Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy

Traditional target detection models face challenges in recognizing urban high-altitude remote sensing targets due to complex background noise and significant variations in target scale. These challenges can result in loss of feature information and missed object detection. In light of this, this art...

Full description

Saved in:
Bibliographic Details
Main Authors: Luxuan Bian, Zijun Gao, Jue Wang, Bo Li
Format: Article
Language:English
Published: Taylor & Francis Group 2024-01-01
Series:Geocarto International
Subjects:
Online Access:https://www.tandfonline.com/doi/10.1080/10106049.2024.2322061
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1846129188636983296
author Luxuan Bian
Zijun Gao
Jue Wang
Bo Li
author_facet Luxuan Bian
Zijun Gao
Jue Wang
Bo Li
author_sort Luxuan Bian
collection DOAJ
description Traditional target detection models face challenges in recognizing urban high-altitude remote sensing targets due to complex background noise and significant variations in target scale. These challenges can result in loss of feature information and missed object detection. In light of this, this article introduces a novel dual-gated feature mechanism and adaptive fusion strategy. First, the dual-gated feature mechanism enables selective suppression or enhancement of multilevel features, thereby reducing the interference of complex environmental noise in remote sensing on feature fusion. Second, the adaptive fusion strategy and module facilitate multilevel scale feature fusion, and by dynamically learning fusion weights, they mitigate scale conflicts during the feature extraction process and preserve feature information. Experimental comparisons and analysis on the RSOD and NWPU VHR-10 public datasets showcase the effectiveness of the proposed method. In comparison to current mainstream detection methods, the improved approach presented in this article demonstrates significant advantages in terms of detection performance and efficiency.
format Article
id doaj-art-9f683a80202048a0a9aa3889161eb809
institution Kabale University
issn 1010-6049
1752-0762
language English
publishDate 2024-01-01
publisher Taylor & Francis Group
record_format Article
series Geocarto International
spelling doaj-art-9f683a80202048a0a9aa3889161eb8092024-12-10T08:23:09ZengTaylor & Francis GroupGeocarto International1010-60491752-07622024-01-0139110.1080/10106049.2024.2322061Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategyLuxuan Bian0Zijun Gao1Jue Wang2Bo Li3College of Information Science and Engineering, School of Information Science and Engineering, Dalian Polytechnic University, Dalian, ChinaCollege of Information Science and Engineering, School of Information Science and Engineering, Dalian Polytechnic University, Dalian, ChinaCollege of Information Science and Engineering, School of Information Science and Engineering, Dalian Polytechnic University, Dalian, ChinaCollege of Information Science and Engineering, School of Information Science and Engineering, Dalian Polytechnic University, Dalian, ChinaTraditional target detection models face challenges in recognizing urban high-altitude remote sensing targets due to complex background noise and significant variations in target scale. These challenges can result in loss of feature information and missed object detection. In light of this, this article introduces a novel dual-gated feature mechanism and adaptive fusion strategy. First, the dual-gated feature mechanism enables selective suppression or enhancement of multilevel features, thereby reducing the interference of complex environmental noise in remote sensing on feature fusion. Second, the adaptive fusion strategy and module facilitate multilevel scale feature fusion, and by dynamically learning fusion weights, they mitigate scale conflicts during the feature extraction process and preserve feature information. Experimental comparisons and analysis on the RSOD and NWPU VHR-10 public datasets showcase the effectiveness of the proposed method. In comparison to current mainstream detection methods, the improved approach presented in this article demonstrates significant advantages in terms of detection performance and efficiency.https://www.tandfonline.com/doi/10.1080/10106049.2024.2322061Remote sensingalgorithm feature extractionimage fusion
spellingShingle Luxuan Bian
Zijun Gao
Jue Wang
Bo Li
Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
Geocarto International
Remote sensing
algorithm feature extraction
image fusion
title Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
title_full Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
title_fullStr Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
title_full_unstemmed Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
title_short Urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
title_sort urban object detection algorithm based on feature enhancement and progressive dynamic aggregation strategy
topic Remote sensing
algorithm feature extraction
image fusion
url https://www.tandfonline.com/doi/10.1080/10106049.2024.2322061
work_keys_str_mv AT luxuanbian urbanobjectdetectionalgorithmbasedonfeatureenhancementandprogressivedynamicaggregationstrategy
AT zijungao urbanobjectdetectionalgorithmbasedonfeatureenhancementandprogressivedynamicaggregationstrategy
AT juewang urbanobjectdetectionalgorithmbasedonfeatureenhancementandprogressivedynamicaggregationstrategy
AT boli urbanobjectdetectionalgorithmbasedonfeatureenhancementandprogressivedynamicaggregationstrategy