Applications > Previous Apps > Cutting Machine Control

Cutting Machine Control

Overview

A manufacturer needed a machine capable of trimming the perimeter of large parts in a batch type operation. The parts consisted of a variety of sizes with significant dimensional variations within several standard size groups. Runs of non-standard sizes were also possible and it was common to mix parts of different sizes. Hard fixturing for part placement simply was not an option. A Machine Vision System was needed to locate the parts and generate a CNC program to trim the parts.

Operation

The parts are placed on a conveyor in a work area with minimum constraints on their placement. The Machine Vision System images the work area, locates the parts, and generates the CNC program. The operator is given the opportunity review and modify a graphical display of the cutting program prior to downloading it to the the cutting machine. Once the program is downloaded, the conveyor indexes the parts into the cutting area. While parts from the previous load are being trimmed, the operator loads up another set of parts.

Machine Vision System Details

A Machine Vision system was designed to image the work area where the parts were placed on the conveyor. Because of work area size and resolution requirements, the installed system utilized two Kodak 1.4i MegaPlus Digital Cameras. The images from the two cameras are combined to provide a 1200 x 2000 pixel image. To simplify the mechanical requirements for mounting the cameras and the associated tolerances, an inter-camera registration procedure was incorporated into the normal camera calibration routine. Additionally, the calibration procedure was required to correct for excessive geometric distortion due to the necessary use of wide angle lenses. Once calibrated, the two cameras are registered with respect to each other to sub-pixel accuracy across the entire work area.

A third camera was located on the cutting machine in order to compensate for conveyor index errors. This was a standard RS-170 analog camera. All three cameras were used by a single machine vision system based on a Windows NT 4.0 platform with custom software written in MS Visual C/C++. Communication with the cutting machine was via an RS-232C interface. The implemented Machine Vision system may be further enhanced to communicate job information to a file server via an Ethernet interface.