Stereo Matching Disparity Merging for Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for stereo matching often result in incorrect disparity values due to horizontal streak noise and high processing burdens, making it difficult to accurately detect object distances.
Innovation Solution
An image processing apparatus and method that includes a stereo matching unit for obtaining disparity images from paired camera images, a filter processing unit for filtering these images, and a merging unit that compares and merges disparity values based on filtered results, using techniques like median filtering and reliability calculations to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If simple algorithm is used in calculating disparity for high speed stereo matching, then processing speed is improved, but many wrong disparity values are calculated
Solution Approach 1:
The patent performs filter processing on disparity images before merging right and left disparity values. This preliminary filtering action removes anomaly values from the disparity data, ensuring that when right and left disparities are compared and merged, the comparison is based on reliable values only, thus preventing wrong disparity propagation while maintaining processing efficiency
Solution Approach 2:
The patent introduces filter processing as an intermediary step between stereo matching and disparity merging. This intermediary process cleans the disparity data by removing anomaly values, allowing the subsequent merging operation to reliably combine right and left disparity information without being corrupted by erroneous values
2Reliability
If constraint for smooth change of disparity is used, then disparity calculation reliability is improved, but horizontal streak noise occurs and processing amount increases
Solution Approach 1:
The patent extracts and removes anomaly values from disparity images using filter processing before performing the merging operation. By taking out the harmful anomaly values that would otherwise cause horizontal streak noise, the method achieves reliable disparity calculation without requiring complex smoothness constraints or increasing processing complexity
3Productivity
If right and left disparity values are compared and merged without filter processing, then processing speed is maintained, but wrong disparity values propagate and reduce merging reliability
Solution Approach 1:
The patent performs filter processing as a preliminary step before the disparity merging operation. This preliminary filtering removes anomaly values from the disparity images, ensuring that when right and left disparities are subsequently compared and merged, the operation works with clean data, maintaining both processing efficiency and high merging reliability
Data Source
AI summary
There is provided an image processing apparatus including a stereo matching unit configured to obtain right and left disparity images by using stereo matching, based on a pair of images captured by right and left cameras, respectively, a filter processing unit configured to perform filter processing on the disparity images, and a first merging unit configured to make a comparison, in the disparity images that have undergone the filter processing, between disparity values at mutually corresponding positions in the right and left disparity images and to merge the disparity values of the right and left disparity images based on a comparison result.