Obstacle Detection Using Dual Motion Compensation and Block Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional monocular-based obstacle detection methods rely solely on motion compensation, which can lead to errors due to low precision in motion parameters and image noise, resulting in incorrect obstacle detection.
Innovation Solution
The method involves partitioning frames into blocks and calculating motion and characteristic obstacle confidences using both absolute and relative motion compensation, along with a characteristic analysis strategy to determine obstacles, reducing error detection and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If motion compensation is used for obstacle detection, then obstacle detection can be performed, but error detection increases due to low precision in motion parameters and image noise
Solution Approach 1:
The patent combines absolute motion compensation (comparing current frame with hypothetical frame) and relative motion compensation (comparing current frame with previous frame) to calculate motion obstacle confidence. This merging of multiple compensation methods resolves the contradiction by compensating for the low precision of individual motion parameters through multiple reference frames, thereby improving detection reliability without requiring higher single-parameter precision
Solution Approach 2:
The patent uses N previous frames to calculate motion obstacle confidence values that serve as feedback for determining whether blocks are obstacles. By incorporating historical frame data and using confidence thresholds, the system creates a feedback mechanism that filters out noise-induced errors while maintaining accurate obstacle detection, thus improving reliability despite limited motion parameter precision
2Reliability
If motion compensation is used for obstacle detection, then obstacle detection can be performed, but false positives increase due to image noise
Solution Approach 1:
The patent merges absolute and relative motion compensation approaches, where absolute compensation (current frame vs. hypothetical frame) and relative compensation (current frame vs. previous frame) are combined to calculate motion obstacle confidence. This dual-approach merging reduces false positives caused by image noise because noise affects both compensation methods differently, and their combination filters out noise-induced errors while preserving true obstacle signals
Solution Approach 2:
The patent performs preliminary motion compensation calculations using N previous frames before making the final obstacle determination. By pre-calculating motion obstacle confidences from multiple historical frames and comparing against thresholds, the system prepares filtered confidence values that reduce the impact of image noise before the final detection decision, thereby reducing false positives
Data Source
AI summary
A method for identifying barriers in images is disclosed. In the method, images of a current frame and N frame which is nearest to the current frame are obtained, the obtained images of the frames are divided in the same way, and the image of each frame obtains a plurality of divided block regions; the motion barrier confidence of each block region corresponding to the current frame and the N frame which is nearest to the current frame is calculated; whether each block region in the image of the current frame is decided successively according to the motion barrier confidence of each block region corresponding to the current frame and the N frame which is nearest to the current frame; the barriers in the images are determined according to each block region.