GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model

Desert encroachment significantly threatens the living and activity space of humanity, and undertaking human-directed vegetation restoration is one of the effective ways to prevent desert expansion. In the process of desert vegetation restoration, counting the number of tree saplings for rapidly ass...

Full description

Saved in:
Bibliographic Details
Main Authors: Shuzhen Hua, Biao Yang, Xinchang Zhang, Ji Qi, Fengxi Su, Jing Sun, Yongjian Ruan
Format: Article
Language:English
Published: MDPI AG 2025-02-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/4/691
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850081222888259584
author Shuzhen Hua
Biao Yang
Xinchang Zhang
Ji Qi
Fengxi Su
Jing Sun
Yongjian Ruan
author_facet Shuzhen Hua
Biao Yang
Xinchang Zhang
Ji Qi
Fengxi Su
Jing Sun
Yongjian Ruan
author_sort Shuzhen Hua
collection DOAJ
description Desert encroachment significantly threatens the living and activity space of humanity, and undertaking human-directed vegetation restoration is one of the effective ways to prevent desert expansion. In the process of desert vegetation restoration, counting the number of tree saplings for rapidly assessing the survival rate of vegetation (such as <i>Haloxylon ammodendron</i>) is a critical task within the restoration process. However, traditional ground-based statistical methods are resource-intensive and time-consuming. This paper proposed a novel unsupervised fine segmentation framework driven by Grounding DINO prompt generation and optimization segment anything model, termed GDPGO-SAM, designed for the segmentation of desert vegetation from UAV-derived remote sensing imagery, thereby facilitating the rapid inventory of tree saplings counts. The framework combines the Grounding DINO object detector and the pre-trained visual model SAM, employing a task-prior-based prompt optimization mechanism to effectively capture the innate features of desert vegetation. This method achieves zero-sample instance segmentation of desert vegetation with an overall accuracy (OA) of 96.56%, a mean Intersection over Union (mIoU) of 81.50%, and a kappa coefficient (kappa) of 0.782, successfully overcoming the limitations of traditional supervised models that rely on passive memorization rather than true recognition. This research significantly enhances the precision of vegetation extraction and canopy depiction, providing strong support for the management of desert vegetation restoration and combating desert expansion.
format Article
id doaj-art-7df16c478daf4cf3a83d796129f48ebb
institution DOAJ
issn 2072-4292
language English
publishDate 2025-02-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-7df16c478daf4cf3a83d796129f48ebb2025-08-20T02:44:47ZengMDPI AGRemote Sensing2072-42922025-02-0117469110.3390/rs17040691GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything ModelShuzhen Hua0Biao Yang1Xinchang Zhang2Ji Qi3Fengxi Su4Jing Sun5Yongjian Ruan6School of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, ChinaSociety of Entrepreneurs & Ecology Foundation, Beijing 100027, ChinaSchool of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, ChinaSchool of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, ChinaSchool of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, ChinaSociety of Entrepreneurs & Ecology Foundation, Beijing 100027, ChinaSchool of Geography and Remote Sensing, Guangzhou University, Guangzhou 510006, ChinaDesert encroachment significantly threatens the living and activity space of humanity, and undertaking human-directed vegetation restoration is one of the effective ways to prevent desert expansion. In the process of desert vegetation restoration, counting the number of tree saplings for rapidly assessing the survival rate of vegetation (such as <i>Haloxylon ammodendron</i>) is a critical task within the restoration process. However, traditional ground-based statistical methods are resource-intensive and time-consuming. This paper proposed a novel unsupervised fine segmentation framework driven by Grounding DINO prompt generation and optimization segment anything model, termed GDPGO-SAM, designed for the segmentation of desert vegetation from UAV-derived remote sensing imagery, thereby facilitating the rapid inventory of tree saplings counts. The framework combines the Grounding DINO object detector and the pre-trained visual model SAM, employing a task-prior-based prompt optimization mechanism to effectively capture the innate features of desert vegetation. This method achieves zero-sample instance segmentation of desert vegetation with an overall accuracy (OA) of 96.56%, a mean Intersection over Union (mIoU) of 81.50%, and a kappa coefficient (kappa) of 0.782, successfully overcoming the limitations of traditional supervised models that rely on passive memorization rather than true recognition. This research significantly enhances the precision of vegetation extraction and canopy depiction, providing strong support for the management of desert vegetation restoration and combating desert expansion.https://www.mdpi.com/2072-4292/17/4/691vegetation canopy segmentationinstance segmentationzero-shotsegment anything model
spellingShingle Shuzhen Hua
Biao Yang
Xinchang Zhang
Ji Qi
Fengxi Su
Jing Sun
Yongjian Ruan
GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
Remote Sensing
vegetation canopy segmentation
instance segmentation
zero-shot
segment anything model
title GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
title_full GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
title_fullStr GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
title_full_unstemmed GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
title_short GDPGO-SAM: An Unsupervised Fine Segmentation of Desert Vegetation Driven by Grounding DINO Prompt Generation and Optimization Segment Anything Model
title_sort gdpgo sam an unsupervised fine segmentation of desert vegetation driven by grounding dino prompt generation and optimization segment anything model
topic vegetation canopy segmentation
instance segmentation
zero-shot
segment anything model
url https://www.mdpi.com/2072-4292/17/4/691
work_keys_str_mv AT shuzhenhua gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT biaoyang gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT xinchangzhang gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT jiqi gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT fengxisu gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT jingsun gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel
AT yongjianruan gdpgosamanunsupervisedfinesegmentationofdesertvegetationdrivenbygroundingdinopromptgenerationandoptimizationsegmentanythingmodel