Optimizing PV Panel Segmentation in Complex Environments Using Pre-Training and Simulated Annealing Algorithm: The JSWPVI

Rui Zhang, Ruikai Hong, Qiannan Li, Xu He, Age Shama, Jichao Lv and Renzhe Wu ()
Additional contact information
Rui Zhang: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China
Ruikai Hong: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China
Qiannan Li: Henan Provincial Key Laboratory of Ecological Environment Remote Sensing, Zhengzhou 450046, China
Xu He: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China
Age Shama: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China
Jichao Lv: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China
Renzhe Wu: Faculty of Geosciences and Engineering, Southwest Jiaotong University, Chengdu 611756, China

Land, 2025, vol. 14, issue 6, 1-20

Abstract: Photovoltaic (PV) technology, as a crucial source of clean energy, can effectively mitigate the impact of climate change caused by fossil fuel-based power generation. However, improper use of PV installations may encroach upon agricultural land, grasslands, and other land uses, thereby affecting local ecosystems. Exploring the spatial characteristics of centralized or distributed PV installations is essential for quantifying the development of clean energy and protecting agricultural land. Due to the distinct characteristics of centralized and distributed PV installations, large-scale mapping methods based on satellite remote sensing are insufficient for creating detailed PV distribution maps. This study proposes a model called Joint Semi-Supervised Weighted Adaptive PV Panel Recognition Model (JSWPVI)to achieve reliable PV mapping using UAV datasets. The JSWPVI employs a semi-supervised approach to construct and optimize a comprehensive segmentation network, incorporating the Spatial and Channel Weight Adaptive Model (SCWA) module to integrate different feature layers by reconstructing the spatial and channel weights of feature maps. Finally, a guided filtering algorithm is used to minimize non-edge noise while preserving edge integrity. Our results demonstrate that JSWPVI can accurately extract PV panels in both centralized and distributed scenarios, with an average extraction accuracy of 91.1% and a mean Intersection over Union of 77.7%. The findings of this study will assist regional policymakers in better quantifying renewable energy potential and assessing environmental impacts.

Keywords: photovoltaic panels; contrast learning; simulated annealing algorithm; image segmentation; aerial remote sensing; adaptive weights (search for similar items in EconPapers)
JEL-codes: Q15 Q2 Q24 Q28 Q5 R14 R52 (search for similar items in EconPapers)
Date: 2025
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/2073-445X/14/6/1245/pdf (application/pdf)
https://www.mdpi.com/2073-445X/14/6/1245/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jlands:v:14:y:2025:i:6:p:1245-:d:1675853

Access Statistics for this article

Land is currently edited by Ms. Carol Ma

More articles in Land from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().