Integration of image processing algorithm and deep learning approaches to monitor ginger plant

Tan, Cheng Yong (2024) Integration of image processing algorithm and deep learning approaches to monitor ginger plant. Final Year Project, UTAR.

Preview

PDF Download (4Mb) | Preview

Abstract

This study aims to integrate image processing and deep learning algorithms to monitor the growth of ginger plants. The proposed system is designed to detect ginger plants and track their growth rate effectively. The deep learning algorithm will undergo training using a dataset containing ginger plant images, which will allow it to accurately identify and categorize various stages of growth. The image processing techniques will be used to pre-process and enhance the quality of the images to making it easier for the deep learning model to identify the ginger plants. One YOLOv8 based model was developed for detecting and segmenting ginger plants in various growth states. Following the successful detection and segmentation of the plants, another YOLOv8 based model was further developed to segment individual leaves from detected plant. In order to improve the monitoring process, a depth estimation model was used to calculate the distance from the camera to the plants, enabling measurements of the height and leaf area of the ginger plants. The integration of these two methods will provide a more efficient and reliable way to monitor ginger plant growth, which is important for farmers and researchers in the field of agriculture.

Actions (login required)