Pick Up Choose
Motivation
Continuing previous research (Paper published in TAAI'21), this project is my graduation project, which used Machine Learning techniques to solve the agricultural workforce scarcity problem that severly affects Taiwan’s agriculture industry. My group member and I discovered that the advancement in cities provides more job opportunities and higher salaries that attract young people to move from the countryside to cities. As a result, agricultural labour has gradually decreased in recent years, posing a massive crisis to Taiwan’s agricultural industry. Therefore, we focused on the issue of labour shortages when classifying date grades in the harvest season.
Approach
I successfully extracted dates and injured parts from images by training two layers of the YOLOv4 model. Then, I divided dates into different grades with the EfficientNet that was trained with multiple data such as injured area percentage and ingury cause weight.
Deployment
Considering the practical situation, agricultural workers cannot afford expensive equipment. Thus, I used Flutter to design a smartphone application that allows people to use our service conveniently. We presented our result as a domestic conference paper and received many positive feedbacks. Besides, our prominent performance and outstanding work earned us 2nd prize in the Graduation Project Competition.
Resource
To download the project and know more details about the project, welcome to visit my Github.
