Institutional-Repository, University of Moratuwa.  

Detection of affected area, pests and classification of pests using convolutional neural networks from the leaf images

Show simple item record

dc.contributor.advisor Premaratne S
dc.contributor.author Suthakaran A
dc.date.accessioned 2019
dc.date.available 2019
dc.date.issued 2019
dc.identifier.uri http://dl.lib.mrt.ac.lk/handle/123/15966
dc.description.abstract Pest infection is the most important problem on vegetable plants. One way to control the pest infection is to use proper pesticides. Early detection of the pest or the initial presence of pests is a key element for crop protection. The identification of the pest was done manually at the beginning. This takes time and also requires ongoing monitoring of experts. An automatic pest detection system is needed to examine the infestation and classify the type of pest. Today, there are many techniques and methods for identifying pests and detecting plant diseases. In these techniques, image processing techniques are very efficient and reliable. First, the proposed model detects whether the leaf is affected or not and calculates the affected area in the image. Next, the region of the detected pest and classification were performed using convolutional neural networks. The severity of the infection can be observed by calculating the percentage of the affected area, which leads to taking the appropriate measures. en_US
dc.language.iso en en_US
dc.subject INFORMATION TECHNOLOGY-Dissertations en_US
dc.subject IMAGE PROCESSING-Applications en_US
dc.subject PESTS-Management en_US
dc.subject NEURAL NETWORKS en_US
dc.title Detection of affected area, pests and classification of pests using convolutional neural networks from the leaf images en_US
dc.type Thesis-Abstract en_US
dc.identifier.faculty IT en_US
dc.identifier.degree MSc in Information Technology en_US
dc.identifier.department Department of Information Technology en_US
dc.date.accept 2019
dc.identifier.accno TH3899 en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record