Machine Vision: End User Solutions
Applications: General Manufacturing
Problem - Fiberglass Drilling
A
manufacturer of fiberglass components utilizing robotic-based drilling
equipment wanted to improve the accuracy of the robot's positioning
prior to commencing the drilling process. Although the robotic drilling
is very precise, the transport mechanism upon which the large size
objects is placed was minutely offset from object to object.
Since the drilling has to be extremely precise, any movement in the transport mechanism (even 2mm) could cause objects to be incorrectly drilled and subsequently thrown away as a very expensive reject. To make matters worse, the offset is never the same in X or Y from object to object. The robotic system needed real-time feedback on each part.
Solution
The vision component of the solution used Sherlock integrated with a DALSA PCVision frame grabber within a standard PC. The PCVision is interfaced to a standard analog camera. The camera is equipped with a zoom lens since the camera could not be located on the robot due to the intense vibrations caused by the drilling process. A high intensity light was also located remotely to illuminate the fiberglass surface target area to be analyzed by Sherlock.
The inspection process would commence upon a sensor detecting the transport mechanism having arrived in front of the robotic drilling station. This would trigger the camera to take an image and Sherlock would then measure the exact X and Y coordinates of the first drilling position relative to the robot's resting position. The XY calculation takes milliseconds to complete on a standard Pentium III processor.
Once Sherlock calculated the first drilling coordinates, a simple "send to serial port" command in Sherlock was used to transmit the XY coordinates from the PC's serial port to the robots existing PLC controller. We also have the possibility of sending such a command using the PCVision's on-board analog or digital I/O ports if the manufacturer's existing infrastructure required it.
By implementing Sherlock in this fashion, the manufacturer reduces product waste since each fiberglass component is drilled to their exacting requirements regardless of its position on the transport mechanism. This Sherlock solution offers a rapid return on investment for the manufacturer.
Problem - Wood Pallet Manufacturing
A
manufacturer of pallets required the primary and secondary break
down and centering of a short block log. Initially, this process
was done manually and was considered too slow for today's competitive
market, at just 1.5 logs per minute. The forest product mill is
chaotic with lots of debris and dust, presenting a challenging environment
for machine vision.
Solution
The vision system uses Sherlock along with a frame grabber, in a standard PC, connected to two analog cameras positioned a few feet away on the lateral side, with the log being back lighted with a flood light bar. This allows a dark contrasting profile of the log. A laser is used during initial calibration to establish a centering line with the end-doggers of the carriage.
The acquisition sequence occurs as follows:
- The operator positions a log onto a centering table.
- The system is triggered and the picture is taken.
- Sherlock processes the image and generates the required dimensions of the log. The calculated distance from the center of the end-doggers to each end of the log is also generated.
- This information is sent to the operator display and to the PLC.
- The PLC uses the acquired information to raise or lower the log to center with the end-dogger carriage and sends the primary and secondary breakdown information to the saws.
Customized
operator interface
Because the operator is relieved from manually centering the short block log and best guessing a diameter, this application has greatly improved accuracy and productivity.
Application courtesy of DCS Systems, Inc. Alabama (334-792-0823)
