Key Technologies Research On Green Plant Species Identification Based On Local Edge Pattern Algorithm

Green plant species identification has important application prospects and potential economic value for the protection and utilization of plant resources,clarifying the evolution of plants,exploring the relationship between plants,protecting the ecological environment,agriculture and horticultural applications,and so on.With the advancement of computer science and technology,the field of pattern recognition and image processing has been developed by leaps and bounds.The use of computer technology instead of traditional manual work which efficiently and automatically identifies plant species has attracted the wide attention of researchers.How to use the digital image processing technology to extract plant features so as to achieve the classification and identification has always become a core research topic in the field of computer vision.Therefore,in this paper the feature extraction of green plant species is focused on with a more in-depth study,and the main research work is as follows:Firstly,by the analysis,the effective feature of the green plant species identification is the texture edge.Then five kinds of texture feature extraction algorithms are described in detail.Because the traditional edge operators are difficult to extract the texture edge feature from the perspective of multi-scale,a multi-scale local edge pattern algorithm based on traditional edge operators is proposed.Secondly,due to the inherent defects of the traditional edge operators on the basis of first derivative,a varied local edge pattern algorithm with multi-scale and multi-direction characteristics is proposed.At the same time,two feature fusion methods are designed to improve the recognition rate of green plant species identification.Finally,a green plant species identification system is designed and set up which integrates the proposed algorithm in this thesis and the traditional classic algorithm in a prototype system.The two aspects of system interface and algorithm embedding was described in detail.The experimental results show that the system can realize green plant species identification and satisfactory result are obtained.