基于Android的植物叶片图像病害检测

Plant leaf image disease detection based on Android

针对目前农业智能系统开发平台多基于Windows Mobile操作系统,开源性、免费性和人机交互性等方面较差等问题,提出了基于Android平台的一种最大类间差法和Canny算子结合的植物叶片图像病害检测方法.该方法首先用自适应中值滤波对叶片图像进行平滑处理,再对图像进行灰度变换,然后用最大类间差法对图像进行二值化,最后基于Canny算子对图像进行边缘检测.实验结果表明,该方法实现了Android平台下的图像检测,能有效减少边缘噪声,并且能够有效提取叶片图像病害部分的边缘,具有很好的鲁棒性、有效性和准确性.

Aiming at the poor performance of agricultural intelligence development platform in open source,free charge and human-computer interaction based on Windows Mobile,under Android platform,a method of leaf image disease detection based on maximum between-cluster variance and Canny operator was proposed. Firstly,adaptive median filter was used to smooth leaf images. Secondly,the images were processed by grey scale transformation. Then,the transformed images were segmented to bi-value images by adopting a method based on maximum between-cluster variance. Finally,the bi-value images images were processed by edge detection based on Canny operator. Experimental result showed that this method achieved leaf image disease detection under Android platform. The method which could reduce marginal noise and extract the edge of leaf image disease effectively provided robustness,effectiveness and correctness.