WOOD BASED IMAGE ANALYSIS OF FIBER USING GRAY LEVEL CO-OCURENT MATRIX (GLCM)
Wood included a relatively high economic commodity. In Indonesia, there are quite a lot of these kinds of a timber. Perhutani Party needs to know information about the type of wood, so that at the time of cutting can be directly known from the type of wood to be processed. To provide good information about the type of wood, we need a system that can provide information on the type of wood. to the presentation type of wood 100% should have no errors. Because in forestry wood type is preferred, they are not allowed to run into the error to determine the type of wood because it will have an impact on those who would buy and sell the wood.
Based on the above background, it would require an analysis to determine the type of wood to be taken with a digital camera. Anaisis are carried out so that the image of wood fibers using greylevel co-ocurent matrix (GLCM) for feature extraction and dikasifikasikan using Backpropagation. The result showed that the results of the classification has been done by using two hidden layers which each layer contains 50 neurons and iteration 3000, the presentation of success is 100% for testing the image of wood fiber camphor, 60% for testing the image of wood fiber kruwing, 100% for testing image meranti fiber, 20% for testing the image of pine wood fiber, 80% for testing image kapok wood fiber.
Kata Kunci : Serat Kayu, GLCM, Ekstraksi Ciri, Backpropagation.
- There are currently no refbacks.