IoT Based Smart Palm Oil Seed Segregator using RGB Color Sensor

Abstract

The manual seed grading process through human vision is tedious and time-consuming due to the tendency for errors and inconsistencies. Shell residues also failed to be properly segregated and required a huge amount of human energy as they were unable to be detected automatically once the waste container was full. The main objectives of this research are to develop an automated palm seed grading and sorting system to increase the seed’s sorting quality. The system can segregate palm oil and shell residues automatically to ease kernel recycling in the future. In addition, the system can implement the waste bins’ level detection, which can notify the user via the Blynk application, reducing the time and manpower required. To develop this system, sensor comparisons are done to determine the best-performing sensor to be used in the operation. The result shows that the RGB color sensor is the best color detecting sensor with an increment accuracy of 30.525%. As for the smart bin for shell residue, system response became 62.96% faster with the use of an ultrasonic sensor. The RGB sensor detects seed with readings <165 color concentration as freshly ripe and readings >165 color concentration as overripe. The janitor can be notified in real-time when the bins for oil and shell residues are full through the Blynk application using the WiFi module (ESP8266). The system has 100% accuracy, which was tested using the confusion matrix formula for both seed categories.