A 2-D Measurement System For Collecting Object Edge Data Based On Machine Vision

This dissertation discusses a measurement system with machine vision on the collecting 2-D object edge data, which is based on the principle of Reverse Engineering.According to the principle of the system and the need of designing work, it researches and discusses deeply in several aspects as follows, and points out new ideas and methods. It includes the following contents:(1) It studies the principle of machine vision calibration, and uses calibration matrix groups of subjunctive grid to realize the method of collecting boundary edge of the object on the condition of using more CCD cameras.(2) It makes the feature extraction in pattern recognition easier by setting up object of reference with regular shape.(3) It studies a method of the image binarization about adaptive threshold value, 4-neighbors is used to eliminate the effect of noise about the boundary.(4) It studies the chain code and average chain code, bring forward a method of using the chain code difference to search for the turning points, then subjunctive beeline and change-length subjunctive beeline are used to distinguish linetype, and presents a practical algorithm.(5) It develops modules including image collection, machine vision calibration, measurement, image binarization, edge extraction, graphics vectorization by means of VC++6.0, and integrates this software system.(6) A machine vision system for collecting 2-D object edge data is developed (including CCD cameras, image card, workbench, etc.), the hardware system combined with the software can measure and record data correctly.By theoretical research and experiment analysis, the author proves this system, compared with other general measure and record system, is more quick, non-contact, and automatic measurement. The shortcoming of the system and the improved suggestion are proposed.