Estudo e implementação de um classificador de tamanho útil de madeira serrada através de técnicas de processamento digital de imagens.

Abstract

The problem of classification of the useful size of wood is a long time challenge in the wood frame industry, and consists basically in finding the right area to be used for the manufacturing. The problem is easy to understand, but very difficult to be solved. Many techniques were experienced by companies, many were abandoned, many were partially utilized, and most used a computer vision and human combination. This paper proposes the use of computer vision for the classification of the useful area through the images capture and image processing of pieces that receive projections of lasers at an angle of 45% on them. Many techniques were utilized in this industrial automation, in this techniques the math morphology. This technique ensures that the spot where the lasers are aimed can change its direction, or even disappear, in the event of any failure in the piece, and thus measuring the useful size each of the products. The proposed system had experienced better results than the currently systems used by frame manufacturing industry, and can prove the process optimization itself through simulation and benchmarking of the two techniques.