Vehicle Obstacle Recognition Using Dual-Threshold Feature Motion Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing obstacle recognition systems for autonomous vehicles face challenges in accurately distinguishing between stationary and moving objects when the vehicle is in motion, leading to insufficient feature points for reliable obstacle detection.
Innovation Solution
The proposed obstacle recognition device employs a system with four fish-eye monocular cameras mounted at different positions on the vehicle, using a processing unit to extract feature points, calculate motion distances, and determine whether these distances exceed specific thresholds to identify obstacles, including intermediate objects between stationary and moving ones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle is moving, then the camera can capture images of the surrounding environment, but it becomes difficult to distinguish between stationary and moving objects leading to insufficient feature points for reliable obstacle detection
Solution Approach 1:
The patent segments the motion analysis into two distinct components: vehicle body motion and object-relative motion. By separating these motions and analyzing them independently through dual-camera setup, the system can accurately distinguish between stationary objects (whose feature points show only vehicle-induced motion) and moving objects (whose feature points show additional relative motion), thereby resolving the contradiction between detection reliability and measurement precision.
2Measurement precision
If feature points from stationary objects are excluded from analysis, then moving objects can be detected, but the number of available feature points decreases reducing detection accuracy
Solution Approach 1:
The patent segments feature points into two categories: those belonging to stationary objects and those belonging to moving objects. By using spatial-temporal analysis and dual-camera motion compensation, the system can identify and separate these categories, allowing all feature points to be utilized while maintaining detection accuracy through proper classification.
Solution Approach 2:
The patent applies dynamic motion analysis to differentiate between stationary and moving objects. By continuously tracking feature point motion patterns across multiple frames and comparing them against vehicle motion patterns, the system dynamically identifies which feature points belong to moving objects, enabling accurate detection while maximizing the use of available feature points.
3Reliability
If multiple cameras are used to monitor areas around the vehicle, then obstacle detection coverage is improved, but the number of cameras required increases system complexity
Solution Approach 1:
The patent segments the monitoring task into two functional components handled by different camera pairs: one pair for capturing images and another for capturing reference images. This segmentation allows the system to achieve comprehensive obstacle detection coverage while maintaining a manageable camera configuration, as each camera pair has a specific functional role.
Solution Approach 2:
The patent makes the camera system multi-functional by using the same camera hardware for dual purposes: capturing both regular images and reference images. This universality reduces the total number of cameras needed while maintaining comprehensive monitoring coverage, as each camera contributes to multiple aspects of the obstacle detection process.
Data Source
AI summary
An obstacle recognition device of a vehicle provided with a camera capturing an image around the vehicle, includes an acquiring unit sequentially acquiring the image captured by the camera; a feature point extracting unit extracting a plurality of feature points of an object included in the image; a calculation unit calculating each motion distance of the plurality of feature points between the image previously acquired and the image currently acquired by the acquiring unit; a first determination unit determining whether each motion distance of the feature points is larger than or equal to a first threshold; a second determination unit determining whether each motion distance of the feature points is larger than or equal to a second threshold; and an obstacle recognition unit recognizing an obstacle.