Aiming at the problem of low accuracy of cone barrel detection in the perception system of driverless Formula racing in the field environment,a cone-barrel detection algorithm based on calibration fusion was proposed by aligning LiDAR and camera with time and space.Firstly,the point cloud collected by the LiDAR was filtered,ground plane was filtered,and Euclid clustering was performed to obtain the spatial position information of the cone barrel in ROI.Secondly,the spatial position information of the cone barrel was projected into the pixel plane by the internal and external reference matrix of the camera obtained after calibration,and the color judgment was made on the pixels of the H channel in a certain area around the spatial position of the cone barrel under the HSV color space,so as to obtain the color information of the cone barrel.Finally,the spatial position information of the cone bucket was projected into the pixel plane through coordinate transformation and fused with the obtained cone bucket color information.The results show that the calibration fusion algorithm can meet the competition requirements of different events on the field,and avoid cone barrel detection of errors when using a single sensor algorithm.Therefore,the proposed fusion algorithm can effectively detect the cone barrel,enrich the information of the detection target,improve the accuracy of detecting the cone barrel,and provide a reference for the use of sensor fusion to achieve the detection of the cone barrel in the Formula competition.