Design Of Video Processing System On Railway Handcarts And Achievement Of The Key Technologies

In recent years, the rapid development of China's railway business, driving a substantial increase in density and speed of traffic, and then make the requirements on the railway safety and reliability requirements more sophisticated. However, the current inspection of the railway situation is also heavily dependent on railway workers' visual check, which resulted in inefficient, heavy labor work, and low reliability, and this is clearly unable to meet the immediate needs of the development of the railway business.In this paper, to the above-mentioned problems, the railway handcarts automatic detection system is proposed. This paper has shown the working principle and methods of the railway handcart detection system, and as well as the methods of the roadbed state and missing fastener detection. All the methods and algorithms that were proposed have been achieved or experimentalized. The pretreatment methods have also been introduced detailedly in this paper.In this paper, a roadbed and fastener location algorithm which is based on apriority of texture analysis is proposed; this algorithm allows the detection system to locate the concerned areas by measuring both horizontal and vertical angles. Besides this, this paper also proposes a roadbed state detection analytical method which is based on gray co-occurrence matrix texture analysis; this method picks up eigenvalues via the gray co-occurrence matrix of the roadbed image, then puts the eigenvalues into the BP neural network to train the classify machine. For detecting the fastener status, we used the PCA algorithm to pick up eigenvalues.