Image Measurement For Geometry

Along with the quick development of image sample and process techniques, the images have already become an important way of getting information from outside. It is also the vital content of image processing to evaluate products' geometrical characteristics using the correlative image information, so some companies develop many image processing instruments in succession, but little image measurement software is seen in the market, especially there is almost no independent of machine tool measurement software.In this paper, an image measuring software is proposed. It integrates the latest image processing techniques and advanced algorithms. It is based on the perspective magnifying principle. It is independent of machine tool. Using the non-orthogonal coordinate system, the true scales of image to parts can be obtained from the known geometries. When measuring, the image acquisition equipment can be hold in hand or fixed at a position. This software can be used to process monochrome or colorized images of planar small size parts by adopting dynamic error compensation and sub-pixel processing technique for each image, and to evaluate images by means of the geometrical feature fitting algorithms. The main research work in this paper is as follows:Non-orthogonal coordinate system in image measurement for geometry is researched. The measuring part needn't to be aligned precisely. This is benefit for operating.Only using single image for measurement, no machine body is necessary. It's convenient for spot measurement or timely measurement.The calibration procedure is proposed to decrease the bad influence of environment and to increase the accuracy of geometrical measurement by image.According to the dynamic properties of this image system, the dynamic compensation method is analyzed to get higher performance.Both colorful image and gray image are applicable to this system for edges verging. By researching the gray properties of the geometrical edges, the sub-pixel precision for locating can be achieved.Using C# language and components technique, it's convenient for maintenance, migration and reuse.By testing software and measuring actual workpiece, the precision of 0.03mm can be achieved. The measurement can be done by manual mode or automatic mode. It is also applicable for batch measurements.