基于图像分析的陕西苹果叶片病害识别系统设计与实现

发布时间：2018-04-20 17:53

本文选题：斑点落叶病 + 花叶病　；参考：《西北农林科技大学》2017年硕士论文

【摘要】：随着苹果种植面积的逐渐增大,果树病害种类随之增多,在某种程度上就有可能导致苹果产成率的降低,从而会影响到果农的经济收益。为解决此问题,实验利用目前比较成熟的图像处理技术和自动识别技术来实现陕西关中地区苹果叶片各种病害的识别。本文介绍利用图像处理技术自动检测和识别苹果叶片病害的方法,优化病害图像预处理算法,并针对图像分割及模式识别部分算法开展研究。为避免在图像处理过程中受叶片背景影响,采用基于灰度形态学图像平滑方法。通过比较不同的图像分割算法,对分割后得到的病斑或病斑轮廓进行特征提取,找出最能代表病斑特征的数据,然后利用模式识别的方法来识别苹果叶片病斑的种类和名称。从病害的种类和性质上来说,斑点落叶病、锈病、花叶病是陕西关中区域较为普遍的果树病害,因此本文以此为研究对象,分析并提取其特征参数,同时,设计相应的苹果叶片病害识别系统。本课题内容主要包括以下几点:(1)收集苹果叶片斑点落叶病、锈病、花叶病3种病害图像,并对所有的病害图像进行图像增强、去噪等预处理操作。(2)对预处理后的图像进行图像分割操作,从多种苹果叶片图像中分离病斑图像,并提取出不同病斑的相关特征参数。(3)通过基于决策理论的模式识别方法对3种苹果叶片病害图像进行较为准确的识别。(4)设计苹果叶片病害识别系统。本文设计出一种基于MATLAB的苹果叶片病害识别系统,能够应用于生产实践中,使苹果叶片病害识别技术较好地服务于广大果农。
[Abstract]:With the increase of apple planting area, the variety of fruit diseases will increase, which may lead to the decrease of apple yield to some extent, which will affect the economic benefits of fruit farmers. In order to solve this problem, the current mature image processing technology and automatic recognition technology are used to realize the identification of apple leaf diseases in Guanzhong area of Shaanxi Province. In this paper, the method of automatically detecting and recognizing apple leaf diseases by image processing technology is introduced, and the image preprocessing algorithm is optimized, and some algorithms of image segmentation and pattern recognition are studied. In order to avoid the influence of leaf background in image processing, a grayscale morphological image smoothing method is adopted. By comparing different image segmentation algorithms, the feature extraction of the disease-spot or disease-spot contour was carried out, and the most representative data was found, and then the type and name of the disease spot in apple leaf was identified by pattern recognition. As far as the species and properties of the disease are concerned, spot leaf disease, rust disease and mosaic disease are common diseases of fruit trees in Guanzhong region of Shaanxi Province. Therefore, this paper takes this disease as the research object, analyzes and extracts its characteristic parameters, at the same time, The corresponding apple leaf disease recognition system was designed. The main contents of this paper include the following points: 1) collect the three disease images of apple leaf spot, rust and mosaic, and enhance all the disease images. Pre-processing operations such as de-noising. 2) segmenting images after preprocessing and separating diseased images from a variety of apple leaf images. The relative characteristic parameters of different disease spots were extracted. (3) the apple leaf disease recognition system was designed by using the pattern recognition method based on the decision theory to identify the three apple leaf disease images accurately. In this paper, an apple leaf disease identification system based on MATLAB is designed, which can be applied to the production practice and make the apple leaf disease identification technology better serve for the majority of fruit farmers.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S436.611;TP391.41

【参考文献】