Visual Inspection On Positions And Dimensions Of Machined Parts

With the development of computer technology and photo-electricity, visual inspection has been used in a wide filed. Based on modern optics, it is an integrated inspection technology including photo-electricity, computer graphics, information management and computer vision etc. Computer vision has high measuring precision, fast measuring speed, steady inspection result and has the capabilities of inspection in real-time, real analysis and control, so it makes a significant contribution to enhance production efficiency, ensure product quality and enhance product precision etc.The purpose of this thesis is to develop a method to obtain the positions and dimensions of work-pieces such as step position, the start position of a thread and the center coordinates of a circle on the work-piece surface. The main research work in this thesis includes the follows.1. The paper studies the situation and trends of visual inspection technology, analyzes the problems faced by the visual inspection technology, defines the primary contents of the paper.2. The composition of visual inspection system is analyzed and improved based on the present detection platform. Appropriate optical devices are selected to construct the visual inspection platform. The image acquisition and data processing system are designed with the tool of Visual C++.3. Image processing techniques are studied in detail. Several image processing methods including median filtering, Roberts edge detection operator and threshold filtering are used to extract the curve formed by the intersection of linear structured light and the surface of the machine part.4. Various edge detection algorithms are studied in detail. Key algorithms of the step position detection based on the least square method and key algorithms of the hole center detection are raised which can be used to test if the distance of two steps and the center distance of two holes can meet the requirements of precision machining. The algorithm of thread starting position detection is put forward which is used in conjunction with the thread probe to measure the thread accuracy. 5. The methods of camera calibration are studied. The calibration of visual inspection system is carried out according to the actual accuracy requirement of the item and the situation.6. Experiments are designed to test the visual inspection system. The stability, safety and the feasibility of the detection algorithm are verified. The various factors which impact the visual inspection accuracy are analyzed in detail.