Enhancing The Efficiency of Agriculture by using Image Processing and Drones

Abstract

Agriculture is an important factor for the development of any country, in addition to providing foodstuffs, agriculture is a primary source of raw materials that are used in several industries, a term known as smart agriculture has recently appeared where technology and modern techniques are used to better plan and manage crops. The research focuses on discovering one of the most dangerous cotton diseases, angular spot disease. which is also known as bacterial blight. It can cause production losses of up to 10% of the crop. The drone is used to take pictures of the agricultural field. It is a mechanical vehicle with four arms, and in each arm, a motor is connected to a propeller. Two of the propellers rotate clockwise, while the other two spin counterclockwise. Based on the images taken by the drone, the technique of filtering neural networks is convolutional Neural Network (CNN), which is used to detect the disease by performing operations on the images. This research contributed to helping to enhance the efficiency of cotton production by classifying the images taken by the drones into healthy images or bacterial blight diseases, training a CNN model using the data set that contains images of diseased and healthy cotton and we obtained an accuracy of 97.2% and thus is successfully classify the cotton images into diseased and healthy and return them to a map showing disease prevalence.