Indentifikasi Jenis Kayu Berdasarkan Citra Digital Menggunakan Algoritma Eigenimage Dan Principal Components Analysis
Wahyu Widodo dan Edi Faizal 29 September 2014 Teknik Informatika

Abstract / Intisari :
Indonesia is a tropical country which has about 4,000 species of trees. Potential tree estimated 400 types of botanical (species), covered in 198 genera and 68 tribes. General characteristics of wood can be recognized directly by the five senses by color, pattern, texture, grain direction, luster, impressions touch, smell and hardness of wood. While anatomical features include the composition, shape, and size of the cells or tissue intruders, which can only be clearly observed using the aid of a magnifying glass or microscope like loops. Along with the development of computerized technology, pattern recognition have been carried out with a variety of applications and algorithms. In this research wood species recognition systems developed using the imagery pore structure of the wood. The algorithm used is Eigenimage and transformation of pattern recognition used by Principal Components Analysis (PCA). Identification performed by pattern matching. The test results showed the percentage of the system's ability to correctly identify the wood image by 96% (sensitivity), percentage of the system's specificity of 59%, positive predictive value of 88% (PPV), negative predictive value by 83% (NPV), and accuracy rate of 87% (accuracy) with an error rate (error rate) of 13%. Keyword: wood, pattern recognition, eigenimage, PCA

File PDF :
   Wahyu Widodo dan Edi Faizal-27-stmikelrahma.pdf

Tag Jurnal :