Authors

Nour Bahaa, Mai Mahmoud, Ibrahim Fawzy, Abanoub George, Supervised by: Dr. Khaled Hussein Eng. Nour El-Huda Ashraf

Publishing Date

December 29, 2020.

Abstract

Wheat is an important factor in every human being life’s. Due to that, studying wheat production is a major thing. However, the existence of Wild Oats and Rust in wheat reduces its production by 93% for every square meter. The spread of Wild Oats seeds destroys both the field and the land, for many more years according to the land condition. Moreover, the Rust disease appear due to the change of the weather. Therefore, if the farmer didn’t recognize it as soon as possible, it will damage and reduce the field. For this reason, we will use image processing and deep learning to decrease their appearance.

1.1 Purpose of this document

The purpose of the Software requirement specification (SRS) document is to give a detailed description of the importance of detecting the impurities’ in wheat. Moreover, the document defines how our stakeholder, team, and audience see the product and its functionality. In this document we will illustrate how the system will work with a better understanding, User scenario, and market need. This document will fulfill the detailed information for every input, output, algorithms used in each stage. Along with the Web application, interfaces, hardware, software requirements, and development process. The main purpose of this system is to automatically detect Wild Oats in the early stages and predict the percentage of the rust in Wheat,

1.2 Scope of this document

This document is targeting the farm owners and agriculture experts ,example: Agriculture Research Center , who maybe interested in our web application and cameras, as it will detect if their wheat crop is healthy or not in its stage whether there is Wild Oats grown in their field or not. Also, it will give them the percentage of the rust according to the wind condition in Wheat crop. As a result, for this detection, it will decrease the harm infection of the Wheat crop. In addition, this will help them to save time, effort, and

money.

1.3 System Overview

First thing, the main machine will collect all the knowledge from the input images(size, date and types), where these input images have different specifications. Then some pre-processing will occur by our system to normalize these images it will have same size and dimensions and have full RGB channels. This phase goes to assist us within the main processing phase once we must use these input images. The main processes stage is where we’ve our data in a very normalized sequence, so we are able to now begin to run the training phase of our model to start feeding the algorithm with the pictures. Then comes our testing phase(Validation phase), where our model is prepared to be evaluated and tested by the pictures. We used a layer or shape of Neural Networks which is that the Mask YOLOv5 deep learning approach to simply & faster detection of the wild oat grass & the rusty wheat from the healthy wheat plants within the sector.

1.4 System Scope

Our Scope is to decrease the number of Wild Oats and rust in Wheat and increase the crop production. Producing it by the best quality and in larger crop weight output. Above all, it will increase the farmer income and reduce the work and the time he takes in the field to recognize the Wild Oats. Besides, the money he pays for using the expensive high-quality chemicals to control Wild Oats.