The Design And Implementation Of Indusrial Wide-format Image Scanning And Vectorization System

Generally, the solution of small industrial parts detection was using the CCD or CMOS cameras, when it came to the wide-format applications, although those camera solutions could work, but made things a little bit more complicated. Because it had to carry a 2D movable platform, and during the 2D platform move, the camera would capture partial images of wide-format object, then using the image mosaic algorithm to obtain global view of the object. The whole process was very complex, and demand highly platform moving precision, also, it became more difficult to integrate to the industrial production line. A new solution was discussed in this thesis. Which was using multiple CIS sensors concatenated together, it looked like zigzag pattern, and in this dissertation, it was called wide-format scan component by us. The wide-format scan component was easy to implement, convenience to integrate with the industrial production line and had a compact structure. And the system discussed in this dissertation, had the function to export the vector format contour set of the shape, which had been used in apparels, leather and other similar manufacture industries. their applications needed to do laser version based on existing plate shape, and export the vector format of the shape.This dissertation principally discussed the design and implementation of system software and algorithms. The main contents as follows:(1) system’s hardware designOutlines the overall composition of the system hardware and the design of each module, mainly including hardware framework, wide-format acquisition scheme, the circuit of hardware control and the transmission protocol.(2) system’s software designDesign and implementation of the system software architecture, and focused on analyzing the wide-format memory management and assembly line acquisition methods of image acquisition.(3) image correction and fusion algorithmAnalyzed the image sensor nonlinear response correction algorithm and stitching algorithm involved in the CIS wide-format image acquisition.(4) Image processing key algorithmsAnalyzed image processing algorithms, including filtering, threshold segmentation, edge detection, contour vector points set optimization algorithm. Deeply analyzed the vector point set optimization algorithm.The first innovation point of this dissertation was the assembly line method introduced to the wide-format image acquisition, which solved the large volume data of real time wide-format image acquisition. Another innovation was, the system discussed in this dissertation, had the function to export material contour vector point set, which satisfied the requirement of the laser version on existing plate shape in the apparel, leather and other similar industries, it solved the traditional solution’s problem which relying on artificial and inefficient, to our knowledge, this had not seen reports of the similar system yet.