บทคัดย่อ

       This research aims to develop an asparagus grader using Artificial Intelligence (AI) technology combined with image processing and a robotic arm to increase the accuracy of asparagus grade sorting and reduce the reliance on human labor in traditional sorting. The system can examine asparagus characteristics, such as stem size and shoot characteristics, to classify grades accurately.

       This research uses 5,927 asparagus image data to train the YOLOv8 (You Only Look Once Version 8) model, which is a high-speed and high-accuracy object detection model. It starts by collecting and preparing images of asparagus of each grade, then labels them to train the AI model, dividing the data into 70% training data and 30% performance testing data. The system's hardware consists of a Raspberry Pi 5, a Raspberry Pi Camera Module 3, an ESP32 board, a Servo Motor, and a robotic arm. In addition, the system is designed to connect to a mobile application developed with Flutter. Users can view the sorting data results in real time.

       The results of the real-time system experiment with 60 asparagus samples showed that the system can sort grades correctly at 85.00 percent, which shows that it can sort asparagus grades as intended.