Pinewood Knot Detection Method Using Color Analysis

Since ancient times the pinewood is the most used construction material for living houses and bridges. The material density for fresh cut tree is 830 kg/m3 and it decreases to 460 kg/m3 for dry tree and it can be found in dedicated stores for construction material as timber. Because pinewood is a low weight material with bending resistance, for wooden houses with ground floor or ground floor and attic height regime, the weight loads are distributed through beams (wood structural components). In this work we propose to detect pinewood knots from an images database. The contributions presented in this paper are as follows: mathematical morphology operators (MMO) applied to color images (particularly dilation and erosion) followed by a RGB value filter (RGBVF). The MMO removes the pixels with weights lower than the structuring element defined in the proposed paper. RGBVF highlights the presence of knots if are detected. Experimental results on several sets of images with regards to proposed method are shown and discussed. The global implication: improvements in construction buildings' safety and financial advantages for pinewood buyers.