Identifikasi Kualitas Kesegaran Susu Sapi Melalui Pengolahan Citra Digital Menggunakan Metode Watershed Dan Klasifikasi Learning Vector Quantization (lvq)

Mohamad Fikri Permana, Bambang Hidayat, Sjafril Darana

Abstract

Abstrak Tingginya minat masyarakat terhadap susu, membuat para produsen melakukan inovasi agar mendapatkan keuntungan yang lebih dengan menurunkan kualitas asli dari susu sapi. Misalnya menambahkan bahan tambahan yang dapat merusak nilai gizi dari susu dengan dicampur air agar volumenya lebih banyak. Cara umum dalam membedakan kualitas susu, yaitu dari aroma dan rasa, namun hal tersebut tidak efektif karena indra perasa setiap orang dapat berbeda. Pada era masa kini diperlukan teknologi yang dapat membedakan susu murni dengan susu yang sudah dicampur bahan lain. Penelitian dilakukan dengan mengidentifikasi kualitas kesegaran susu sapi melalui pengolahan citra digital menggunakan metode Watershed, dimana proses ekstraksi ciri menggunakan Local Binary Pattern serta diklasifikasikan menggunakan Learning Vector Quantization. Sistem tersebut telah diaplikasikan melalui penggunaan perangkat lunak Matlab dengan mengidentifikasi dan mengklasifikasikannya pada tekstur susu sapi. Pengambilan data dilakukan dengan cara mengambil sampel susu sapi murni dan sampel susu sapi yang dicampur air sebanyak 25%, 50%, dan 75%. Hasil penelitian identifikasi kualitas kesegaran susu diperoleh tingkat akurasi sebesar 92.5% dan waktu komputasi 0.4791 detik. Kata kunci: Susu Sapi, Watershed, Learning Vector Quantization Abstract The high public interest of milk which give many benefits for human body, urge the producers for doing an innovation to get more profit by lowered the quality of the milk adding an additional substance to the milk can impair the quality of the milk itself. A general way to distinguish whether the quality of the milk is good or not is from its aroma and taste. On the other hand, this kind of way is not effective. In this era of technology, it is needed a kind of technology which can distinguish whether the milk is pure or not. Research done by identifying the quality of fresh cow's milk through digital image processing using Watershed method, where the extraction process characteristics using Local Binary Pattern and classified using Learning Vector Quantization. The system has been applied through the use of Matlab software by identifying and classifying on texture of cow's milk. Data retrieval is done by taking a sample of pure cow's milk and cow's milk samples were mixed in the water as much as 25%, 50%, and 75%. The research of identification the quality freshness obtained accuracy of 92.5% and computational time 0.4791 seconds. Keywords: Cow’s milk, Watershed, Learning Vector Quantization

Full Text:

PDF

Refbacks

  • There are currently no refbacks.
max_upload :0