In-vehicle Stereo Camera Abnormal Region Impact Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing in-vehicle stereo cameras fail to maintain autonomous driving functionality when obstructed by blocking matters or when the camera fails, as they cannot effectively continue sensing functions necessary for traveling control.
Innovation Solution
An in-vehicle stereo camera system that acquires paired images with overlapping regions, performs object recognition using both stereoscopic and monocular vision, and adjusts processing based on the degree of impact from any abnormalities detected in the images to ensure continuous object recognition and autonomous driving.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the stereo camera uses both imaging units for object recognition, then the accuracy of object recognition is improved, but when one imaging unit has an abnormal region, the system cannot maintain reliable sensing functions
Solution Approach 1:
The patent dynamically switches the object recognition processing mode based on the degree of impact of abnormal regions. When abnormal regions are detected, the system transitions from using both imaging units to using only the normal imaging unit, and adjusts the processing content accordingly. This dynamic adaptation maintains reliability while preserving object recognition functionality under degraded conditions.
Solution Approach 2:
The system changes the processing content parameter of object recognition based on the diagnosed degree of impact. When abnormal regions are present, the processing content is adjusted to account for the reduced input quality, allowing the system to maintain accurate object recognition despite the degradation in imaging conditions.
2Duration of action of stationary object
If the system continues autonomous driving with blocked or failed cameras, then the continuity of autonomous driving is improved, but the safety of traveling control may be compromised without proper impact assessment
Solution Approach 1:
The degree-of-impact diagnosis unit provides continuous feedback about the state of imaging units to the object recognition unit. This feedback loop allows the system to monitor the quality of input images and adjust processing accordingly, ensuring that autonomous driving continues safely even when imaging conditions deteriorate, by maintaining appropriate safety margins through informed decision-making.
Solution Approach 2:
The system prepares for potential imaging failures by having pre-defined processing content adjustments ready to apply when abnormal regions are detected. This beforehand cushioning ensures that when imaging quality degrades, the system can immediately adapt its object recognition processing to maintain safe autonomous driving operations without interruption.
3Device complexity
If the object recognition unit processes images with abnormal regions using fixed processing content, then the system complexity is reduced, but the object recognition accuracy deteriorates
Solution Approach 1:
The processing content is made dynamic rather than fixed, allowing the object recognition unit to adapt its processing strategy based on the diagnosed degree of impact from abnormal regions. This dynamic approach maintains relatively simple system architecture while achieving accurate object recognition under varying imaging conditions through flexible processing adjustments.
Data Source
AI summary
Provided is an in-vehicle stereo camera which enables the continuation of automated driving if failures occur during imaging. A pair of captured images 301, 302, which is captured by a pair of imaging units so that both contain an overlapping region, are acquired (S201), it is determined whether an abnormal region is present in at least one of the pair of captured images, and if an abnormal region is present, the degree of impact from the abnormal region on an object recognition process of an object recognition unit is diagnosed (S202), and the processing content of the object recognition process is updated according to the degree of impact (S203).