MSC-YOLO: Improved YOLOv7 Based on Multi-Scale Spatial Context for Small Object Detection in UAV-View

Accurately identifying small objects in high-resolution aerial images presents a complex and crucial task in the field of small object detection on unmanned aerial vehicles (UAVs). This task is challenging due to variations in UAV flight altitude, differences in object scales, as well as factors lik...

Full description

Saved in:
Bibliographic Details
Published in:Computers, materials & continua Vol. 79; no. 1; pp. 983 - 1003
Main Authors: Tang, Xiangyan, Ruan, Chengchun, Li, Xiulai, Li, Binbin, Fu, Cebin
Format: Journal Article
Language:English
Published: Henderson Tech Science Press 2024
Subjects:
Aerial photography
Airspeed
Algorithms
Context
Drone aircraft
Flight altitude
Image resolution
Modules
Object recognition
Target detection
Unmanned aerial vehicles
ISSN:1546-2226, 1546-2218, 1546-2226
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Accurately identifying small objects in high-resolution aerial images presents a complex and crucial task in the field of small object detection on unmanned aerial vehicles (UAVs). This task is challenging due to variations in UAV flight altitude, differences in object scales, as well as factors like flight speed and motion blur. To enhance the detection efficacy of small targets in drone aerial imagery, we propose an enhanced You Only Look Once version 7 (YOLOv7) algorithm based on multi-scale spatial context. We build the MSC-YOLO model, which incorporates an additional prediction head, denoted as P2, to improve adaptability for small objects. We replace conventional downsampling with a Spatial-to-Depth Convolutional Combination (CSPDC) module to mitigate the loss of intricate feature details related to small objects. Furthermore, we propose a Spatial Context Pyramid with Multi-Scale Attention (SCPMA) module, which captures spatial and channel-dependent features of small targets across multiple scales. This module enhances the perception of spatial contextual features and the utilization of multiscale feature information. On the Visdrone2023 and UAVDT datasets, MSC-YOLO achieves remarkable results, outperforming the baseline method YOLOv7 by 3.0% in terms of mean average precision (mAP). The MSC-YOLO algorithm proposed in this paper has demonstrated satisfactory performance in detecting small targets in UAV aerial photography, providing strong support for practical applications.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1546-2226
1546-2218
1546-2226
DOI:10.32604/cmc.2024.047541