Research On Vehicle Recognition And Monitoring Algorithm Based On Computer Vision

In recent years,with the rapid development of deep learning,the research and development of intelligent vehicles has gained extensive attention.For this reason,this paper takes object detection as the basis and studies the application of computer vision technology in vehicle recognition and monitoring.The specific contents are as follows:Firstly,this paper studies the object detection method based on based on deep learning,analyzes the structure and principle of convolutional neural networks and mostly used object detection models.Also proposes an improved vehicle detection algorithm combined with traditional methods.Then this paper studies camera model which is the principle of distance measurement based on vision.Typical methods of monocular ranging are introduced,such as geometry model,inverse perspective transform and mathematical regression.A method of distance measurement based on object detection and perspective transform is proposed.Then this paper implements a method of vehicle recognition and monitoring based on computer vision combined with the actual driving condition.This method includes two main functions which are object detection and data measurement.For lane line detection traditional detection method is used.Road objects are detected by deep learning model.The method by using perspective transform to calculate distance is used in data measurement.Finally the test results are basically in line with expectations.So the algorithm has certain application value.