We describe a method to detect knots and faults in wood logs based on microwave measurements and the depolarising properties of wood. The depolarisation is an effect of the anisotropy in the dielectric properties of wood, which is caused by the fibre structure. The measurements are made in the frequency domain where a discrete reflec-tion spectrum is sampled. By transforming the frequency spectrum into the time domain the spatial distribution of the reflections appear. The wave is transmitted in a pure linear polarisation, tilted to 45° For each discrete frequency all available information is measured for the reflected wave and therefore it is possible to calculate the complex polarisation ratio and the state of polarisation which are the quantities containing the most visible information. The measurements are one dimensional and by combining measurements from different directions we create tomographic slice images of the inner structure of a log. Our measurements has showed the possibility to follow the extension of knots in a log using only a iso-surface visualisation. The used equipment is working at low intensity and is therefore virtually harm-less to human beings, furthermore it is portable and can be operated by one single person.

The ability to measure the presence of decay in trees using three different methods is studied. The methods are based on drill resistance, sound wave propagation measurements, and microwave polarimetry. All methods have been tested on samples with artificial cavities as well as on decayed samples. The experiments with both types of samples have showed positive results.

When water is removed from the paper during paper making, a dimensional change occurs in which the paper shrinks in the direction perpendicular to the direction of processing. The dimensional changes vary across the web and influence, e.g., the surface and compression properties of the paper; they also complicate the control of the paper machine. In this article, a robust method for estimating the relative shrinkage profile is presented. The method is based on a one-dimensional recording of the imprints from the forming fabric, using a fluorescence technique. The recording is transformed into a time-frequency spectrum, on which three different frequency estimators have been evaluated. In simulations on synthetic data and measurements on paper profiles the estimator that maximizes the correlation energy showed the most robust and accurate performance of the methods evaluated, even at a low signal-to-noise ratio.