Recognition of plants by Leaf Image using Moment Invariant and Texture Analysis

This paper presents a simple and computationally good method for plant species recognition using leaf images. Recognition of plant images is one of the research topics of computer vision. The use of shape for recognizing objects has been actively studied since the beginning of object recognition in 1950s. Several authors suggest that object shape is more informative than its appearance properties such as texture and color vary between object instances more than the shape. Initially we have scanned leaf images which are two dimensional in nature and segmented the images by mathematical morphological segmentation and then extracted the high frequency feature of image. For removing the noise, the image has been converted into binary, than complemented and multiplied by filtered image. We quantitatively establish the use of texture for detection various leaf images of same tree that are difficult by other classical methods of image processing. Further we use Nearest Neighborhood classification method to classify plant leaf. In this paper we focuses mainly on image enhancement, image segmentation, high frequency feature extraction, noise remove from background, volume fraction, inverse difference moment, moment invariant and morphological feature such as area convexity.