A Method for Pedestrian Intrusion Detection of Railway Perimeter Based on HOG and SVM

Abstract

In order to ensure the safety of railway operation, it has become urgent to strengthen the detection and protection of railway perimeter safety. In this paper, a method for pedestrian intrusion detection of railway perimeter based on histograms of oriented gradients (HOG) plus support vector machine (SVM) is proposed. By establishing the perimeter intrusion image sample set of the railway scene, Gaussian filtering is performed on the sample image to obtain the training set of positive and negative samples, and then, the HOG features are extracted for training. Finally, the field image is used to test this algorithm. The results show that the proposed algorithm is better than other algorithms, such as the inter-frame difference method and the mixed Gaussian background modeling method in the detection of pedestrian intrusion detection of railway perimeter. This paper describes in detail the above process and puts forward the deficiencies for further research in the future.