Precision-Agriculture: An Image Net – Based Multilayer Convolution Neural Network for Leaf Disease Detection in Coffee Plant in Early-Stage System
Secured Plants often are affected by various fungal infections and leaf diseases that hinders the growth of healthy crop and eventually reducing yield in plants, lead, buds, flowers and fruits, Plant diseases are the leading source of agricultural production losses in terms of quantity and quality. These reductions have a severe influence on agricultural production. Farmers and plant pathologist have always relied on their eyes to diagnose diseases and make decisions based on experiences, which necessitates a significant amount of time and human work., It is sometimes inaccurate and sometimes prejudiced because many diseases appear to be much like the early stages, proper results may not be recorded, and accuracy may be low at times [1-4]. This technique leads to the usage of pesticides that aren't necessary, resulting in greater production costs. According for that knowledge, a precise disease detector linked to a trustworthy database to assist farmers is required, particularly in the presence of naive and inexperienced farmers [5-6]. Rapid development of computer vision open the way for this using province Deep learning (DL) and machine learning (ML) techniques. An early disease detector is also required to protect your crop in a timely manner. There have been numerous previous studies undertaken for this aim . With CNN's most popular model, we used the "Plant Village" dataset, a well-known database that really is available on the internet. The CNNs, on the other hand, require a big amount of information for training. In this paper, we offer the following strategy: CNN models have been improved with ImageNet algorithm learning, which employs an Artificial Neural Network (ANN) with Feature Selection (FS) to address multi-class classification problems for four types of diseases, namely Cercospora, Miner, Phoma, and Rust The proposed frame work aim at increasing the models’ accuracy when the data is limited
How to Cite
Copyright (c) 2021 Journal of Research Proceedings
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.