Appearance Inspection System Of IC Chips Based On Machine Vision

As the rapid development of electronics industry, more and more IC chips are needed, and the inspection of IC chips is becoming an indispensable part in the electronics industry. At present, most production lines are still using the traditional method to detect with technician's eyes. When workers find the defects, they discard the unqualified chips. Such quality detection method is easy to cause visual fatigue, of which the efficiency is low, and has high cost.At present, some technician use computer image processing technology in production. They detect the position and the count of pins on the chip through the camera. But these detections are only about the locations of pins, not including all of the quality testing of chip's appearance, such as the color of the pins, and the surface printing information of the chip, which can not reach the request of a comprehensive detection of IC chip defects.The paper devotes to the study of appearance inspection system of IC chips based on machine vision, which takes the software's factor as the goal, and concentrates in the study of the intelligent inspection algorithm and the development of the software system. The main contents of the paper include as follows:(1) The paper presents several pre-processing algorithms of IC chips inspection, including automatic adjustment of light source, chip positioning and image rotation, which are more conducive to the post processing.(2) The paper study a series of algorithms about pin positioning, including the algorithm based on template match, and the positioning algorithm based on regional identity. According to the characteristics of the pins'region, the paper presents a simple algorithm based on scan positioning in a binary image, which is applicable to most SOP chips. The algorithm calculates each parameter of pins and compares these parameters with the standard parameters to implement the inspection of position defects of pins. The defects such as lack of pin, offset of pin's position, and offset of pin's height can be detected.(3) The paper utilizes the theory of HSI color space to convert the color space from RGB space to HSI space in the image of pins, which implements the inspection of color defects of pins. The defects such as solder layer damage of pin, and pin-oxidation can be detected.(4) The paper presents a preliminary treatment algorithm which can conclude whether the printing information is clear enough. The algorithm calculates the integrality and clarity of printing information to find out whether the printing information is clear enough. (5) The inspection software system of IC chips is developed in the paper. The software system can basically detect some defects of SOP chips, including the defects of pins and defects of the printing information clarity.