Study On Feature Extraction Of Leaf Edge In Vegetation Community

In modern agricultural production,leaf shape is the most common index to distinguish different vegetation communities,and the leaf shape is shown as the leaf edge in the image processing process of the existing research.Therefore,the extraction of leaf edge features is of great significance in distinguishing and accurately identifying different plant vegetation communities.In this paper,study on the extraction algorithm of leaf edge feature in vegetation community to extract leaf edges from vegetation communities,which provides a theoretical basis for the accurate identification of different plants.But the actual situation of the field vegetation community is very complex,the leaf overlap cover is very serious,and mixed with a variety of crops and weeds.Therefore,the study began with the plant leaves,and finally extended to the vegetation community.To extract the leaf edge of the plant image,the first step is to collect the plant image.After the image acquisition is completed,the image will be processed.Image segmentation is not only the first step of image processing,but also the basis of image processing.The effect of image segmentation directly affects a series of subsequent image processing processes.In this paper,the threshold segmentation algorithm is used to segment plant images.The segmentation features are the non-normalized super-green features,the normalized green features and the green features.The grayscale image of leaf was obtained by segmenting the plant image.Then the edge of the leaf gray image is extracted based on the existing edge extraction algorithm.The result of the binary map of leaf edge is very unsatisfactory,and the leaf edge of overlapping leaf is not extracted.By analyzing the segmentation effect of different segmentation features and the factors affecting the segmentation effect,it is found that the change of illumination intensity has great influence on the image.The leaf edge and leaf edge of plant image will have different trends under different light intensity.Therefore,this paper proposes a method to extract leaf edges from plant images based on different light intensities.Under different light intensity,the variation trend of leaf and leaf edge of plant image is different.In this paper,sobel template and perwitt template are used to process the grayscale image of leaf,showing different trends in leaf and leaf edges.It was found that the edge of the leaf was unstable under different light intensity and the change trend was large,but the leaf was stable under different light intensity and had no change trend.Therefore,this paper present the variance of grayscale eigenvalue of leaf grayscale image under different illumination intensity is proposed to extract the leaf edge of overlapping leaf in plant image.The plant images with different illumination intensity were collected,and leaf edge extraction from plant image based on leaf edge extraction algorithm proposed in this paper,and the leaf edges of overlapping leaves in plant images were obtained.Finally,based on the different illumination intensity of the leaf edge extraction algorithm for the vegetation community image leaf edge extraction.The leaf edges of overlapping parts in vegetation community images were successfully obtained,which provides a theoretical basis for the accurate identification of different plant vegetation communities.