Sightech Machine Vision - Self Learning Eyebot
Self-Learning PC-Eyebot
Sightech Home Page
About Sightech
Contact Sightech
Frequently Asked Questions
News, Events, and Shows
Dealers and Resellers
Vision Products PageApplications PageSupport, Documentation, etc.Intelligent Vision ExplainedSightech Careersfiller

Applications - Easy to Install and Maintain General Inspection for Canning/Bottling Applications

PC-Eyebot can be effectively general inspection of containers and labels for bottling and canning lines. Sightech's equipment can be easily mounted directly on the conveyer line and provide multiple inspection tasks all with one or two camera! The ability to define multiple inspection Areas and apply sophisticated decision consolidation provides all of the defined views (Areas) to result in a single PASS/FAIL decision per can or bottle.


The following example shows the ease of setup, training, and operation as well as the high degree of inspection ability.


PC-Eyebot can easily detect for the following 11 conditions:
A) lot code presence, B) missing, incorrect, skewed neck label, C) flaps in neck label,
D) incorrect, missing, skewed front and back labels, E) incorrect, low fill level, or missing


If you want complete general filled bottle inspection, be sure to contact Sightech Vision Systems at!

To only produce the utmost in quality, automated inspection for a range of potential defects is important. PC-Eyebot offers the benefits of:


a) easy setup,

b) intuitive operation,

c) general inspection performing many task simultaneously, and

d) powerful and sensitive full-speed inspection.


Running: The image (right) shows PC-Eyebot inspecting for 5 different defect types on the front of the dressing bottle:


1) Absence of lot code printed on top of cap

2) Incorrect, missing or misplaced neck label

3) Incorrect content color and/or texture

4) Missing content

5) Incorrect, missing, or misplaced lower front label


We have slowed down the frame rate for visibility - much higher speed is possible.

Running: The image (left) shows PC-Eyebot inspecting for 7 different defect types on the back of the dressing bottle::


1) Cap height

2) Puffed out neck label

3) Skewed or Flapped neck label

4) Missing neck and/or lower label

5) Content type color and texture

6) Incorrect lower label

7) Misplaced lower label


We have slowed down the frame rate for visibility - much higher speed is possible.


Step 1 - Setup: Place six Areas - one to each aspect of the general inspection. This examples inspects label correctness, label placement, and contents. (above).

Step 2 - Learn: Simultaneously train all Areas on several good containers. Training is very easy to perform and normally 5 to 60 seconds is all that is required. Sightech can configure an external "Learn" button that can be accessed by the operators without directly controlling the system. When training is needed, the operator can just push the button. (above)
Step 3 Run: Running in Inspect mode, his example, (above), shows a PASS condition. The back view of the container shows no defects. (above) The (above) image shows 1) missing lot code on the top of cap, and 2) incorrect lower label in the front view. The red Area outlines show which Areas are signaling problems. This is a FAIL condition.
The (above) image shows a puffed-out label on one side from the back view of the container. The purple markup shows where the defect is detected. This is a FAIL condition. The (above) image shows a label flap - an overlap misalignment. By dedicating a separate Area for each inspection category, you can set the decision sensitivity separately for each type of defect.
We have shown several important points:
1) Once setup, simple training prepares PC-Eyebot to go into operation for a given bottle type;.
2) Many inspection tasks are performed simultaneously.
3) Using multiple Areas, each for a desired defect type, allows you to set the decision sensitivity as you wish.
4) Sightech's self-learning technology lends itself well to container inspection applications because it absorbs all of the natural variability.
5) Line change-over's and product variance are easily accommodated due to a simple start-from-scratch or cumulative training operation.
6) Additionally, PC-Eyebot has an option that can save up to 20,000 images of failed containers! These images can be accessed over ethernet from a remote location as well and analyzed to aid in process improvement.

Home | About Sightech | Contact | Products | Applications | Support | Technology | Careers | Site Map | Resources

Copyright © 2005 Sightech Corporation — All rights reserved
Privacy Statement

Valid XHTML 1.0!