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.

YOLOv3 Tiny Vehicle: a New Model for Real-Time Vehicle Detection

Real-Time Object Detection Based on YOLO-v2 for Tiny Vehicle Object

A lightweight Tiny-YOLOv3 vehicle detection approach

LittleYOLO-SPP: A Delicate Real-Time Vehicle Detection Algorithm

Fast vehicle detection algorithm based on lightweight YOLO7-tiny

Improved Vision-Based Vehicle Detection and Classification by Optimized YOLOv4

Embedded YOLO: A Real-Time Object Detector for Small Intelligent Trajectory Cars

Vehicle target detection method based on improved YOLO V3 network model

Real-time Detection of Vehicle and Traffic Light for Intelligent and Connected Vehicles Based on YOLOv3 Network

A Vehicle Detection Method Based on an Improved U-YOLO Network for High-Resolution Remote-Sensing Images

Neighborhood physical disorder in New York City

An Improved YOLO for Road and Vehicle Target Detection Model

M-YOLO: A Nighttime Vehicle Detection Method Combining Mobilenet v2 and YOLO v3

A lightweight vehicles detection network model based on YOLOv5

Vehicle and Pedestrian Detection Based on Improved YOLOv7-Tiny

Light-YOLOv3: License Plate Detection in Multi-Vehicle Scenario.

YOLOv4-dense: A smaller and faster YOLOv4 for real-time edge-device based object detection in traffic scene

DV3-IBi_YOLOv5s: A Lightweight Backbone Network and Multiscale Neck Network Vehicle Detection Algorithm

Vehicle–Pedestrian Detection Method Based on Improved YOLOv8

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

Real-Time Vehicle Detection Based on Improved YOLO v5