Research On Obstacle Detection Technology Based On Radar And Camera Of Driverless Smart Vehicles

In a sense, unmanned intelligent vehicle is a kind of intelligent autonomous mobile robot,it mainly uses some of the sensors installed to collect the roads around their ownenvironmental information, these information processing and control intelligent vehicle speedand steering, so that the intelligent vehicle can safely and reliably achieve the goal ofunmanned in the road.This topic mainly use vehicle-mounted camera, laser scanning radar and other sensors tocollect real-time environmental information, for good processing and analysis detected thespecific location and shape of the obstacle,we put forward the theoretical framework for theintegration of the two sensors technology to detect obstacles; related calculation model isestablished, and to study how to improve the computational efficiency of the drive controlunit so as to achieve high-speed real-time achieve the purpose of avoiding obstacles to thesuccess of unmanned intelligent vehicles.This thesis mainly includes the following research contents:(1) According to the requirements of unmanned intelligent car for obstacle detection,then identify and build unmanned intelligent vehicle system platform; the selection of thevarious sensors are analyzed and compared, the design of this article obstacle detection radarand in-car camera processing module; and then to achieve real-time data acquisition andprocessing, obstacle target detection and multi-sensor data fusion, and other functions.(2) In order to achieve the detection and identification of obstacles goal, three coordinatesystems are defined in the smart car system: the world coordinate system, the imagecoordinate system, the sensor coordinate system, the world coordinate system is to determinethe relative position of objects in space, therefore, in order to reflect the space obstaclesspecific information to the data collected by the radar and the camera, this article alsoidentified the transformation of relations between the three coordinate systems, laser radar andcamera calibration algorithm are analyzed. It was also suggested herein with respect to theradar and camera acquisition data processing algorithms, and through specific data for thevalidation of the proposed algorithm. (3) Laser radar data collected can provide the specific shape of the obstacle, but can notget the specific shape features of obstacles, over camera images collected to provide thespecific shape of the obstacle, but can not judge the distance. This article first by radar andcameras to collect data, the data preprocessing; with a laser radar data collected by judge theobstacle area, combining with the camera image information acquisition, extraction andidentification sense of xing area location, finally detect obstacles, realizes the funtion ofunmanned intelligent car around obstacles automatic autonomously, to achieve the purpose ofsecurity and stability of driving.