Long Su and Zhi Qiao
Sobel edge detection (ED) Hough algorithm, pinhole model, patrol,robots, avoiding obstacles, fireproof
The complexity and danger of the coal mine environment make it challenging for people to thoroughly inspect fire hazards. The research aims to develop a patrol robot for coal mine fire prevention, addressing the limitations of manual inspections. It enhances Sobel edge detection and the Hough algorithm to improve obstacle identification and fire monitoring efficiency. The method expands directional sensitivity to eight directions and optimises noise suppression. It also refines Hough parameters to reduce noise impact and calculates geometric relationships for real-time obstacle distance. In the experiment, the number of pixels of the proposed method reaches 33,647, and the processing speed of the algorithm is only 17.2 ms. When the patrol robot runs in the actual coal mine environment, the recognition accuracy of the patrol robot can reach 92.2%, which is significantly higher than the traditional Sobel algorithm’s 71.6% and Canny algorithm’s 82.3%. The distance measurement error and angle measurement error are both less than 0.05, which is significantly lower than 0.15 of the traditional Sobel algorithm and 0.10 of the Canny algorithm. The research method can improve the ability of boundary recognition and environment understanding, and improve the practical value of patrol robot in coal mine fire protection application.
Important Links:
Go Back
