Sistem Klasifikasi Kualitas Pupuk Berdasarkan Fitur Warna HSV Menggunakan Metode K-Nearest Neighbors

Naranakubar, Rizki Bimo Alvrido (2026) Sistem Klasifikasi Kualitas Pupuk Berdasarkan Fitur Warna HSV Menggunakan Metode K-Nearest Neighbors. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221079-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

PT Petrokimia Gresik masih melaksanakan inspeksi warna produk pupuk secara visual oleh operator pada lini pengantongan. Proses inspeksi dilakukan secara berkala pada awal shift, sehingga perubahan warna produk selama proses produksi berpotensi tidak terdeteksi. Kondisi tersebut dapat menyebabkan produk di luar spesifikasi (off-spec) lolos hingga ke tahap distribusi. Penelitian ini bertujuan merancang dan mengimplementasikan sistem inspeksi warna produk pupuk secara real-time berbasis Computer Vision menggunakan metode K-Nearest Neighbors (KNN). Sistem dikembangkan melalui tahapan akuisisi citra menggunakan IP Camera, cropping berdasarkan Region of Interest (ROI), ekstraksi fitur warna HSV (Hue, Saturation, dan Value), serta klasifikasi menggunakan metode KNN. Penentuan parameter dilakukan melalui pengujian nilai K ganjil dari 3 hingga 21 menggunakan Stratified 5-Fold Cross Validation. Hasil pengujian menunjukkan bahwa parameter terbaik diperoleh pada nilai K=5 dengan rata-rata akurasi 99,36%, sedangkan nilai precision, recall, dan F1-score berada pada rentang 0,984–0,999. Pengujian implementasi menghasilkan rata-rata kecepatan pemrosesan sebesar 24,543 frame per second (FPS) dengan rata-rata delay 0,74 ms/frame atau 18,5 ms/detik. Selain itu, warning system aktif ketika sistem mengklasifikasikan produk sebagai Produk Tidak Sesuai, sedangkan pada kelas Subsidi, Non-subsidi, dan Sedang Tidak Ada Produk, sistem hanya menampilkan hasil klasifikasi tanpa mengaktifkan peringatan. Hasil tersebut menunjukkan bahwa mekanisme deteksi dan pemberian peringatan berjalan sesuai dengan rancangan sistem.
=====================================================================================================================================
At PT Petrokimia Gresik, fertilizer color inspection on the bagging line is still performed visually by operators. The inspection is conducted periodically at the beginning of each shift, allowing color changes that occur during the production process to go undetected. As a result, off-spec products may pass the inspection process and proceed to the distribution stage. This study aims to design and implement a real-time fertilizer color inspection system based on Computer Vision using the K-Nearest Neighbors (KNN) algorithm. The proposed system consists of image acquisition using an IP Camera, cropping based on the Region of Interest (ROI), color feature extraction in the HSV (Hue, Saturation, and Value) color space, and classification using the KNN algorithm. The optimal parameter was determined by evaluating odd K values ranging from 3 to 21 using Stratified 5-Fold Cross Validation. The experimental results showed that the best performance was achieved at K=5 with an average accuracy of 99.36%, while the precision, recall, and F1-score values ranged from 0.984 to 0.999. The implementation test achieved an average processing speed of 24.543 frames per second (FPS) with an average delay of 0.74 ms/frame or 18.5 ms/second. In addition, the warning system was activated when the system classified a product as Non-Conforming Product, whereas for the Subsidized Fertilizer, Non-Subsidized Fertilizer, and No Product Detected classes, the system displayed only the classification result without triggering a warning. These results indicate that the detection and warning mechanisms operated in accordance with the designed system.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Computer Vision, HSV, K-Nearest Neighbors, inspeksi warna, pupuk, Computer Vision, HSV, K-Nearest Neighbors, color inspection, fertilizer.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. S Agriculture > S Agriculture (General) > S633.5 Fertilizers T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. T Technology > TS Manufactures > TS156 Quality Control. QFD. Taguchi methods (Quality control)
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Rizki Bimo Alvrido Naranakubar
Date Deposited:	10 Aug 2026 02:34
Last Modified:	10 Aug 2026 02:34
URI:	http://repository.its.ac.id/id/eprint/144228

Actions (login required)

View Item