Brake discs are one of the key components, and their surface defects directly affect the braking performance and safety performance of vehicles. Traditional defect detection methods have poor robustness due to their reliance on manual feature extraction. A detection instrument was designed to focus on the detection of surface defects on brake discs. The defect features of brake discs were extracted using the improved Gaussian difference algorithm and Hough transform algorithm(IGD-IHT). An identification method for brake disc surface defects was designed in this paper based on the Perception-based Image Quality Evaluator and Dempster rule-improved Bayes particle swarm optimization-Nonlinear echo state network(PIQEDS-IBPSO-NESN) to better identify defects. The experimental results showed that the accuracy was more than 97%, the false alarm rate was less than 1.5%, and the false alarm rate was less than 1.5%. The algorithm proposed in this paper achieved the best detection performance.
You may also start an advanced similarity search for this article.