Development Of The Online Measuring System Of Grinding Surface Roughness Based On DSP

Roughness means the micro-geometry shape characteristic composed of tinyspaces, peaks and valleys in the machining surface, and it has an important influenceon the performance and lifetime of mechanical equipments. The advance inmanufacturing automation has created more and more high demand to develop onlinemeasurement of surface roughness, thus the online measurement of roughness hasbeen paid more and more attention. Nowadays, most of the roughness measuringsystems are based on PC and image acquisition card, which have some disadvantages,such as high cost, inconvenient installation and inapplicability in industrial workshop,and so on. Recently, a rapid development has been taken place in the field ofembedded system. The processing speed and storage capacity of the embedded systemis improving continually, especially for DSP chips. DSP chips have many advantages,such as abundant peripheral interfaces and high programmed ability, and theseadvantages provide a basis for the online measurement of surface roughness which isbased on DSP.Supported by the sub-topic of the Key Science and Technology Project of China,which named as “The online measuring technique and device of the surface roughnessof the typical grinding parts (Project Approval Number:2011ZX04016-041)”, thisthesis designed an online measuring system of surface roughness based on DSP.The main research works of the thesis are as follows(1) The definition of surface roughness and the light scattering principle wereintroduced. The qualitative and quantitative relationship between the surfaceroughness and the scattering intensity distribution were analyzed. Machine visiontheory was expounded, and then, the measuring scheme of surface roughness based onthe light scattering principle and machine vision theory is proposed.(2) The hardware platform of the surface roughness measuring system was put up,including the optical measurement module, the image acquisition module, the dataprocessing module and the communication module.The introductions of every module,which include the component selecting and how does it work, were presented.(3) A complete set of algorithms of the surface roughness measurement wasdesigned, including the image preprocessing algorithm, the feature extractionalgorithm and the surface roughness grades classifying algorithm based on themulti-class support vector machine. The simulation results in MATLAB platform show that these algorithms are effective in roughness grades identification.(4) Based on the deep study of the DSP chip, the designed surface roughnessgrades classifying algorithm was transplanted from MATLAB to DSP. Firstly theprograms which include the hardware programming and roughness detectionalgorithm are designed and written by C language.Then, the programs are debuggedin the CCS software. When debugged successfully, the programs are stored to theFlash chip in which data is difficult to be lost, and the system based on DSP can workindependently. At last, the experiment results show that the system is reliable andreal-time.Based on the DSP chip, a system for the online measurement of surfaceroughness is designed. At first, the optical measuring system is built up and thesurface scattered images are captured by an industrial camera. Then, the capturedimages are preprocessed and feature parameters are extracted by the high performanceDSP chip. Finally, images with different surface roughness grades are classified by themulti-class support vector machine model. Experiment results show that the designedonline measuring system can be used to identify the surface roughness gradeeffectively and conveniently.