A visual information assessment tool for resin canal identification and property measurement

Abstract

Resin canal studies rely on the accurate measurement of various resin canal properties. The conventional approach of manually marking resin canals on a wood sample is not only labour intensive but, more importantly, unable to quantify the visual information of resin canals, such as the size and contrast of a resin canal. This paper describes a semi-automatic visual information processing approach that was developed to identify resin canals and growth rings from colour-scanned images of the transverse surface of polished wood samples, and to quantify the characteristics of resin canals by ring-based statistical measures in terms of the area of a resin canal occupied, colour contrast of a resin canal to its surrounding material and resin canals occurrence frequency within a growth ring. The proposed approach endeavours to automate the analysis as far as practicable. Interactive facilities are provided for correcting decisions made by the automated analysis, wherever necessary. The method of resin canal detection, based on image processing techniques, provides an effective solution to many problems that require feature identification and assessment. The proposed approach was used to assess variation in axial resin canals in radiata pine across different sites.