The Study Of Digital Forage Image Processing Techniques

Pasture is an important one of the growing crops, and alfalfa as a representative of the northern pasture varieties has a huge cultivated area. Developing the forage crop fanning vigorously is the important security and basic needs for the development of animal husbandry bases, it can improve the supply capacity of livestock products.Inner Mongolia region has a spars population, and has a huge pastureland area. It is very difficult for a lot of remote areas to judge the growth status of pasture using traditional field observation. Analyzing the image which after shooting using the digital image processing technology, that can get the information of pasture growth quickly. The technology realized remote observations, and provides a data support for future forage production and management.This paper analyses digital image processing technology which be used after the pasture and the background image separation. The results as follows:(1) This paper analyses preprocessing and segmentation algorithms of alfalfa image, including the image gray of image preprocessing, images de-nosing, edge detection, image mathematical morphological processing for image segmentation algorithm, including single-threshold segmentation、clustering segmentation based on the Euclidean distance, FCM segmentation, watershed segmentation, the optimal threshold segmentation. That completed the alfalfa image from the original color image to binary image.(2) The peaks and valleys dividing line is not obvious for the use of common gray method. This paper provides a improved algorithm on gray method, which reduced the difficulty for the subsequent image segmentation algorithm processing.(3) This paper improved the optimal threshold algorithm of segmentation algorithm, and improved the algorithm implementation efficiency significantly.(4) The paper get a complete image segmentation method through the combination of improved image segmentation algorithm and image gray method. The paper found the best segmentation methods through corresponding experiments to segmentation methods and different types of pasture image. It proves a effective algorithm choice for the next research of digital image processing technology.