基于Android的玉米病虫害机器视觉诊断系统研究

为了使农业智能诊断系统更加廉价、便捷,有效地为普通农户服务,提出了一种基于Android手机的农业病虫害智能诊断系统。该系统使用Android智能手机对玉米病虫害部分进行图片拍摄,并将图像利用无线网上传至Web服务器,利用分割和匹配算法对病虫害部分进行智能化分析,最终将结果传输到手机用户端。为实现图像匹配的特征点提取,采用高斯差分的方法对图像进行分割和精确定位,使用聚类算法对匹配效果进行优化,并利用特征点的无限逼近,完成病虫害图像的匹配,从而诊断病虫害的类型。上传后的图像和Web服务器的规则库的图像进行匹配后可以生成病虫害的匹配结果信息,该信息可以通过Android智能系统接收,最终反馈给农户的手机客户端。通过测试发现:玉米病虫害诊断系统可从多幅图像里有效地对病虫害类型进行匹配,匹配成功率较高,系统的稳定性较好,具有很好的推广前景。

Research on the Diagnosis System of Maize Plant Diseases and Insect Pests Based on Android

In order to make the agricultural intelligent diagnosis system more cheap, convenient, efficient service for the ordinary farmers,it puts forward a kind of intelligent diagnosis system of agricultural plant diseases and insect pests based on Android mobile phone. The system uses Android intelligent mobile phone on plant diseases and insect pests of maize Part pictures. By using the wireless network to the Web server, it uses the segmentation and matching algorithm of pest and disease part of intelligent analysis. The final result is transmitted to the mobile phone user end. In order to realize im-age matching feature points extraction, it uses Gauss differential method for segmentation and precise location of the im-age,uses clustering algorithm to optimize the matching effect, and makes use of the characteristic points of infinite ap-proximation, image,then it completes plant diseases and insect pest diagnosis of plant diseases and insect pests. It can generate matching results of pests and diseases information through image rule and web server. The information can be re-ceived by Android intelligent system, ultimately back to the farmer's mobile phone client. Finally, through the test,it was found that maize pest diagnosis system can effectively realize the aim from multiple images of plant diseases and in-sect pests type matching. And the matching success rate is high, the system is stable, which has good popularizing pros-pect.