Reseach Of Leaf Image Processing Based On M-PCNN

From the 1990 s, PCNN(Pulse Coupled Neural Network) has been studied by more and more scientists. Moreover, the application of PCNN attracts a growing number of attentions especially. And there are many achievements so far. A great deal of improved models of PCNN have been presented and one of them is the m-PCNN — a multi-channel PCNN model based on the PCNN. And it is firstly used in the image fusion. It makes up the disadvantages of PCNN that the process of the operation is too long. Then it makes the number of the external input controllable, and it also changes the single-layer model to the multi-layer model. Therefore it simplifies the computational process and is very valuable in the research of improving algorithm.The work of this paper is studying the application of m-PCNN on the plant leaf images processing. Leaf is one of the most important organs of plant and it’s often used as one of the basic characters for plant classification. The developmental condition of leaf can provide us with lots of critical information, such as the plant’s health condition, the prospection of crop yield and so on. Image processing of leaf offers an effective platform for plant classification and growth observation.This paper aims to solve two problems: one is to detect the petiole, and the other is to extact the vein. The m-PCNN and the HSI(hue-saturation-intensity) color space are used in the research of the algorithms for the two problems. Finally,the experimental results prove the efficiency and feasibility of the algorithms.