EU PVSEC Programme Online
EU PVSEC 2021, 6 - 10 September 2021
Presentation: 5CV.2.18 Solar Photovoltaic Power Plant Inventory and Solar Modules Counting by Convolutional Neural Network from Aerial Imagery
Type: Visual
Date: Wednesday, 8th September 2021
13:30 - 15:00
Author(s): Y.-R. Lin, C.-Y. Huang, J.-S. Zhang, H.-L. Chen
Presenter / Speaker: Y.-R. Lin, ITRI, Tainan, Taiwan
Event: Conference Conference
Session: 5CV.2 Operation, Performance and Maintenance of PV Systems
Topic: 5. 3 Operation, Performance and Maintenance of PV Systems
Keywords: Aerial Imagery, Computer Vision, PV, Convolutional Neural Networks
Summary / Abstract: Taiwan is an island located in the subtropics and near the Pacific Ocean, as a result, Taiwan is hit by typhoons almost every summer. Solar photovoltaic power plants suffer from storm damage and parts of solar panels could be lost. Nowadays, Taiwan is committed to promoting installations of solar photovoltaic plants. However, the government doesn’t verify those existing plants are completeness or not. In this work, Aerial imagery or highresolution satellite imagery are used to take inventory of photovoltaic systems. A training model for PV plants detection is set up by CNN (Convolutional Neural Network). After capturing the region of PV plants, a counting method, a consequent image processing techniques, is applied to count the amount of solar panels. By clarifying the amount of solar modules, those would be compared with the original information. A successful case at Liujia in Tainan verifying the counting method, the recognition rate has reach over 98%.