Position detection

from Wikipedia, the free encyclopedia

The position detection is one of the most important tasks in the field of machine vision of machine to automate processing. The task arises from the practical task that although parts are usually delivered clearly identified by the conveyor system , their exact position and alignment are still undetermined. The task of position detection is to determine the exact three-dimensional position and alignment of the components with the help of an optical image acquisition system. On the basis of this data, precise position processing can be carried out on the article, such as further processing of bores, application of labels, etc.

The position detection of an object also includes a check by object detection . This is done using reference markers . The task is simplified in that only a single object has to be identified and its position in space has to be determined.

A main challenge of position recognition is to ensure the highest possible precision with the shortest possible recognition time and maximum recognition reliability with the least possible effort in the multi-axis coordinates of the processing system. For this purpose, local recognition algorithms are required for soft materials that are not completely dimensionally stable, for dust-free processing to be guaranteed, for measures to increase precision at the maximum detection area to be taken, and for adjustment procedures to be used to compare the detection system and the processing system.

Application examples

One application is the subsequent application of a laser-burned label to tires that have been selectively provided with additional leak protection equipment. The core task is the exact three-dimensionally correct engraving of the storage, production and material-related, individually slightly warped, brand-new rimless tires. The label should be aligned exactly with an existing lettering. Due to the low-contrast dark color of the tires, special laser-assisted scanning of the tires is required. This is done using a laser fan that is recorded at an angle by a camera.

A second application is the determination of the position of fastening holes on aluminum rims with an accuracy of 0.1 mm. The rims delivered by the conveyor system have different angles of rotation. The misaligned machine determines the exact centering position of the rims and their alignment. Then the painted rims are stripped of paint on the contact surfaces of the wheel nuts. This is done by a machining head that is guided by a multi-axis system. If article data is available for a rim, the machining height can be taken from the article data. Therefore, a one-camera system is sufficient, which determines the hole positions by means of a projection screen in transmitted light. The non-aligned machine is aligned using an image-based alignment procedure and a test specimen. Various system parameters such as camera distance and axis positions are determined here.

Image analysis software and peripherals

There are application software packages for identifying the object positions. Bores or other features can also be optimally determined using individual software kernels. Current GigE Vision , Camera Link , USB or FireWire cameras, together with a PLC machine control, enable reliable, precise processing.

literature

  • Christian Demant, Bernd Streicher-Abel, Peter Waszkewitz: Industrial image processing: How optical quality control really works . 2nd Edition. Springer, 2001, ISBN 3-540-41977-2 , Chapter 3: Situation detection, p. 81-102 .