Research On The Vision System Of Gantry-Type Placement Machine

Placement machine is the most important equipment in the Surface Mount Technology. It has been widely used to mount electronic components fast and accurately onto the Printed Circuit Board. Placement machine's design is a multidisciplinary integrated technology, in which the machine vision system plays a key part. This dissertation focuces on the study and implementation of the vision system. The main research contents are illustrated as below.Firstly, the general scheme of the on-the-fly vision system is proposed and the function of key modules is analysed. Also, the system performance indices are put forward on the basis of much analysis.Secondly, an image acquisition system is designed and developed to capture the moving electronic components. The acquisiton system is composed of a optical imaging system and an LED strobe lighting system. The lighting system can supply strobe lighting of various angles and intensities. So the each kind of components can be illuminated with the best lighting condition. And the imaging system can capture high quality images in extemely short exposure time.Thirdly, some foundmental algorithms of image processing and feature extraction are studied. For example, edge extraction with pixel or subpixel precision, geometric primitives extraction and region feature extraction.Fourthly, the numerous surface mount components are classified into three classes from the standpoint of image processing technique. Consequently, the generality and flexibility of the algorithm are much improved. Based on the previous studies, components locating and detection algorithms based on component lead are researched and implemented for each class of the components. The algorithms exhibit high execution speed, high locating precision and strong robustness in a number of experiments.At last, the algoritm for detecting and locating the widely used PCB circular marks is studied. The research has taken various possible interferences from the production field into consideration and lots of experiments have been done. Experiments show that the algorithm not only runs fast and locates precisely, but also possesses strong robustness towards those interferences.