Research On Plant Leaf Recognition Method Based On Sparse Representation Classification

Plant identification is an important branch of plant science,which is the basis for further development and utilization of plant resources.The traditional way of plant identification is to make artificial recognition based on the physical form,chemical composition and biological structure of plants,and the disadvantages of low accuracy and poor efficiency are found.Since the blade image has the advantages of two-dimensional shape,large amount of information and easy access,it has become a very important research direction for plant identification.Firstly,the pretreatment of plant image and feature extraction method are studied,and the basic process of sparse representation classification is introduced.Image pretreatment includes gray-scale,image segmentation,morphological processing,etc.The feature extraction of plant image includes shape feature and texture feature extraction.Then,studied the botanical garden leaf recognition method based on sparse representation classification,first of all,the original sparse representation method is applied to the plant leaves recognition,USES the original leaf image dictionary initialized,using the OMP algorithm directly sparse expression of blade under test sample in the dictionary.by the least deviation degree identification classification,found that there are large amount of calculation and low recognition rate shortcomings;Then in view of the sparse representation classification is improved,using leaf texture feature matrix instead of the original leaf images to dictionary initialized,and the K-SVD dictionary learning methods to study optimization of dictionary,has the stronger adaptability,finally using the OMP algorithm under test sample texture matrix sparse expression on the optimal dictionary,complete plant leaves recognition classification.Finally,Flavia data sets and ICL data sets were used to verify this method,the experimental results show that the method has high recognition rate,and missing data block blade has stronger robustness;The design of plant leaf identification system was completed by using matlab GUI platform..