透過光画像に画像相関法を応用した紙の物性解析技術

Abstract

The pattern matching technique is applied to light transmission images to analyze paper properties and its versatility is demonstrated. A prototype system developed for obtaining light transmission images consists of a CCD camera, diffuse illumination from the backside of a specimen and software. In pattern matching, one of the two images to be compared is divided into small rectangular blocks. The other image is searched for a block with a similar pattern to one of the blocks of the first image taken as a template block while the block frame is moved pixel for pixel. The similarity is judged based on the inner product of the two block images as a vector. Ten-fold interpolation with the high-resolution bi-cubic spline interpolation function is successfully applied to estimate the deformation distribution with ten-fold resolution. In-plane distribution of tensile deformation is measured well even for paper with good formation. This technique is applied to in-plane distribution of restraint drying shrinkage of wetted copy paper. When a paper specimen is set and constrained vertically, every block around the center of the specimen totally shifted upward. It is because water moves downward by gravitational force, resulting in that the top part that dries and shrinks earlier draws the fiber network mostly from the flexible wet part immediately below. Thermography proves the water movement as a portion of low temperature due to absorption of evaporation heat remains to the last moment around the bottom part. On the contrary, when set horizontally, the specimen does not exhibit any total shift in the restraint direction, but shows shrinkage only in CD as well as observed with the specimen set vertically. For free drying of the copy paper, a flatbed image scanner was used to acquire images. The shrinkage tends to occur in a complicated manner, distorted by mixture of CD and MD shrinkages.