IMPLEMENTATION OF COMPUTER VISION ON PICK-ANDPLACE MACHINE FOR SMALL-SCALE ELECTRONIC BOARD ASSEMBLY

• Main Author: Chandra, Richard
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/54189
• Institution: Institut Teknologi Bandung

Pick-and-place equipment for assembly of electronic boards is generally equipped with a computer-vision module to increase accuracy and prevent assembly errors. In this final project, a computer vision module was developed which is part of the development of a pick-and place machine prototype. The implemented computer vision aims to obtain the center-point (centroid) position of the component package to be assembled. The goal is to position the nozzle of the head at this centroid point when picking up the component. The image of the component was captured using a digital microscope with a resolution of 640x480 pixels. The processes carried out by the computer-vision module include several steps: (i) image-capture, (ii) image processing (iii) the centroid determination, and (iv) position the nozzle to the component's centroid. The digital-microscope is located at a height of 3 cm. The capture area of the computer-vision is 10 x 7.5 mm. Image processing is carried out using OpenCV library. The image is taken from above the component, which is known as look-down vision. After that, the captured image is processed by the contour identification method followed by the determination of its centroid. The processes of controlling camera operation, image processing, centroid determination and head movements are carried out using Raspberry Pi 4 board. The implemented computer vision algorithm has succeeded in recognizing the contours and determining the midpoints of two types of components: SOIC-8 (Small Outline IC 8 pin) and SMD 2012 resistor components, withe the size of 2 mm x 1.25 mm.