Defect Location and Classification of Wire Bonding Solder Joints

Abstract

The invention discloses a method for locating and classifying defects of wire bonding solder joints, which comprises the following steps: 1) using an industrial camera to obtain images of bonded solder joints; 3) Use the gray projection algorithm to remove redundant non-solder joint areas; 4) Use the region growing algorithm to initially extract the area where the solder joints are located, and then use the level set method to segment defects; 5) Utilize core principal component analysis to extract the linearly separable main features of solder joints; 6) send the extracted main features into the random forest classifier to classify the defect types, and provide welding parameter adjustment suggestions according to the results of multi-classification; the present invention is based on Compared with other solder joint detection technologies, the wire bonding solder joint defect location and classification method based on image processing and machine learning has the characteristics of high precision, fast speed, and high level of intelligence. It has a relatively large application prospect in the actual electronic industry production .

Description

technical field [0001] The invention relates to the technical field of wire bonding solder joint defect location and classification, in particular to a wire bonding solder joint defect location and classification method based on image processing and machine learning. Background technique [0002] In the fields of scientific research and engineering applications, ultrasonic bonding has always played a very important role. Ultrasonic bonding is that the head of a welding tool with ultrasonic energy is pressed against the surface of the metal wire to be bonded with a certain pressure, and the vibration brought by the ultrasonic energy causes the metals in contact with each other to rub against each other and finally bond together tightly. The quality of solder joints directly determines the quality and reliability of chip manufacturing. For the detection of solder joints, the method widely used at home and abroad is to obtain the appearance of solder joints by imaging with ind...

AI Technical Summary

Problems solved by technology

For the detection of solder joints, the method widely used at home and abroad is to obtain the appearance of solder joints by imaging with industrial cameras, but the resolution of the collected original solder joint images is relatively high and contains a large number of background areas, so it is necessary to extract the area where the solder joints are located , the general methods include optical flow method, feature description method and gray level consistency matching (NCC) algorithm, but the above methods are only good for solder joint images with regular geometric shapes, and generally require similar machine vision problems. Design special algorithms according to actual scenarios

Images