AI-image processing and image recognition for intelligent prawn farming
This project aims to explore the untapped potential of Convolutional Neural Networks (CNNs) within the realm of Prawn Farming. In the domain of deep learning, CNNs are recognized as a set of neural networks primarily designed for processing spatial data, such as images and videos. This research proj...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2023
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/6035/1/fyp_CS_2023_KJC.pdf http://eprints.utar.edu.my/6035/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This project aims to explore the untapped potential of Convolutional Neural Networks (CNNs) within the realm of Prawn Farming. In the domain of deep learning, CNNs are recognized as a set of neural networks primarily designed for processing spatial data, such as images and videos. This research project is dedicated to applying various types of CNNs to detect the growth stages of the Giant Freshwater Prawn and find out the most suitable. Simultaneously, it endeavours to uncover the broader contributions that CNN models can make to the Prawn Farming industry and the broader aquaculture ecosystem.
The CNN models employed in this research include You Only Look Once (YOLOv7), Faster-RCNN ResNet101, SSD ResNet101, Centernet Hourglass 104, SSD MobileNet V1, and Faster-RCNN ResNet50 v1. The research process involves an extensive literature review and in-depth research to gain a comprehensive understanding of the application of Machine Learning in Prawn Farming and the broader aquaculture system. Through this review, it becomes evident that, in contrast to traditional Artificial Intelligence (AI) models, CNNs have emerged as a prominent trend in image processing and recognition. These insights underscore the rationale for conducting this project.
Furthermore, data acquisition stands as a pivotal aspect of this research project due to the unavailability of the required image data from existing sources. Consequently, a dedicated dataset has been meticulously curated and made accessible on Kaggle for use by fellow researchers in the future. The outcomes of this research project offer valuable insights into the selection of suitable CNN models for Giant Freshwater Prawn growth stage classification, believing will be able to provide guidance for future new entrants into this field. |
|---|