基于fcm和svm相结合的作物病害图像分割方法研究-信号与信息处理专业论文.docxVIP

万方数据 万方数据 摘要 (3)在图像分割结果评价准则方面，本文选用“分类正确率”和“直方图” 进行效果评定。 (5)为了满足用户的需要，本文利用 MATLAB 语言，研究开发出了作物病害 图像分割界面的开发。 关键词：图像分割 特征提取 自动选取训练样本 FCM；SVM II Abst Abstract Abstract Image segmentation is to separate the target of interest segmentation from background, and the result directly influences the latest image analysis and recognition. Based on The segmentation technology based on crop disease images is to segmented disease spot from the disease images, in order to the following diagnose and recognize the disease. A learning method based on statistical learning theory is proposed, it is Support Vector Machine(SVM).It has been widely used in many areas, such as data classification, pattern recognition. In recent years, the segmentation based on SVM has been used in the image segmentation, and can achieve good effect. SVM algorithm applied to image segmentation of disease spot is largely based on classification using some image pixels gray and other features as training samples to train SVM, and in the following steps, use the trained classifier to segment the whole image. the choice of training samples affects the classifier and further affects image segmentations effect of the disease spot. Conventional SVM algorithm based on artificial in selecting training samples, but if this, it is largely influenced by subjective, and the samples is random, time consuming and affecting the segmentation results. therefore, how to select the training samples automatically and making the trained samples widely representative of the samples type, will be the key in the process of SVM applied in the disease spot image segmentation. Through a large number of simulation experiments, in the thesis the algorithm based on SVM and Fuzzy C-Means(FCM) is proposed. The main works are concluded as follows: The thesis analyzed and summarized the image segmentation approaches based on SVM，and in the process of image segmentation using SVM exits the problem, the training samples cant selected automaticall