الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 111 |  |  |
| | | from | 111 | | |
Abstract
Agriculture is regarded as a key economic pillar in the world and provides for one of humankind’s fundamental needs, namely food. It is regarded as the main source of employment in most of the countries.
By offering detailed advice and insights about the crops, machine learning is a current technology that helps farmers reduce farming losses. The use of machine learning in agriculture enables more precise, efficient, and high-quality productive farming with less damage.
In this thesis, we present a design of an AI model for the early detection of existence of Red Palm Weevil (RPW) larvae in palm trees to avoid the complete devastation of the yield. The proposed methodology is based on recording the RPW sounds during different activity states then extracts the important features in the sounds by applying MFCC technique. The recorded sounds has been transformed to spectrogram and an ANN has been applied to automatically detect if RPW larvae exist or not. We achieved through the proposed strategy 99.2% accuracy using Alexnet.
The presented thesis consists of four main chapters which are summarized as follows:
Chapter 1: Introduction chapter in which the importance of the dates cultivation is introduced and the problem encountering this field is stated.
Chapter 2: Literature Survey chapter in which a survey is presented of different early detection techniques of the RPW infestation and different machine learning techniques.
Chapter3: Methodology chapter in which the proposed methodology is explained in details and the results are presented.
Chapter 4: Conclusion chapter in which the final conclusion of this thesis is presented and the possible implementation of the model in real field applications.