Abstract:During automatic driving, the complex background and mutual occlusion between multiple targets hinder the correct judgment of the detector and miss detection. When a close-range target is captured again, the vehicle may not be able to respond in time and cause a fatal accident. Therefore, in the application of auxiliary systems, a model that can accurately identify partially occluded targets in complex backgrounds and perform short-term tracking and early warning of completely occluded objects is required. This paper proposes a method to improve detection accuracy while supporting real-time operations based on YOLOv3 and realize real-time warnings for those objects that are completely blocked. First, we obtain a more suitable prior frames setting through class-wise K-means clustering. To solve the problem that the maxpool operation of original CBAM easily introduces background noise, we proposed AS-CBAM(Adaptive Selection Convolutional Block Attention Module) and innovatively combined the HDC(Hybrid Dilated Convolution) to maximize the receptive field and fine-tune the characteristics. The 1×1 convolution operation is used to suppress the increase of the parameter amount. In this study, DIOU-NMS was used to replace traditional NMS. Besides, a tracking algorithm based on Kalman filtering and Hungarian matching is introduced to improve the system's ability to recognize occluded objects. Compared with the traditional YOLOv3, the proposed method can increase the mAP by 1.32% and 1.47% on KITTI and UA-DETRAC, respectively. Nevertheless, it shows a processing speed of 35.07FPS and a more significant improvement in accuracy (90.36% vs. 85.71%) on the Object-Mask, a dataset that focuses on occlusion conditions. Therefore, the proposed algorithm is more suitable for autonomous driving applications.

A Part-Based Gaussian Reweighted Approach for Occluded Vehicle Detection

Part-Aware Region Proposal for Vehicle Detection in High Occlusion Environment

Research and Implementation of Anti-occlusion Algorithm for Vehicle Detection in Video Data.

A Multi-view 3D Vehicle Detection Method Based On Novel 3D Proposal Generation Method

Pseudo-Image and Sparse Points: Vehicle Detection with 2D LiDAR Revisited by Deep Learning-Based Methods

An Efficient Wide-Range Pseudo-3D Vehicle Detection Using A Single Camera

Vehicle Detection with a Part-Based Model for Complex Traffic Conditions

Quad Gaussian Networks for Vehicle Detection in Aerial Images

Probabilistic Inference for Occluded and Multiview On-road Vehicle Detection

3D Bounding Box Estimation for Autonomous Vehicles by Cascaded Geometric Constraints and Depurated 2D Detections Using 3D Results

Research on Moving Vehicle Detection in the Presence of Occlusion

On-road Vehicle Detection Through Part Model Learning and Probabilistic Inference

Enhancing YOLO for occluded vehicle detection with grouped orthogonal attention and dense object repulsion

YOLOv3-MT: A YOLOv3 using multi-target tracking for vehicle visual detection

A Multi-View Vehicle Detection Method Based on Deep Neural Networks

3D Detection for Occluded Vehicles From Point Clouds

Image Guidance Based 3D Vehicle Detection in Traffic Scene.

Robust LiDAR-Based Vehicle Detection for On-Road Autonomous Driving

Vehicle Detection Based on Deep Dual-Vehicle Deformable Part Models

3-D Vehicle Detection Enhancement Using Tracking Feedback in Sparse Point Clouds Environments

Beyond Bounding Box: Fine-Grained Vehicle Detection via Single Stage Detector with Hierarchical output