MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance

The detection of distant tiny objects in aerial imagery plays a pivotal role in early warning, localization, and recognition tasks. However, due to the scarcity of appearance information, minimal pixel representation, susceptibility to blending with the background, and the incompatibility of convent...

Full description

Saved in:
Bibliographic Details
Main Authors: Nan Su, Zilong Zhao, Yiming Yan, Jinpeng Wang, Wanxuan Lu, Hongbo Cui, Yunfei Qu, Shou Feng, Chunhui Zhao
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/16/23/4485
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850261351003324416
author Nan Su
Zilong Zhao
Yiming Yan
Jinpeng Wang
Wanxuan Lu
Hongbo Cui
Yunfei Qu
Shou Feng
Chunhui Zhao
author_facet Nan Su
Zilong Zhao
Yiming Yan
Jinpeng Wang
Wanxuan Lu
Hongbo Cui
Yunfei Qu
Shou Feng
Chunhui Zhao
author_sort Nan Su
collection DOAJ
description The detection of distant tiny objects in aerial imagery plays a pivotal role in early warning, localization, and recognition tasks. However, due to the scarcity of appearance information, minimal pixel representation, susceptibility to blending with the background, and the incompatibility of conventional metrics, the rapid and accurate detection of tiny objects poses significant challenges. To address these issues, a single-stage tiny object detector tailored for aerial imagery is proposed, comprising two primary components. Firstly, we introduce a light backbone-heavy neck architecture, named the Global Context Self-Attention and Dense Nested Connection Feature Extraction Network (GC-DN Network), which efficiently extracts and fuses multi-scale features of the target. Secondly, we propose a novel metric, MMPW, to replace the Intersection over Union (IoU) in label assignment strategies, Non-Maximum Suppression (NMS), and regression loss functions. Specifically, MMPW models bounding boxes as 2D Gaussian distributions and utilizes the Mixed Minimum Point-Wasserstein Distance to quantify the similarity between boxes. Experiments conducted on the latest aerial image tiny object datasets, AI-TOD and VisDrone-19, demonstrate that our method improves AP50 performance by 9.4% and 5%, respectively, and AP performance by 4.3% and 3.6%. This validates the efficacy of our approach for detecting tiny objects in aerial imagery.
format Article
id doaj-art-bd5f903ef2b1477da62c90c161f4c2df
institution OA Journals
issn 2072-4292
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-bd5f903ef2b1477da62c90c161f4c2df2025-08-20T01:55:27ZengMDPI AGRemote Sensing2072-42922024-11-011623448510.3390/rs16234485MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein DistanceNan Su0Zilong Zhao1Yiming Yan2Jinpeng Wang3Wanxuan Lu4Hongbo Cui5Yunfei Qu6Shou Feng7Chunhui Zhao8College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaThe Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100190, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaSchool of Light Industry, Harbin University of Commerce, Harbin 150028, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaCollege of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, ChinaThe detection of distant tiny objects in aerial imagery plays a pivotal role in early warning, localization, and recognition tasks. However, due to the scarcity of appearance information, minimal pixel representation, susceptibility to blending with the background, and the incompatibility of conventional metrics, the rapid and accurate detection of tiny objects poses significant challenges. To address these issues, a single-stage tiny object detector tailored for aerial imagery is proposed, comprising two primary components. Firstly, we introduce a light backbone-heavy neck architecture, named the Global Context Self-Attention and Dense Nested Connection Feature Extraction Network (GC-DN Network), which efficiently extracts and fuses multi-scale features of the target. Secondly, we propose a novel metric, MMPW, to replace the Intersection over Union (IoU) in label assignment strategies, Non-Maximum Suppression (NMS), and regression loss functions. Specifically, MMPW models bounding boxes as 2D Gaussian distributions and utilizes the Mixed Minimum Point-Wasserstein Distance to quantify the similarity between boxes. Experiments conducted on the latest aerial image tiny object datasets, AI-TOD and VisDrone-19, demonstrate that our method improves AP50 performance by 9.4% and 5%, respectively, and AP performance by 4.3% and 3.6%. This validates the efficacy of our approach for detecting tiny objects in aerial imagery.https://www.mdpi.com/2072-4292/16/23/4485tiny object detectionaerial imageryMMPW distanceglobal context self-attentiondense nested connection
spellingShingle Nan Su
Zilong Zhao
Yiming Yan
Jinpeng Wang
Wanxuan Lu
Hongbo Cui
Yunfei Qu
Shou Feng
Chunhui Zhao
MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
Remote Sensing
tiny object detection
aerial imagery
MMPW distance
global context self-attention
dense nested connection
title MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
title_full MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
title_fullStr MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
title_full_unstemmed MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
title_short MMPW-Net: Detection of Tiny Objects in Aerial Imagery Using Mixed Minimum Point-Wasserstein Distance
title_sort mmpw net detection of tiny objects in aerial imagery using mixed minimum point wasserstein distance
topic tiny object detection
aerial imagery
MMPW distance
global context self-attention
dense nested connection
url https://www.mdpi.com/2072-4292/16/23/4485
work_keys_str_mv AT nansu mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT zilongzhao mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT yimingyan mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT jinpengwang mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT wanxuanlu mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT hongbocui mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT yunfeiqu mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT shoufeng mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance
AT chunhuizhao mmpwnetdetectionoftinyobjectsinaerialimageryusingmixedminimumpointwassersteindistance