Obstacle And Slope Terrain Detection Based On 3D LADAR And Monocular Camera

The unmanned platform is a kind of intelligent system,which integrates the functions of environment perception,decision making and motion control into a whole.Environment perception system is a key part of unmanned platforms,and obstacle and slope topography detection are important phases of environmental perception.How to obtain reliable and accurate external environment information is the hot and difficult point of this research field.In this paper,the fusion information of 3D laser radar and monocular camera to detect the target information in unknown environment is used.The main contents are as follows:1.In order to solve the problem of data fusion between a monocular camera and a 3D LADAR,a novel simplified scheme is proposed in this paper.In this method,only one target calibration plate is used in the whole calibration process,and the calibration of the calibration plate can be reduced by reducing the number of placement positions to achieve the joint calibration.In order to verify the accuracy of the joint calibration results,the image reprojection method is used to display the calibration data.Experimental results show that the minimum number of positions required to meet the joint calibration accuracy in the proposed method is 5,which is smaller than 12 in traditional methods.Therefore,this method can simplify the calibration process and improves the efficiency of the joint calibration under the premise of ensuring the accuracy.2.An algorithm for detecting the specific size information of obstacles is proposed,which based on the fusion data of the above two sensors.In the first step of the algorithm,the laser point clouds are clustered according to the European clustering method.Secondly,detected image is divided into the corresponding small regions according to the number of clusters.In the final,the image processing technology is used to detect the obstacle areas,the obstacle size information is calculated to complete the obstacle information detection.The experiments in the real world are implemented to show that the proposed approach can effectively detect the size of the obstacles in unknown environment.3.In order to detect the slope topography,the color point cloud is generated based on the fusion of the laser point cloud and the corresponding camera image information.Firstly,the laser point cloud data is filtered,and the HSV method is also used to extract the environmental features of the corresponding image area.Then,the slope of the laser point cloud and the color point cloud are respectively analyzed to determine the accessible area.4.The integrated experiments including obstacles information and slope detection is designed.Based on the experimental platform of the intelligent car,the validity of proposed algorithm is proved.