<![CDATA[The collection of wood datasets is carried out in a cross-sectional position to record images of the shape, size, anddistribution of wood vessels and fibers. The image is taken using a smartphone camera which is equipped with anadditional 60 times magnification loop. The condition of taking pictures with non-ideal means and equipment makesit difficult to get good quality and consistent images, so that the texture of pores and wood fibers is not maximallyexposed, this will cause the quality of the dataset to decrease. The decline in the quality of the dataset will result in adecrease in classification accuracy. Therefore in this paper, research is conducted to improve the quality of theimage, by strengthening the boundary edges of the objects in the image. The edge boundary strengthening method isperformed using kernel convolution with edge enhancement filter matrix, then performance testing is performedusing two types of pattern extraction algorithms namely LBP and HOG with SVM pattern classification algorithm.The highest results achieved with the edge boundary reinforcement method found an increase in classificationaccuracy of +7.14% compared to before.]]>
Copyrights © 2019
