SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images

With special consideration for complex scenes and densely distributed small objects, this frequently leads to serious false and missed detections for unmanned aerial vehicle (UAV) images in small object detection scenarios. Consequently, we propose a UAV image small object detection algorithm, terme...

Full description

Saved in:
Bibliographic Details
Main Authors: Shenming Qu, Chaoxu Dang, Wangyou Chen, Yanhong Liu
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/14/2421
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849419040838123520
author Shenming Qu
Chaoxu Dang
Wangyou Chen
Yanhong Liu
author_facet Shenming Qu
Chaoxu Dang
Wangyou Chen
Yanhong Liu
author_sort Shenming Qu
collection DOAJ
description With special consideration for complex scenes and densely distributed small objects, this frequently leads to serious false and missed detections for unmanned aerial vehicle (UAV) images in small object detection scenarios. Consequently, we propose a UAV image small object detection algorithm, termed SMA-YOLO. Firstly, a parameter-free simple slicing convolution (SSC) module is integrated in the backbone network to slice the feature maps and enhance the features so as to effectively retain the features of small objects. Subsequently, to enhance the information exchange between upper and lower layers, we design a special multi-cross-scale feature pyramid network (M-FPN). The C2f-Hierarchical-Phantom Convolution (C2f-HPC) module in the network effectively reduces information loss by fine-grained multi-scale feature fusion. Ultimately, adaptive spatial feature fusion detection Head (ASFFDHead) introduces an additional P2 detection head to enhance the resolution of feature maps to better locate small objects. Moreover, the ASFF mechanism is employed to optimize the detection process by filtering out information conflicts during multi-scale feature fusion, thereby significantly optimizing small object detection capability. Using YOLOv8n as the baseline, SMA-YOLO is evaluated on the VisDrone2019 dataset, achieving a 7.4% improvement in mAP@0.5 and a 13.3% reduction in model parameters, and we also verified its generalization ability on VAUDT and RSOD datasets, which demonstrates the effectiveness of our approach.
format Article
id doaj-art-6d794a4e74e54657a8baf53d49df4856
institution Kabale University
issn 2072-4292
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-6d794a4e74e54657a8baf53d49df48562025-08-20T03:32:15ZengMDPI AGRemote Sensing2072-42922025-07-011714242110.3390/rs17142421SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV ImagesShenming Qu0Chaoxu Dang1Wangyou Chen2Yanhong Liu3School of Software, Henan University, Kaifeng 475004, ChinaSchool of Software, Henan University, Kaifeng 475004, ChinaSchool of Software, Henan University, Kaifeng 475004, ChinaSchool of Software, Henan University, Kaifeng 475004, ChinaWith special consideration for complex scenes and densely distributed small objects, this frequently leads to serious false and missed detections for unmanned aerial vehicle (UAV) images in small object detection scenarios. Consequently, we propose a UAV image small object detection algorithm, termed SMA-YOLO. Firstly, a parameter-free simple slicing convolution (SSC) module is integrated in the backbone network to slice the feature maps and enhance the features so as to effectively retain the features of small objects. Subsequently, to enhance the information exchange between upper and lower layers, we design a special multi-cross-scale feature pyramid network (M-FPN). The C2f-Hierarchical-Phantom Convolution (C2f-HPC) module in the network effectively reduces information loss by fine-grained multi-scale feature fusion. Ultimately, adaptive spatial feature fusion detection Head (ASFFDHead) introduces an additional P2 detection head to enhance the resolution of feature maps to better locate small objects. Moreover, the ASFF mechanism is employed to optimize the detection process by filtering out information conflicts during multi-scale feature fusion, thereby significantly optimizing small object detection capability. Using YOLOv8n as the baseline, SMA-YOLO is evaluated on the VisDrone2019 dataset, achieving a 7.4% improvement in mAP@0.5 and a 13.3% reduction in model parameters, and we also verified its generalization ability on VAUDT and RSOD datasets, which demonstrates the effectiveness of our approach.https://www.mdpi.com/2072-4292/17/14/2421UAV imagesfeature fusionsmall objectsYOLOv8SMA-YOLO
spellingShingle Shenming Qu
Chaoxu Dang
Wangyou Chen
Yanhong Liu
SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
Remote Sensing
UAV images
feature fusion
small objects
YOLOv8
SMA-YOLO
title SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
title_full SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
title_fullStr SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
title_full_unstemmed SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
title_short SMA-YOLO: An Improved YOLOv8 Algorithm Based on Parameter-Free Attention Mechanism and Multi-Scale Feature Fusion for Small Object Detection in UAV Images
title_sort sma yolo an improved yolov8 algorithm based on parameter free attention mechanism and multi scale feature fusion for small object detection in uav images
topic UAV images
feature fusion
small objects
YOLOv8
SMA-YOLO
url https://www.mdpi.com/2072-4292/17/14/2421
work_keys_str_mv AT shenmingqu smayoloanimprovedyolov8algorithmbasedonparameterfreeattentionmechanismandmultiscalefeaturefusionforsmallobjectdetectioninuavimages
AT chaoxudang smayoloanimprovedyolov8algorithmbasedonparameterfreeattentionmechanismandmultiscalefeaturefusionforsmallobjectdetectioninuavimages
AT wangyouchen smayoloanimprovedyolov8algorithmbasedonparameterfreeattentionmechanismandmultiscalefeaturefusionforsmallobjectdetectioninuavimages
AT yanhongliu smayoloanimprovedyolov8algorithmbasedonparameterfreeattentionmechanismandmultiscalefeaturefusionforsmallobjectdetectioninuavimages