Plant Leaves Recognition Based On Leaf Contour And Corner Points

Plants are the pillar of human survival, the diversity of plants plays a very important role to human life, and protection of plant diversity, the first is to solve the problem of plant classification. To identify the plant rapidly and accurately, a plant identification technique which based on CBIR is proposed. Relatively to other plant organs, leaf which has a stable form, long lifecycle, easily to collect and two D structure, is often chosen as the basis for plant identification. Leaf classification is an important component of computerized living plant recognition. The leaf contains important information for plant species identification despite its complexity. Therefore, it is very important to study leaf recognition technology.To extract leaves features and recognize leaves types by image analysis are the research concentration and main content of the paper. The corner point of plant contains the local shape of important information, this paper corner point is detected by the improved method based on the traditional Harris detection.These points are existed in multi-scale, besides the corner detection is more stable and accurate positioning At the same time This paper introduces a method of recognizing leaf images based on triangular representation composed of corner points and edge contour of leaf, it calculate parameters of a triangle to represent the shape of the leaf shape feature,such as the area, angles and so on. And it can represents the spatial correlation between salient points of the leaf and leaf margin and then combined with Kernelized Locality-Sensitive Hashing matching algorithm for classification.Based on the leaf shape feature, the paper builds a plant identification and verification simulation system. First adopted the pretreatment method on leaf, then get the Triangle information using the algorithm above-mentioned finally using Kernelized Locality Sensitive Hashing matching algorithm to get classification recognition rate. From the experiment data how that the effectiveness and the efficiency of the proposed method.