Semi-Automated Weeds Identification and Watering System Using Machine Learning: Based on Cabbage Crops
No Thumbnail Available
Date
2021
Journal Title
Journal ISSN
Volume Title
Publisher
Uva Wellassa University of Sri Lanka
Abstract
Fast-growing novel technologies influencing the world's different industries with innovative
solutions for the issues arising to simplify the human workload and to avoid time consumption. In the
Sri Lankan context usage of novel technology in the agricultural industry is not at a satisfactory level
and weeds can be identified as a major problem agricultural field. Manual weed identification and
removal methods are still using and it needs a lot of human workload and time. This Semi-
Automated system was developed for plant and weeds identification and watering the crops using
machine vision and machine learning technologies. Raspberry pi model B+ was used as the
controller with camera and python with OpenCV is used for programming. This system was
developed based on the cabbage crops and Bayesian classification and Mahalanobis distance for non-
diagonal classification was used for recognition of plants. The plant watering mechanism works on
the identification of plants this is an autonomous mechanism. This robotic system can be controlled
through Bluetooth and it is able to ride through different field conditions. 85% of accuracy in the
identification of plants and weeds is achieved in conducted trial runs using this system.
Implementation of the system will enhance crop cultivation and cost, time reduction of the farmers in
Sri Lanka. The system is successful in the trial runs and it will develop with the robotic arm for
weeds removal in the future with this identification. Results manifest that farmer can use this system
for their crop cultivation as an effective method in Sri Lanka.
Key words: Artificial Intelligence; Machine learning technologies; Watering; Plantation; Weeds
Management
Description
Keywords
Engineering Technology, Engineering, Mechanical Engineering, Weeds Management, Crop Production