The Research Of Computer-aided Plant Species Identification Based On Leaf Feature

As the development of digital image processing and pattern recognition, the plant species identification based on computer is likely to be accessed. Since the characteristics can be obtained in the forms of numerical images, the efficiency of plant species identification can be improved with image processing and pattern recognition. Computer-Aided Plant Species Identification technology tries to recognize the known plant species by salient features of the leaf. The focus of system is to extract the stable features of plants, which are discriminable from others, then to classify and recognize plant species. It acts significantly on plant digital museum system and systematic botany which is the groundwork for research and development of plant.The basic concept and theory of computer-aided plant species identification is first described in this paper. According to different ways and techniques to conclude and compare the existing algorithms of computer–aided plant species identification. Then based on the shortcoming of the existing algorithms, we proposed a new method for plant species identification using leaf feature and the neural network. This algorithm is improved on feature extraction. It firstly apply preprocess on leaf image, extracted the geometrical feature of leaf shape by the 2-D moment invariants, then Wavelet statistical features are used to extract leaf texture information. Self-Organizing Feature Map (SOM) neural network has the advantages of simple structure, ordered mapping topology and low complexity of learning. It is suitable for many complex problems such as multi-class pattern recognition, high dimension input vector and large quantity training data. So this paper use SOM neural network to identify the plant species. The experimental result illustrates the effectiveness of this method.The result of the proposed algorithm not only takes a further step to the technique of computer-aided plant species identification, but also remarkably enhances the correct rate of computer-aided plant species identification. At the end of this paper we have a prospect of technique of computer-aided plant species identification. The extensive application in the future of the plant species identification technique will also attracts more algorithms.