Research And Implementation Of Digital Fingerprint Technology On Program Identification In CNC Machine Tool Fittings

With the rapid development of numerical control technology, diversified workpiece machining requirements make it critical to improve the machining program efficiency. The increase in machining program complexity will consume more time and decrease production efficiency. Through image search technique to retrieve the related existing machining program to copy or improve on, the industry can save the laborious repetition work so as to reduce programming time and improve production efficiency. Therefore, how to efficiently and accurately retrieve the required image has become a major challenge for high-quality and low-cost CNC workpiece machining. Thus, the present study probed into the application of image retrieval technology in workpiece image retrieving and proposed the use of digital fingerprint characteristics in image searching and matching in CNC machining field o. The main work is as follows:The first step is image pre-processing and image filtering. Based on analyses the effect of several common filtering methods, the study chose the median filtering method by showing the advantages it has in not only eliminating useless information and noise, but also retaining relatively well the edge information of the image.The next step is the image feature extraction. This paper analyzes the commonly used edge detection algorithms, e.g., Soble operator, LOG operator and Canny operator, etc. Analyses and experiments were performed for different edge detection methods. Results showed that showed that Canny operator-based edge detection had a better performance than others.Based on the contour information after the above image processing, this paper analyzed several shape-based image feature extraction techniques. In order to obtain workpiece image featured fingerprint with strong robustness, scale, rotation and translation invariance, the study designed moment-featured image feature extraction method and the corresponding feature image matching algorithm. Experimental results showed that it satisfied the scale and rotational invariance requirement and is suitable for image retrieval system.Finally, the present study proposed the workpiece-image-based digital fingerprint retrieval system. A though analyses and engineering design and application were attempted for its system framework design, database management, integration of systematic functions. The proposed system can satisfy the basic function required for workpiece image retrieval and can be of significance for the application of digital fingerprints in CNC machining field.