Environment perception is one of the key technologies of the intelligent vehicle. However, there are limitations in object detection and object positioning by visual sensor or lidar only. Based on the image and lidar object detection, a fusion algorithm of vehicle object shape-position using stereo vision and lidar is proposed in this paper. Firstly, the deep learning methods are used for object detection on image and point cloud. Then, the shape-position of object is determined by the object shape-position estimation method based on 3D points and object types. Finally, the image object and point cloud object are fused simultaneously after data association and the shape-position of the object are acquired. The proposed algorithm is evaluated on the KITTI data set and actual road scenarios. Experimental results show that the detection accuracy of the proposed method is 5. 72% and 1. 8% higher than those of the YOLOv3 network and the Point-GNN network, respectively. In addition, the average error of object shape and position within 20 m is 4. 34% and 4. 52% , respectively.