Study On Plant Species Recognition Based On Leaf Characteristics And Realization Of Recognition System

Plant classification and identification is the basis of the Plant Protection and Research. With the rapid development of information science and technology, image processing and pattern recognition technology has been widely applied in many fields.In recent years, with the rapid development of mobile terminal photography and promotion of mobile Internet, people become more and more interested in the cognition of plant, which provides an opportunity for the development of APP about " Plant identification ".This thesis chose plant leaf image as research object, used digital image processing technology to preprocess the image of plant leaves,and proposed a AdaboostLDA classification algorithm,which was used in the classification and identification of plant leaves. At last we implement a mobile application running on Android devices. The main contributions are as follows:(1) In order to obtain more useful information of leaf images, in this thesis, we makes a series of processing of the image, including the Standardized processing,grayscale processing, binary processing, morphological processing, image shapening processing.(2) Based on the research on the characteristic information of plant leaf image,this thesis extracted five basic geometric features,including the leaves of perimeter,area, diameter and so on.Then we further extracted 12 morphological characters of aspect ratio, shape factor and texture feature, and then used PCA to reduce the dimension of the feature.(3) We proposed a AdaboostLDA classification algorithm, by changing the weights of input data, the performance of extreme learning machine can be improved.Compared with BP neural network, the training speed is much faster,and the classification accuracy is similar.(4) Finally, plant species identification system is realized based on android platform,which converting research results to available APP.