D.H. Li, J.L. Sun, P. Wu, W. Tao, A. Chen

Pages: 113-122

Abstract
Aiming at the problem of large error and long time of traditional dangerous driving behavior recognition, this paper designs a dangerous driving behavior recognition method based on Apriori algorithm in rush hour. Firstly, the basic coordinate system and the world coordinate system of the camera are established, and the driving behavior images of the driver are collected comprehensively by means of homogeneous matrix transformation. After the grayscale processing of the image, median filtering method is used to suppress the noise in the image, and histogram equalization algorithm is used to enhance the image. Finally, dangerous driving behavior is located by determining the face and hand change characteristics of dangerous driving behavior, and the Apriori algorithm is used to determine the frequent item set of dangerous driving behavior, and the principal component of dangerous driving behavior image is extracted to achieve effective recognition. Experimental results show that the proposed method can keep the recognition error below 2.0%, and its global maximum recognition time is only 0.21s, indicating that the proposed method effectively achieves the design expectation.
Keywords: apriori algorithm; dangerous driving behavior; behavior recognition; histogram equalization; median filtering