A study of Feature Extraction of leaf shape and Texture Quality Surface | Original Article

ABSTRACT:

The aim of this paper is to the field of leaf Disorder recognition for plant identification has experienced an increased need for fast and efficient classification algorithms to aid in keeping track of the most precious plants on earth. This requirement resulted in a number of techniques revolutionizing the automatic classification area. The increasing number of techniques has led to a dilemma in deciding which of these methods have the best qualities and potential to efficiently classify. In the botanical industry where the information distortion can produce inaccurate diagnosis, this problem is particularly important. Thus, the urgent necessity of the botanical field is automated tools that help identify factories. The main process of the machine's training includes building a leaf database, image enhancement, segmentation (leaf extraction) features, extraction and classification. Leaf disorder recognition (CAPLDR) consists of four stages. This research aims primarily at proposing techniques for improving every plant identification operation through leaf disease. A system for improving the leaf image was proposed, called 'Enhanced wavelet-based demoizing with built-in edge enhancement and automatic contrast adjustment algorithm. This 197 method combines the wavelets, CLAHE (contrast adjustment), corner enhances and a relaxed middle filter (noise removal), with a single procedure to increase the visual quality of the leaf image.