Detection And Early Warning Of Dangerous Goods Transportation Vehicles Based On Hierarchical Detection Algorithms

In recent years,with the continuous development of the national economy,the demand for dangerous goods in China has also increased year by year.As of 2017,the total amount of dangerous chemicals transported in China has exceeded 300 million tons per year.The dangerous chemicals transported mainly include cyanide,liquid ammonia,liquid chlorine and oil products.Once a traffic accident occurs,it will have a huge impact on the natural environment,road safety and the safety of people's lives and property.In addition,advanced driver assistance systems based on sensor technology and advanced control technology provide a good solution for vehicle driving safety(Advanced Driver Assistance Systems,ADAS).Sensors play an important role in advanced driver assistance systems.Commonly used sensors mainly include cameras,millimeter wave radars,laser radars,etc.,which can be used to obtain information inside and outside the vehicle.This information can help drivers to complete driving tasks more safely.Therefore,this paper summarizes the research status of relevant aspects at home and abroad,builds a framework for vehicle detection and recognition algorithm,and develops a machine vision-based identification and early warning system for dangerous goods vehicles in front.Finally,through the real vehicle road test test system function.The specific research work is as follows:1)By comparing various vehicle identification and detection algorithms,the Haar feature and AdaBoost cascade classifier algorithm are used to identify dangerous goods transport vehicles.A total of four classifiers were obtained,and the number of positive samples of each classifier was 800,1200,1600 and 2000,respectively.By comparison analysis,the classifier trained by 1600 positive samples is the best.2)Hierarchical detection algorithm is used to identify dangerous goods transport vehicles.Because the existing sliding window detection algorithm is too large and the real-time performance is poor,the source code of the sliding window detection algorithm is interpreted,and under the premise of accurately mastering the working principle,a layered detection idea is proposed.Then modify and rewrite the code related to the sliding window detection algorithm in OpenCV according to this layering idea.The resulting stratified detection algorithm is more suitable for the detection of dangerous goods transport vehicles,which not only improves the speed of detection,but also solves problems such as flashing and flashing to some extent.3)The front vehicle ranging function is completed by using the geometric ranging model.In this paper,the monocular vision ranging method based on geometric model is adopted.Firstly,Zhang Zhengyou calibration algorithm is used to complete the calibration of the camera and obtain the parameters in the camera.Then,according to the geometric ranging model and the mutual conversion relationship between the four coordinates,the pixel coordinates of the selected two ranging points are converted into road plane coordinates.Finally,the road plane coordinates of the two ranging points are substituted into the distance calculation formula to obtain the vehicle distance.Implemented in the Visual Studio 2013 development environment via C++.4)Complete the development of the early warning system and conduct the actual road test.Based on the Visual Studio 2013 development environment and OpenCV library,the system is developed using C++ language,and the system is finally transplanted into the Arm development board through cross-compilation.The system function is tested and analyzed through real vehicle test.