Object Classification via Edge Analysis in Stereo Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing collision avoidance systems cannot differentiate detected objects, such as automobiles and pedestrians, which limits their ability to deploy appropriate safety measures in case of an impending impact.
Innovation Solution
A method and apparatus that classify objects in an image by detecting edges within a region of interest, using edge analysis and edge scores to distinguish between different types of objects, such as cars and pedestrians, through a stereo vision system with cameras and an image processor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a sensor system is used to detect objects in front of a vehicle, then object detection capability is improved, but object classification capability deteriorates
Solution Approach 1:
The patent divides the detected object into multiple sub-regions and analyzes edges in each sub-region separately. By segmenting the object detection process into multiple analysis stages (edge detection, sub-region analysis, score generation), the system maintains detection precision while extracting additional classification information from the segmented data.
Solution Approach 2:
The patent transitions from binary object detection to multi-dimensional classification by introducing edge scores as a new dimension of analysis. The system analyzes edge characteristics across multiple sub-regions and generates composite edge scores that enable classification into different object types (pedestrians, vehicles, bicycles), thereby adding informational dimensions without sacrificing detection accuracy.
2Measurement precision
If edge analysis is performed on multiple sub-regions, then object classification accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the region of interest into multiple sub-regions and performs edge analysis on each segment independently. This segmentation allows the system to manage computational complexity by processing smaller, manageable portions of the image rather than analyzing the entire region as a single complex unit, while still achieving high classification accuracy through aggregated results.
Solution Approach 2:
The patent applies edge detection and analysis to multiple sub-regions (excessive action) to ensure comprehensive classification coverage. By analyzing edges in all sub-regions and generating edge scores for each, the system over-samples the analysis process to guarantee accurate classification, accepting the increased computational effort as necessary for reliable object type differentiation.
Data Source
AI summary
A method and apparatus for classifying an object in an image is disclosed. Edges of an object are detected within a region of interest. Edge analysis is performed on a plurality of sub-regions within the region of interest to generate an edge score. The object is classified based on the edge score.