Application of Data-Driven Iterative Learning Algorithm in Transmission Line Defect Detection

Table 1

Statistical table of data set quantity.

Datasets	Lack of locking pin (lsqxz)		Rusted-on nut (lmsx)
	Number of pictures	Number of targets	Number of pictures	Number of targets
Initial training data	2682	3969	1621	2446
The first iterative learning + sample mining	19603	30841	7803	14514
The second iterative learning + sample mining	57272	73415	16238	24840
Test data	1899	2484	1412	1901