Tugas Akhir
Rancang bangun alat uji kekasaran permukaan bahan berbasis pengolahan citra
Proses grinding merupakan tahapan awal dalam uji metalografi untuk menghasilkan sampel yang rata dan halus. Penentuan kekasaran sampel secara visual membuat waktu preparasi sampel cukup lama. Untuk itu, dibuat alat uji kekasaran permukaan bahan berbasis pengolahan citra untuk mengukur tingkat kekasaran sampel. Citra diambil dari sampel hasil grinding dengan melalui proses akuisisi citra. Citra digital sampel melewati tahapan preprocessing, ekstraksi ciri metode statistical texture dan Grey Level Co-occurrence Matrices (GLCM), klasifikasi jaringan syaraf tiruan dan ditampilkan dalam bentuk Graphical User Inteface (GUI). Hasil pengujian akurasi didapatkan nilai akurasi rata-rata pengujian untuk data hasil ekstraksi ciri metode statistical texture sebesar 81.25% dan nilai akurasi rata-rata pengujian untuk data hasil ekstraksi ciri metode GLCM sebesar 87.50%. Berdasarkan pengujian tersebut, alat ini dapat digunakan dengan baik untuk membedakan tingkat kekasaran pembukaan bahan dengan model jaringan syaraf tiruan metode ekstraksi ciri GLCM. rnKata kunci : Kekasaran permukaan, Pengolahan citra, Statistical texture, GLCM, Jaringan syaraf tiruan rnrnABSTRACT rnThe grinding process is an early stage in a metallographic test to produce a flat, smooth sample. The determination of sample roughness visually makes sample preparation time long enough. To that end, a roughness-based material surfacebased test material was developed to measure the roughness of the sample. The image is taken from the grinding sample through the image acquisition process. The digital image of the sample passes through the preprocessing stage, extraction features statistical texture method and Gray Level Co-occurrence Matrices (GLCM), artificial neural network classification and displayed in the form of Graphical User Inteface (GUI). The result of accuracy test is got the average accuracy value of test for the extraction data characteristic of statistical texture method equal to 81.25% and the average accuracy value of test for data extracted feature of GLCM method is 87.50%. Based on the test, this tool can be used well to distinguish the level of cracking openness of the material with the artificial neural network model GLCM characteristic extraction method. rnKeywords: Surface roughness, Image processing, Statistical texture, GLCM, Artificial neural network
S18-0182 | 25/TA/K/18 681.176(043) SIT r | Perpustakaan Poltek Nuklir | Tersedia |
Tidak tersedia versi lain