Design And Research Of Front Vehicle Recognition And Distance Measurement System Based On OpenCV

With the idea of “'Internet Plus",more and more enterprises have begun to focus on combining their products with Internet information technology to upgrade existing products or create emerging products.Intelligent connected vehicles are produced by the traditional automobile industry in the era of "Internet Plus".The vehicle recognition and distance measurement system is an indispensable part in the process of the development and application of intelligent connected vehicle.In this paper,a front vehicle recognition and distance measurement system based on OpenCV is designed.Visual Studio 2015 with OpenCV is used as the program development environment.After loading the video image captured by the camera,the video image is preprocessed.Then the trained vehicle classifier is introduced to recognize the vehicle in the video image,and the distance between the driving vehicle and the target vehicle is calculated by algorithm.Then the recognition result and the distance detection result are displayed on the screen.Finally,transplanting the system to Android and communication between the Android client and the cloud server are all realized.Android client can transmit the detected vehicle distance to the cloud server in real time.In this paper,the acquisition of positive and negative samples for training vehicle classifiers is completed.The algorithm for graying and normalization of all samples at one time is designed.The OpenCV machine learning function is used to complete the training of the Adaboost cascade classifier based on Haar feature.And the vehicle distance calculation model is built.The algorithm of vehicle recognition and distance measurement is designed,including opening of the camera,importing and preprocessing video image,importing classifier to identified front vehicle,completing distance measurement and displaying the recognition result and distance on the screen.Then,transplanting the system to Android and communication between the Android client and the cloud server are all realized.Finally,in this paper,real vehicle experiments are carried out in order to verify the function of the system and the system transplanted to Android.The results show that the system and the system transplanted to Android can recognition the front vehicles well and the result of distance calculation are also relatively accurate,and the communication between the Androidclient and the cloud server can be well realized.