Highly accurate and repeatable measurements of the subject. The manufactured parts are coming in a continuous stream at speeds of 1200 parts per minutes. As these parts move through the vision system their images are captured and dimensions of specified features are obtained and checked. These are classified to be within or outside user-defined specifications.
The measurements are within 10µm form the actual and repeatable within a range of 3µm.
High-speed inline precision gauging systems have been Zentron’s first product launches. They have been deployed in measurement based quality inspection of electronic terminals in leading domestic and multi-national connector manufacturing companies.
OBJECT DETECTION / FEATURE DETECTION:
In a particular object certain physical parts (eg. stem of a fruit or metal pin on a PCB board) or characteristics (eg. crests and troughs on the surface) can be identified based on the appearance of the object in the image.
These characteristics can be called features. Identification of these features using image processing techniques is in the realm of feature-detection.
SURFACE DEFECT DETECTION:
An anomaly on the surface of manufactured products are cause the part to visual inspection. Visual inspection is a quality control method commonly used to certify a product.
This is one of the early ways that identify possibility of part failure.
Automotive parts like bearings, seat-belt inserts etc use this criteria to mark a part as NG.
Simply put, algorithm is the procedure for solving a problem.
The algorithm forms the core part of the vision system,after the analysis it outputs pass/fail Boolean flag. These flags collective enable a decision of OK or NOK.
The categories of algorithms include image handling, search area definition, edge detection, particle analysis and overlay presentation etc. Generally, the kit of algorithms enable the user obtain the GD&T measurements.
Automation is used for many things in the industry. Vision based inspection systems like the precision gauging and surface defect detection techniques need in-line rejection methods, with automation, to implement their quality control.
This feature allows the inspected part to be removed from the production line when it does not meet the quality criteria.
In the packaging industry, labels are applied with the date-of-manufacture, batch-number and price.
These labels have to be inspected for accuracy of the data, print-quality and presence of stray particles. Zentron’s solution is a product that inspects boxes at the rate of 3 per minute moving on a conveyor. Further, the box which contains a NG label is pushed out of production line that is a part of the vision-based inspection system
Electronic connector, post assembly, have to be inspected whether the right pins have been installed in the connector. The definition of right-pins includes the height of the pins, location and its bend.
Using the shine from the tip of the connector pins the location of the pins are identified.
Then the algorithm compares there expected and actual locations of the pins and declares whether the connector is OK or NG.
Cameras capture the color information of an object. Color information can be perceived as a mix of red, blue and green colors. Color identification involves comparing the color content of an image with a particular, pre-defined, combination of red, blue and green via special algorithm.
At times a need arises to identify, separate and quantify objects based on their color. For such applications, the color of the sample gives vital information.
Analytics is the identification and presentation of meaningful patterns in the data. Availability of inspection data enables descriptive, diagnostic, predictive and prescriptive analytics.
Essentially availability of real-time data in the customized format eases the decision making process and keep it current. It will definitely be a powerful tool if you inspection systems enable with actionable analytics serving it to you laptops and mobile.