A method of image segmentation based on mutual information and threshold iteration of multi-spectral image fusion

摘要

The surface color of some mature fruit has little difference with the background color or the entire fruit surface color has inconsistent under the nature growth condition. This fruit image segmentation is very difficulty with the color threshold or the segmentation result is the serious error of the goal outline. A new mutual information threshold iteration segmentation method based on the multi-spectrum image fusion is proposed. First, gather visible light color image and the near-infrared image of the same target, fusion processing the multi-spectrum image using image fusion algorithm (PCA-LWT) based on PCA and lifting wavelet transformation; Then, carries on mutual information optimal threshold value iteration division to the fusion image. Through target extraction experiment of many tomato images, the result indicate carrying on mutual information optimal threshold value iteration segmentation to the visible light and the near-infrared PCA-LWT fusion image, may obtain good segmentation effect, and can solve this kind of image segmentation problem well.