Annals of Warsaw University of Life Sciences - SGGW Forestry and Wood Technology No 69, 2009: 433-436 (Ann. WULS - SGGW, For. and Wood Technol., 69, 2009) Assessment of wood surface roughness based on image histogram analysis JACEK WILKOWSKI, KACPER WOLIŃSKI Wood Mechanical Processing Department, Warsaw University of Life Sciences– SGGW Abstract: Assessment of wood surface roughness based on image histogram analysis. In work was undertaken the probe of surface roughness assessment for chosen species of wood (pine, birch, oak and alder) after sawing, on base of histogram image analysis. Standard model of colors RGB for received digital images was assumed. Standard deviation of pixels intensity distribution in image histograms for each components canal. Simultaneously, the contact measurement of wood surface roughness with usage of profilometer was carried out. Parameters obtained by means of contact method were correlated with indicators which characterize surface image. These indicators to a high degree reflect attributes of real wood surface after sawing. Key words: roughness parameters, images histogram, machine vision, wood sawing INTRODUCTION Machined surface determines very strong product quality. The contact methods are the most common way of surface roughness measurement, which belongs to the group of direct methods. This methods are perceived as standard which can be use to reproduction of the profile as well as to calculate his parameters [Nowicki 1991]. In spite of the fact, that this procedure was developed on very high level, the devices based on contact methods have some serious disadvantages and restrictions. The influence of needle roundness and his pressing which is the source of profile reproduction deformation, slow feed speed of needle which constrain speed of the whole measurement, sensitiveness on external vibrations and high costs belong to the most important [Zakrzewski, Staniszewska 2002]. Due to this faults scientists make efforts to find out ways of measurement based on indirect methods which would give satisfactory results in industry conditions. Among others man can distinguish vision systems. These procedure allows to received information which can be useful in quantitative and qualitative inspection of surface [Gadelmawla 2004, Myshkin et al. 2003]. The advantages of this solutions consist in this, that the measurement is realized in contact, remote and non destructive mode [Storch 2006, Zawada-Tomkiewicz 2007]. According to Demanta et al. [1999] the usage of vision methods during production process enable realizing remote and fast measurement. Thanks to this systems is possible to avoid a lot of problems which are connected with contact method. The aim of this work was surface roughness assessment for chosen species of wood (pine, birch, oak and alder) after sawing, based on histogram parameters analysis of surface image. The results of experiments should give answer on the question how strong pixels intensity distribution in image histograms of wood surface (exactly standard deviation of pixels intensity distribution in image histograms of each canal in RGB model) is correlated with roughness parameters obtained by means of profilometer (contact method). 177