Open AccessStudy on Hot Spot Detection Algorithm based on K-Means
Author(s) -
ChenghaoDeng,
YuShen,
KanjianZhang
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/793/1/012017
Subject(s) - hot spot (computer programming) , rectangle , cluster analysis , photovoltaic system , algorithm , computer science , energy (signal processing) , domain (mathematical analysis) , power (physics) , k means clustering , artificial intelligence , engineering , mathematics , electrical engineering , mathematical analysis , statistics , physics , geometry , quantum mechanics , operating system
Solar energy is a widely used clean energy at present, and it is inexhaustible. As the main power generation device, photovoltaic modules play an important role in solar power generation. Among the various abnormalities of photovoltaic modules, the hot spot effect will not only reduce the power generation efficiency of photovoltaic modules, but even cause fires due to high temperatures. In this paper, k-means algorithm is adopted for photovoltaic hot spot detection. K-means is an unsupervised machine learning clustering algorithm, which clusters the pixel points of the image, and extracting the connected domain with the points with the shape similar to rectangular rectangle in the clustering, that is, obtaining the hot spot area. At the same time, this paper also uses the traditional computer vision MSER algorithm for experiments, compares the MSER algorithm with the K-means algorithm, and draws the conclusion that the traditional computer vision algorithm has low accuracy but high speed. Machine learning algorithms are accurate but fast.