses segmentation to simplify the computation by dividing pixels into manageable regions. This allows the system to apply relatively simple calculation methods to the majority of pixels while using more complex fitting only where necessary, thereby achieving reliable estimation without excessive computation complexity.
Solution Approach 2:
The patent applies local quality by matching the complexity of the calculation method to the specific needs of each pixel region. For regions where pixel locking is more likely to occur, more sophisticated fitting functions can be applied. For other regions, simpler methods suffice, reducing overall computation complexity while maintaining estimation accuracy where it matters most.
Data Source
AI summary
An image processing apparatus includes a calculation circuit configured to calculate a disparity between images, a determination circuit configured to determine a pixel position at which intensive disparity retrieval is to be performed based on distribution statistics of the disparity, an interpolation circuit configured to generate an interpolated image by performing interpolation at pixel positions at which intensive disparity retrieval is to be performed, and an output circuit configured to output a corrected disparity with which the distance to the object is determined. The calculation circuit calculates a first disparity based on the first and second images and generate distribution statistics from the first disparity in the first image, and calculates a second disparity based on the second image and the interpolated image generated from the first image by the interpolation circuit. The output circuit generates the corrected disparity based on differences between the first and second disparities.