Cotton Image Segmentation Based on Histogram Unit-linking PCNN

Abstract

Aiming at the misjudgment of picking target when cotton is covered by branches and leaves in natural scenes, based on the improved local threshold segmentation method based on Otsu, this paper proposes a pulse-coupled neural network(PCNN) based on equidistant threshold drop and histogram-based Cotton image segmentation method. Through comparison experiments, it is found that the histogram-based Unit-linking PCNN method has higher segmentation precision and faster segmentation speed based on retaining the contour features of cotton images, which is more suitable for the real-time requirements of cotton-picking robots.