CD Skripsi
Rangkaian Interface Arduino Dan Matlab Untuk Sistem Penyortiran Otomatis Tandan Buah Segar Kelapa Sawit Menggunakan Jaringan Syaraf Tiruan (Jst)
An automated sorting system using computer vision method induced by artificial neural network (ANN) has been made to classify the ripeness levels of fresh fruit bunches (FBBs) oil palm. This sorting system consists of grading system prototype, arduino UNO microcontroller, an amscope 3.2 megapixel RGB camera, and MATLAB software. In this research the interfacing of Arduino UNO board and MATLAB programs and the use of artificial neural network (JST) have been completed. The samples of FFB used were varietes of tenera clones marihat, topas, and lonsum. Traditionally, defined ripeness levels are F0 and F1 (unripe), F2 and F3 (ripe), F4 and F5 (over ripe). These ripeness levels were used as a database of artificial neural network (JST) and were also tested on the automatic sorting system. The images of FBB samples were recorded by the camera and was extracted into RGB and HSV matrix values. Matrix information was used to create the artificial neural networks (ANN). Program artificial neural networks sorting system was made into two programs namely, RGB and HSV JST. The results of the automatic sorting system showed that RGB JST program successfully sorted 79 of 216 FBB or equal to 36.57% and HSV JST managed to sort 106 of 216 FBB or equal to 49.07%. The HSV JST more accurate than RGB JST because information more presented by the system perception and therefore HSV can be used in color matching and comparison. The results of the interfacing arduino UNO board and MATLAB programs showed that the programs have succeeded to control the AC motor, and the sensors present in the prototype grading system.
Keywords: Computer Vision, RGB, HSV, Artificial Neural Network, Arduino, MATLAB
Tidak tersedia versi lain