Camera Obstacle Detection by Road Surface Contour Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing obstacle detection methods in mobile robots and blind guide systems suffer from limited detection accuracy, particularly failing to detect small obstacles and requiring manual specification of areas of interest, which limits their effectiveness in complex environments.
Innovation Solution
An obstacle detection method that divides images into road surface and non-road surface areas based on pixel information, identifying non-road surface areas as obstacles, thereby improving detection accuracy without the need for manual area specification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If non-visual detection mode using infrared rays and ultrasonic waves is used, then detection coverage is improved, but detection accuracy deteriorates and small obstacles cannot be detected
Solution Approach 1:
The patent replaces non-visual detection methods (infrared rays and ultrasonic waves) with visual detection using a camera. This substitution allows the system to capture optical images for obstacle detection, thereby improving detection accuracy while maintaining comprehensive coverage through image processing algorithms.
Solution Approach 2:
The patent transitions from non-visual detection dimensions (infrared, ultrasonic) to visual detection dimension (optical imaging). By capturing images in the visible spectrum and processing them through computer vision algorithms, the system achieves both wide coverage and high accuracy in detecting obstacles of various sizes.
2Device complexity
If single camera non-stereoscopic vision detection mode is used, then device complexity is reduced, but automation deteriorates as manual area specification is required
Solution Approach 1:
The patent enables the single camera system to automatically detect and identify road surfaces and obstacles through image processing algorithms. The system performs self-service by automatically segmenting the image into road surface areas and non-road surface areas, eliminating the need for manual area specification while maintaining device simplicity.
Solution Approach 2:
The patent changes the detection parameters by using pixel brightness values and color information from the single camera image. By analyzing these parameters through processing algorithms, the system automatically distinguishes road surfaces from obstacles, achieving automation without increasing device complexity.
3Measurement precision
If manual area of interest specification is required, then detection precision in specified areas is improved, but ease of operation deteriorates and adaptability to complex environments worsens
Solution Approach 1:
The patent makes the system self-service by automatically identifying and analyzing the entire image area without requiring manual specification of areas of interest. The processing algorithms automatically distinguish road surfaces from obstacles based on pixel characteristics, making the system easy to operate and adaptable to complex environments.
Solution Approach 2:
The patent enhances the system's universality by enabling it to automatically adapt to various complex environments without manual intervention. The image processing algorithms can handle diverse scenarios including different road surfaces, lighting conditions, and obstacle types, making the system versatile and easy to deploy across different applications.
Data Source
AI summary
An obstacle detection method and apparatus are provided. The obstacle detection method provided includes: obtaining a to-be-detected image; determining a road surface area and a non-road surface area in the to-be-detected image according to pixel information contained in the to-be-detected image; respectively determining an outermost layer contour line of the road surface area and a contour line of the non-road surface area; and when the contour line of at least one non-road surface area is located in the area contained in the outermost layer contour line of the road surface area, determining a physical object contained in the at least one non-road surface area as an obstacle. The present application is applied to a process of detecting an obstacle.