Research On MELF Component Detection System In Mounting Machine

This paper is based on the placement machine project as the engineering background to study the detection algorithm of the MELF chip. Because of the development of electronic technology, the electronic products are more and more powerful, and their circuit are more and more complex. To meet the needs of production, the placement machine must have the ability of detecting many kinds of chips. And the MELF chip in this paper is one of them.The detection algorithm of chips use some kinds of image processing algorithm to process the chips’ image. After all of these steps we get the coordinate and the rotation angle of the chip center. All of these will be used during the chip mounting. So these results will affect the precision of chip placement directly. Therefore the processing algorithm must have high precision, and at the same time more efficient. In this paper, the processing algorithm has two main steps. At first, the chip is located coarsely. Then, based on the result of first step, we will locate the chip precisely.During the first location, we use template matching. Due to the special cylinder packaging form of the chip, its curve surface has different brightness in its picture. So we can not finish the binarization to the picture. Thus we get the edge image of the chip directly. Use the edge image to match the template. At last we find out the best matching position and best template. Through these we get the coarse position and rotation angle. During the matching processing, we use pyramid algorithm to reduce the calculation amount.During the second location, we use the result of the first step to get a ROI image from the original picture. And use the ROI image to get precise position coordinate. At first, according to the geometrical relationship we can get the key edge points by using edge tracking algorithm. Because the images of MELF chips are a kind of rectangle, so refer to these key points, we can abstract the minimum bounding rectangle of the chip. We can know the position and rotation angle of the minimum bounding rectangle, so we will know the chip’s.In the paper, we made experiment and get these results. Analyzed these results and used them to place the chips. We got a pretty good result. So the detection algorithm is feasible.