Single Leaf Image Superresoution using NEDI Technique

Image Processing

The concept of Superresolution is to increase the input image resolution. The paper focuses to study the two different iterative curvature based methods on single diseased leaf image having low resolution. For agricultural countries like India, Crop plays a vital role in country’s economic growth. Plants gets viral, bacterial or fungal diseases which have a significant reduction quantity as well as quality of agricultural products. The traditional approach of expert’s naked eye observation is time consuming. For this leaf identification system, leaf diseases detection system, plant diseases diagnosis system are developed which demands high resolution leaf images as input for better recognition rate. The cheapest solution is to use superresolution technique which is related in both with the statistical relationship between high resolution output and low resolution input images and with the human perception of image quality. However superresolution algorithms are being affected by artifacts such as over smoothed, jaggies, blurred or over sharped. Fast Curvature Based Interpolation (FCBI) Technique was proposed for this but results are not satisfactory. The paper have described ICBI: Iterative Curvature Based Interpolation combined with NEDI: New Edge Detection Interpolation which gives superresolved image for a single leaf image. Fine edges in SR images are preserved without applying complex mathematical algorithms based on wavelet, fast curvelet, etc. This concept can be useful for agricultural expert to help farmers for exact leaf disease detection and accurate remedial actions. The experimental result shows the best visible SR result of an infected leaf along with Mean Square Root (MSE) and Peak Signal to Noise Ratio (PSNR) statistical results.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.