The Position And Detection Technology Of LED Chip Based On Machine Vision

The rapid development of electronic market sets higher requirements for the manufacturing of electronic products. The requirements of high precision and fast development promote the continuous updating of production technology. With the continuous research of machine vision, the vision technology is more and more widely applied in the field of industry. With fast development of automation and product update in practical production, the machine vision positioning and detecting technology are having more and more advantages over traditional technology. Machine vision is playing an significant role in positioning and detecting process of LED chip assembling.The development trend of LED industry and related technology of machine vision are introduced. The visual positioning algorithm of LED chip and the lamp cap visual detection are accomplished. The main contents are as follows:In the aspect of LED chip visual positioning, this article has studied on some commonly used principles of image processing algorithms. Through study on camera model and camera calibration technology, image distortion has been corrected and connection between image coordinate and world coordinate has been established. In order to highlight the characteristics of LED chip, reduce the complexity of algorithms and improve positioning precision, based on different image background, this article put forward two sets of visual positioning schemes:(1) Scheme No.1 realizes the separation of chip and background through threshold segmentation. Through the comparison edge detection algorithm, it selects Canny edge detection to extract the contour features of rectangular chip. Based on characteristics of rectangular contour, a dynamic template detection algorithm is put forward, which realizes the chip identification of the positioning. By contrast experiment between the Hough line detection and minimum external moment algorithm, it shows the high speed and precision feature of this algorithm. (2) Scheme No.2 is about the pin features of the chip. It extracts two symmetrical pin profile by using the perimeter and area features of pin profile. And it determines the center of pin by using centroid method, and realizes visual positioning and posture identification of the chip by using the location and direction features of two pins.In the aspect of LED lamp cap appearance detection, this article analyzes the gray feature of the chip and the PCB. In view of the two gray contrast in traditional gray-scale image, this article increases the gray-scale contrast through the conversion of color model, and realizes the segmentation processing of chip and PCB. In view of the defects of chip edge breakage, leakage and fouling, the visual detection for perimeter, area and Hu moments are realized. For the chip positioning error defect, dynamic template positioning identification algorithm is used to realize the detection work.Through actual repeated test and contrast, the visual positioning and detection scheme raised in this article is workable. The identification positioning algorithm has the characteristics of fast speed, high precision and good stability, which will make a foundation for the further study of machine vision in the future.