Obstacle Detection Credibility Evaluation via Dual Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advanced Driver Assistance Systems (ADAS) face errors in obstacle detection due to interference from headlights, dirt, and backlit objects, leading to misleading results in contour tracing and texture analysis.
Innovation Solution
A method for evaluating the credibility of obstacle detection using a processor connected to an image capturing device, which performs dual classifications and calculates a penalty score based on distinct obstacles detected in images, updating a credibility score to account for inconsistencies between classification results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If single classification is used for obstacle detection, then detection speed is maintained, but detection reliability deteriorates due to interference from headlights, dirt, and backlit objects
Solution Approach 1:
The obstacle detection system is segmented into two independent classification paths: a first classification path that performs comprehensive detection, and a second classification path that performs verification detection. Each path processes images independently and produces separate detection results, which are then compared to determine final obstacle presence. This segmentation allows the system to maintain high reliability through cross-validation while managing complexity through modular design.
Solution Approach 2:
The system implements a feedback mechanism where the second classification result is used to verify and adjust the first classification result. When the second classification detects an obstacle that the first classification missed, or vice versa, the system generates a discrepancy signal that triggers credibility evaluation and potential re-detection. This feedback loop continuously improves detection reliability by learning from classification inconsistencies.
2Measurement precision
If dual classification is performed to improve detection reliability, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The first classification is performed as a preliminary detection step that quickly identifies potential obstacles. Based on the confidence level and characteristics of the first detection result, the system dynamically determines whether a second verification classification is necessary. For high-confidence detections with clear obstacle characteristics, the second classification may be skipped or performed with reduced processing, thereby reducing overall processing time while maintaining accuracy for critical cases.
Solution Approach 2:
The system performs partial dual classification by selectively applying the second classification only when needed based on the first classification results. Rather than always performing both classifications equally, the system applies the second classification partially - only for detections that require verification or when the first classification shows uncertainty. This approach achieves improved accuracy where necessary while minimizing unnecessary processing time for clear, unambiguous detections.
3Reliability
If credibility score calculation is implemented to evaluate detection reliability, then system reliability is improved, but computational complexity increases
Solution Approach 1:
An intermediary credibility evaluation module is introduced between the two classification paths. This module receives detection results from both classifications, compares them, and calculates a credibility score based on the degree of agreement or discrepancy. The intermediary acts as a mediator that translates complex classification comparisons into a simple, interpretable credibility metric that can be used for decision-making, thereby managing complexity through abstraction.
Solution Approach 2:
The system changes the parameter representation of detection reliability from multiple detailed comparison metrics to a single aggregated credibility score parameter. By transforming the complex multi-dimensional comparison data into a unified credibility parameter, the system simplifies subsequent processing and decision-making while preserving the essential reliability information. This parameter transformation reduces computational complexity in downstream applications.
Data Source
AI summary
A method is to be implemented by a processor connected to a storage device and an image capturing device, and includes steps of: performing a first and a second classifications on an image received from the image capturing device for respectively obtaining first and second results of obstacle detection, and storing the same in the storage device; determining whether there is a distinct obstacle, which is an obstacle indicated by one of the first and second results of obstacle detection, and not indicated by the other one of the first and second results of obstacle detection; and calculating a penalty score based on the distinct obstacle, and updating a credibility score by subtracting the penalty score therefrom.