New anatomical method of grain angles measurement using confocal microscopy and image cross-correlation

Abstract

Some tropical trees with indistinct growth rings have a distinct interlocked grain that reveals their internal growth rhythm. To determine their growth rhythm, it is necessary to accurately measure the wood grain angle. The usual methods for grain angle measurement are radial splitting using wood disks, which occasionally provides inaccurate data, and serial tangential sectioning, which requires preparation and analysis of many sections. The present report proposes an easier but accurate method to measure grain angle using a single xylem transverse section. A confocal microscope was used to obtain two optical sections of different depths from a transverse section of a 7-year-old Hopea odorata Roxb. The tangential lag between the optical images was then calculated using image cross-correlation and transformed into grain angle. Radially consecutive sampling revealed distinct radial fluctuations in the grain angle. The fluctuation data were compared to data obtained by radial splitting and serial tangential sectioning. There was a strong correlation between grain angle using the three methods. In the region close to the cambium, however, the present method revealed an abrupt change in the grain angle, although radial splitting showed a smooth undulation throughout the radius. Using the present method, the analysis of a radial range of 5 cm required a single transverse section compared to 1,000 tangential sections 50-μm thick. In conclusion, the present method using a single transverse section, confocal microscopy, and image cross-correlation analysis provides more accurate data than radial splitting, and is less time-consuming than serial tangential sectioning.

The use of laser scanning confocal microscopy in the study of wound healing phenomena in plants

“…Here, some similarities of the results were noticed in comparison to another 3-D microscopical technique, namely the investigation of resin-cast replicas by scanning electron microscopy (Kitin et al, 2001). In mature wood the LSCM has not yet been applied for imaging complex tissue structures, but rather for performing quantitative measurements, e.g., of tracheid cell dimensions and grain angles (Moell and Donaldson, 2001;Ogata and Fujita, 2005).…”

Z-Axis Calibration in Optical Sectioning from Xylem Cross Sections for Grain Angle Measurement

“…The fluorescence from the deeper section might be the less detectable due to the interference by the cell wall above and it is possible that the optical section with the maximum intensity is obtained from the surface. There are, however, some reports that the cell wall information from the deeper optical section is less distinct (Wilson 1990;Pawley 1990;Donaldson & Lausberg 1998;Grey et al 1999;Ogata & Fujita 2005). Supposing the sample surface is perfectly flat, an optical section from the surface shows the sharpest and the brightest.…”

“…Confocal microscopy has been applied to wood anatomical studies (Donaldson & Lausberg 1998;Kitin et al 1999;Kitin et al 2000;Ogata & Fujita 2005). Fluorescence from lignin in the cell wall or from stain reagents can be used to observe xylem tissue.…”

“…Template matching using this function is well developed to detect image strain. Ogata and Fujita (2005) measured the tangential shift between two optical cross sections from different depths in 8-year-old Hopea odorata Roxb., a tropical tree, to estimate radial variation in its interlocked grain. Although computing two-dimensional cross-correlation function in large images requires complex calculations, Ogata and Fujita (2005) achieved a more rapid calculation of the cross-correlation function using the fast Fourier transform, by applying the description by Hino (1977) to a two-dimensional image.…”

“…Ogata and Fujita (2005) measured the tangential shift between two optical cross sections from different depths in 8-year-old Hopea odorata Roxb., a tropical tree, to estimate radial variation in its interlocked grain. Although computing two-dimensional cross-correlation function in large images requires complex calculations, Ogata and Fujita (2005) achieved a more rapid calculation of the cross-correlation function using the fast Fourier transform, by applying the description by Hino (1977) to a two-dimensional image. This technique enabled detection of changes in the wood grain using a less-destructive transverse section of Hopea odorata, without using the radial splitting method, which is conventional for measuring wood grain angle and requires a relatively larger wood portion (Ogata & Fujita 2005).…”