Monocular Camera Obstacle Detection via Feature Point Distance Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional obstacle detection methods using a monocular camera are inefficient and inaccurate, particularly when detecting multiple types of obstacles, as they require multiple classifiers and a tedious training process, leading to low efficiency and accuracy.
Innovation Solution
A method that involves obtaining and processing frame images to identify stable feature points, dividing them into subsets, and judging changes in distance between ground projection points to determine obstacles, thereby avoiding mismatching and the need for multiple classifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple classifiers are used to detect different types of obstacles, then the detection coverage is improved, but the device complexity and training process become tedious and complicated
Solution Approach 1:
The patent employs a universal detection approach using a single camera system that can detect multiple types of obstacles (pedestrians, vehicles, animals, etc.) through unified geometric and motion analysis. Instead of deploying multiple specialized classifiers, the system uses a single detection framework that analyzes spatial relationships, motion patterns, and geometric features to identify various obstacle types, thereby achieving multi-functionality without increasing device complexity
Solution Approach 2:
The system detects different obstacle types by changing and analyzing multiple parameters including spatial coordinates, motion vectors, distance changes, and geometric relationships between feature points. By dynamically adjusting and evaluating these parameters, the system can distinguish between different obstacle types using a single detection mechanism rather than multiple fixed classifiers
2Measurement precision
If multiple classifiers are trained for different obstacle types, then the detection accuracy for specific obstacles is improved, but the training process becomes tedious and complicated
Solution Approach 1:
The system performs preliminary feature extraction and geometric relationship establishment before actual obstacle detection. By pre-defining spatial relationships, motion patterns, and geometric constraints that characterize different obstacle types, the system eliminates the need for time-consuming training processes while maintaining high detection accuracy through rule-based geometric analysis
3Ease of manufacture
If conventional detection methods are used, then the implementation is simple, but the detection efficiency is low
Solution Approach 1:
The patent segments the obstacle detection process into distinct stages: feature point extraction, spatial relationship calculation, motion pattern analysis, and obstacle type classification. This segmentation allows each stage to be optimized independently while maintaining overall system simplicity, thereby improving detection efficiency without significantly increasing implementation complexity
Data Source
AI summary
Method and apparatus for detecting obstacle based on monocular camera are provided. The method includes: obtaining a target frame image and its adjacent frame image shot by the monocular camera; deleting an unstable feature point from an initial feature point set of the adjacent frame image to obtain a preferred feature point set; dividing a target feature point set to obtain several target feature point subsets; judging whether the target feature point subset corresponds to obstacle based on a change of a distance between points within a ground projection point set of the target feature point subset from adjacent frame time instant to target frame time instant; and determining a union of all the target feature point subsets which are judged as corresponding to obstacles as an obstacle point set of the target frame image.