Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Deep Learning Based Classification of Papaya Disease Recognition.

Md. Sagar Hossen
Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Imdadul Haque
Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Md. Saif Islam
Dept. of Software Engineering Daffodil International University Dhaka, Bangladesh

Md. Tanvir Ahmed
Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Md. Jannati Nime
Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Md. Ashiqul Islam
Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Abstract/ Executive Summary:

Every year many of farmers had to face a huge amount of loss due to the papaya disease. From the last several year's thousands of researchers pay attention to the papaya disease recognition system to reduce the losses of farmers. Farmers have no prior knowledge about the detection technique to mitigate the disease. They found the diseases when the papaya already affected and wasted. Due to the loss of cultivation, many of them are now frightening to take steps about planting papaya. On this matter, we have performed a research by the advancement of deep learning technology to detect and classify the papaya disease. We have used CNN model according to Keras API. This model reliable with fully connected where classification is completed and all the process is a deep learning based. It took the fixed size of image 200x200 RGB image as input.

Keywords: Papaya Diseases, Deep Learning, Keras, Classification

Location: Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Start Date:

End Date:

Program Area: Pest Management

Commodity: Papaya

Objectives:

Methodology:

In this paper, we have performed research for recognition of papaya disease following the Deep Learning based approach. Our model is implemented by Convolution Neural Network (CNN) with Keras API which directly takes the papaya images and then fit the model for checking the accuracy and finally predict the disease affected and fresh papaya. We mainly classify disease to take RGB image where we fixed the input shape as 200x200. Now in hidden layer, the learnable convolutional layer first provides output shape as (198,198,16) and not learnable layer MaxPool just decrease the output shape by divide 2 and new output shape is (99,99,16). In the same rule, it is used to learnable layer and also MaxPooling layer reduce by half. This is the equation of the output changes in CNN model. Where nH is “size of height” and nW is “size of weight” for input image data and it is fixed for our model. Padding is the relevant term to CNN (Convolutional Neural Network) which refers the amount of pixel on input image in CNN model and in proposed model, padding is zero means the added pixel value is zero. Kernel shape (f) is known as the batch size which is 3. Now Stride defines number of pixels which shift over the input matrix and we have use 1 as stride that’s mean it can filters to 1 pixel at a single time. After the Hidden Layer, it is main part to classify the image both fresh papaya nor disease affected papaya and it is known as “Classification” part. For learnable layer, it changes the parameters and below the calculations and also rule is given, Rules: {(Previous channel * Filter shape) + 1}*Current channel.

Source of Information: Dept. of Computer Science and Engineering Daffodil International University Dhaka, Bangladesh

Article Link (Online Journal):

Funding Source:

Budget:

Total Budget (Tk.) :

Achievement/Findings:

Papaya disease classification and prediction techniques using the Convolutional Neural Network. Our model has performed relatively better compared to other’s provided model and the accuracy is 91% which is too better. The main purpose of presenting this model is for the village farmers and these farmers who are deprived of suggestions from the agriculture office. So that they can diagnose papaya disease at the right time and get the expected yield with proper care. In addition, using this model will save the farmer from wasting his time as well as it is able to accurately diagnose papaya is affected or not. The main part of our research is to help those farmer’s so that they can easily get their papayas condition details through our model and take necessary steps.

Knowledge Management: Journal