Healthy and Unhealthy Leaf Classification using Convolution Neural Network and CSLBP Features
Harmandeep Kour1, Lal Chand2
1Harmandeep Kour*, CSE, Punjabi University, Patiala, India.
2Lal Chand Panwar, Assistant Professor, CSE, Punjabi University, patiala, India.
Manuscript received on September 02, 2020. | Revised Manuscript received on September 15, 2020. | Manuscript published on October 30, 2020. | PP: 25-31 | Volume-10 Issue-1, October 2020. | Retrieval Number: 100.1/ijeat.F1634089620 | DOI: 10.35940/ijeat.F1634.1010120
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Once applied to real world images, most machine learning models for the automated identification of diseases have limited efficiency. Plant diseases cause major agricultural production and economic loss. These illnesses also show visible signs, including lines, streaks and shift in color, on leaf surfaces. Many researchers have recently researched the potential use of image treatment and computer processing in plants and leaves to diagnose disease. There is space for improved performance though several methods and computer procedures have been developed in this area of investigation. Several previous models only deal with a few morphological features of the diseased regions. A new method for detecting plant leave’s disease using the segmentation, and CNN approach based on GLCM and LPQ features of the Basil and Guava leaves feedback imagery has been established in the present paper. The findings revealed that the suggested model is as effective as possible, for both basil and guava leaves, to better distinguish healthy and unhealthy leaves. The overall accuracy of the Guava dataset is 97.1% and the basil dataset is 92.1%.
Keywords: Bilateral Filter, CNN, GLCM, Leaf Disease Classification, LPQ.
Scope of the Article: Classification