Fast segmentation algorithm of PCB image using 2D OTSU improved by adaptive genetic algorithm and integral image

2D OTSU achieves good image segmentation performance in thresholding-based segmentation tasks. However, for the real-time detection of printed circuit board (PCB) defects, this method is complicated and cannot meet the real-time requirements. In view of the above phenomenon, this paper proposes an improved 2D OTSU combining adaptive genetic algorithm and integral image algorithm. The adaptive genetic algorithm transforms the threshold selection of 2D OTSU into the optimization of an inter-class variance measure. The integral image algorithm reduces a lot of repeated calculations in the optimization process of an inter-class variance measure. Experimental results show that the proposed algorithm greatly reduces the amount of computation and time on the basis of ensuring the performance of PCB image segmentation. Under the condition of low contrast between line and background and uneven illumination, the proposed algorithm has better segmentation performance on PCB images.

Automated summary

2D OTSU achieves good image segmentation performance in thresholding-based segmentation tasks. However, it is not suitable for real-time detection of PCB defects due to its complexity and inability to meet real-time requirements. In this paper, an improved 2D OTSU method combining adaptive genetic algorithm and integral image algorithm is proposed. The adaptive genetic algorithm transforms the threshold selection into an optimization problem, while the integral image algorithm reduces repeated calculations. Experimental results show that the proposed algorithm reduces computation and time while ensuring the performance of PCB image segmentation, especially under low contrast and uneven illumination conditions.