Strip Steel 3D Object Detection Based on Monocular Camera and LiDAR

Unmanned operation of factories has become a popular research direction nowadays. In order to realize the automatic sorting, grasping, loading and unloading of scattered steel by excavators, this paper uses the fusion of monocular camera and LiDAR to perceive and identify the scattered steel targets. Using the texture and edge information of the camera image and the spatial geometry of the LIDAR point cloud, the two features are fused to solve the problem of the accuracy and robustness of the detection of scattered strip steel targets. A feature fusion module and a object detection network are proposed to improve the detection accuracy of strip steel targets, which are relatively slender. Evaluated on a self-constructed dataset, our method achieves a good accuracy of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{6 0. 2 8 \%}$</tex>.