Fast Partial Shading Detection on PV Modules for Precise Power Loss Ratio Estimation Using Digital Image Processing

Table 1

Summary of IRT methodology to detect PV module faults.

No	Detection output	[Cit] year	Algorithm	Input data	Detection rate
1	Detect modules and hot spots	[23] 2015	DIP filtering	Thermal infrared image	Prec: 97.9% and 97.4%
2	Detect defective PV module	[24] 2017	DIP statistical and histogram analysis	Aerial thermal image	Acc: 97%
3	Detect PV module	[25] 2017	DIP	Aerial thermal image	Prec: 82%
4	Binary classification of defective and nondefective modules	[26] 2018	n-Bayes classifier	Thermal infrared image	Acc: 98.4%
5	Detect modules and hot spots	[27] 2018	Geometric transformation and probabilistic analysis	Thermal infrared image	Prec: 96.52%
6	Detect and localize hot spots	[28] 2019	Masking and thresholding in HSV	Thermal infrared image	Acc: 100% (only 3 images)
7	Detect modules and hot spots	[29] 2020	DIP (Sobel and Canny operator) and CNN	Aerial thermal image	Prec: 90.91%
8	PV panel faults classifications: healthy, delamination, and EVA discoloring	[30] 2020	Fuzzy rule-based classification	Thermal infrared image	Acc: 94%
9	Detect modules and hot spots	[31] 2020	Developed R-CNN	Aerial thermal image	Acc: 99% and 92.25%
10	Binary classification of defective and nondefective modules	[32] 2020	DIP and SVM	Thermal infrared image	Acc: 97%