Research And Design On Relief CAD/CAM Based On The Digital Image Processing

In order to meet the increasing demand of the market for the relief products,Zibo Zhi Heng Industry & Trade Co., Ltd. has commissioned the CNC technology and equipment research group to develop image relief CAD/CAM software and three axis CNC engraving machine. This paper makes a detailed study on the key parts of image relief CAD/CAM and three axis CNC engraving machine. Specific research contents are as follows:1. In this paper, the depth gray map is selected as the data source of relief CAD/CAM,and do the pretreatment. In order to achieve the processing requirements of fine relief, for the details blurred image, the high contrast preserving and bilateral filtering algorithm are used to enhance the image details.CB morphological filtering, Gauss filtering, linear light mixing algorithm are used to achieve the filtering of preserving image details. These treatments can make the relief product surface become smoother. In order to match the image and product size, the nearest neighbor interpolation, bilinear interpolation or three convolution interpolation algorithm are used to scale the image.2. In this paper, the gray level of the image pixels is obtained according to the gray psychology formula. The image gray values and processing depth values are processed by linear or nonlinear (power exponent) transform to achieve relief model height data.3. In order to obtain the relief model which is smooth and closer to the actual situation. and the tool can pliant move on the surface of the workpiece, so the SFS algorithm is introduced. According to the intensity of the image, the height of the pixel is obtained. Because the linearized method is used to solve the SFS problem, so the RGB color space is converted to LUV color space, and get the pixel brightness value. In this paper, the image brightness value is assigned to the image intensity.4. Because image processing involves a large amount of data, The data processing of CAM module discards the traditional form of data structure and chooses the way to operate the document. Image gray data, height data, tool path data and other documents are established, and according to the custom data format to read and write documents to achieve data communication. In this paper,the acquisition of relief CNC codes by direct method and 7 times bezier curve fitting method.5. This paper focuses on the five layer architecture mode of the open CNC engraving machine system and the interpretation and execution of CNC codes.The CNC system modules are separated from the operating system by software bus, and the system function is simple and effective. In order to improve the efficiency and stability of the CNC machine, thro ugh the establishment of intermediate files to change the process of interpretation and execution of the traditional CNC codes and the pretreatment and interpolation of CNC codes are separated. And other key parts of CNC engraving machine system are studied in detail. For example, PCI data card operation, axis electrical design, design of machine tool spindle safety link, electrical design, design graphics, multi task real-time processing design.6. The differential principle and the 2D CNC differential interpolation theory are introduced. Based on the 2D CNC differential interpolation theory,this paper studies the 3D CNC differential interpolation theory. The interpolation error is studied in detail, and the error is not more than 0.707 pulse equivalent.At the same time, the acceleration and deceleration control and interpolation efficiency are also studied. The interpolation algorithm is simulated, and good results are obtained.7. The three axis CNC engraving machine and the CAD/CAM software are developed. The Machining experiments of the five pointed star, oblique Chinese characters and the image contours verify the accuracy of the 3D line differential interpolation algorithm and the stability of the CNC engraving machine. The experiments also verify the accuracy and robustness of relief CAD/CAM algorithm based on the gray- height model or the SFS height model. And the two methods are compared and the latter is better than the former.