Image Processing.
Localizing. Qualifying. Identifying.
Various methods that make assembly production perfect.
HIGH-PRECISION PRODUCTION REQUIRES OPTICAL MEASUREMENTS.
If high precision is required, it is inevitable to make use of optical measurement methods. For this purpose, we use high resolution cameras. The images created are processed by a professional software which has been designed by our engineers.
THREE VERIONS ARE SUPPORTED BY AMADYNE.
Our software supports Localising Image Processing, Qualifying Image Processing and Identification. And there’s more benefit for AMADYNE customers. As all these versions were born in our own R&D department any alternation (if required) can easily be done.
Localising Image Processing.
Pick & Place-Process with Highest Precision.
Five clever measurement methods can do the job.
THE LOCALISING IMAGE PROCESSING PLAYS A MAJOR PART.
To run a pick-and-place process with highest precision it is necessary to optically measure the position of all components before production start. To do this successfully AMADYNE offers several methods from its own image processing library.
FIVE CLEVER METHODS.
There is a distinction between those methods which supply position and angle information and those which only supply position information. AMADYNE has five methods in its portfolio: Pattern Recognition, Fiducial Recognition, SMD-Recognition, Circle Recognition, and Surface Focus.
Pattern Recognition.
Subpixel-Precise Information.
This clever feature represents an automatic stabilizer and enables great system robustness.
PATTERN RECOGNITION AND EDGE PROFILES.
The AMADYNE feature-based pattern recognition analyzes reference images and very precisely compares coherent edge profiles (brightness criteria) there. These are subsequently recorded and only their position and direction is stored. The search-image is treated accordingly, and the system looks for the position of the best accordance with the edge profiles.
EXTRACTION OF CONTRAST TRANSITION.
The system’s robustness is increased by this step of extraction of coherent contrast transition. Robustness in terms of changing illumination processes, disturbance and noise. By the pattern recognition users can get subpixel-precise position and angle information.
Fiducial Recognition.
Precision is Our Credo.
Discard probability reduced to nearly zero.
EXTREME SMALL PASSER BRANDS AND SYNTHETICAL REFERENCE IMAGES.
Reference passer brands which calculate the very important positions in a micrometer range and are designed for automatic corrections are called passer fiducials. The AMADYNE feature-based technique uses an ideal synthetic reference image which has been selected from a strongly equipped sample library and which can be also automatically be adjusted. This synthetic reference image is the foundation for a subsequent process: the normal pattern recognition.
FLUCTUATION OF SRUCTURE RECOGNITION CAN BE AVOIDED.
More accuracy and a better reliability can be reached by this method if there are normal fluctuations in terms of structure, form and size recognition. There’s subpixel-precise information, too.
SMD Recognition.
Standard Component Measurement of JEDEC-Housed Items.
AMADYNE’s own and special solution.
PASSIVE SMD (SURFACE MOUNTED DEVICE) MEASUREMENT.
A precise optical measurement of passive JRDEC-housed SMD-components often represents a major challenge. Therefore, AMADYNE offers a special solution which is optimized with respect to geometric and optical specifications of such units. There is recognition of contact surfaces and subsequently of bodies and rotations.
SUBPIXEL-PRECISE INFORMATION AVAILABLE.
This kind of measurement supplies subpixel-precise position and angle information.
Circle Recognition.
Reliable Recognition …
… of circles larger than the camera-produced image.
THIS KIND OF RECOGNITION SUPPLIES SUBPIXEL-PRECISE RESULTS.
Based on the Hough-Transformation this method has been designed for the recognition of circles. Even circular forms which are larger than the camera-created image can be recognized by this technology. It is enough to visualize a rudimental image section. The centre is always outside of the FOV. The circle recognition supplies a subpixel-precise position but logically no angle information.
Qualifying Image Processing.
Qualifying is Also a Decision-Making Process.
And such (automatic) decisions are always good ones.
HOW SUCH PROCESSES ARE DEFINED. AND WHAT THEIR TECHNICAL FONDAITION IS.
Automatic qualifying, recognition and identification comes in the fist place. And, yes, very good decisions (the machine is able to make) follow this step. Qualification is a term which describes what a component’s role during the further production process is. As a rule, qualifying activities are prior to the very production process. Eventual damaged components can be excluded –automatically. For every manufacturer discard (inclusive associated other cost) is a big problem. Automation, as we understand it, avoids such problems.
HOW THESE FUNCTIONS CAN BE USED. AND WHERE USERS CAN FIND MORE ADDED VALUE.
AMADYNE offers a broad spectrum of measurement methods – always suitable for the corresponding job. A rather simple method is Ink Dot Recognition. More complex techniques are Particle Inspection and Chipping Inspection. The latter can also serve as pure documentation if customers choose just this function and, hence, are great instruments for quality control, production planning and wafer map creation.
Reference Images Give Information About Quality.
Quick and easy detection of eventual component damage.
A COMPARISON WITH GOLDEN SAMPLES.
Particles and structural damages at a component’s surface can have a heavy influence on the functionality and reliability of semiconductors. Our particle inspection has the ability to detect such failures by comparison with a great deal of reference images of intact components (so called ‘golden samples’). This name is usually given to components which have been defined as 100% perfect by the microchip designers.
VISUALISATION OF INSPECTION RESULTS.
The inspection results can be documented as text or graphic and affected components can be excluded from further production. Sensibility and minimal size of the defects to be found can be defined in the setup menu. Other methods are Inkdot Recognition and Chipping Inspection.
The Chipping Inspection.
Vulnerable Edges.
How damage can be easily detected.
THE CHIPPING INSPECTION IS A SUITABLE METHOD.
The vulnerability of component edges is sometimes a problem. The chipping inspection can detect, analyze, quantify and document damages.
AUTOMATIC COMPONENT EXCLUSION FROM THE PRODUCTION.
If an edge damage is beyond predefined limit it is automatically excluded from the production. Other methods are: Ink Dot Recognition and Particle Inspection.
THE VULNERABILITY OF COMPONENT EDGES IS A CHALLENGE.
Customers can use the chipping inspection for automatic detection and exclusion.
Identification.
Identification and Documentation.
Complete Record of all Important Information.
IN FOCUS: THE FLOW OF MATERIALS.
Especially with high-quality and otherwise critical products (security etc.) the complete documentation of all flows is strongly required. From a software perspective AMADYNE customers are supported by the SCT-option and the Identification.
OCR AND 2D-CODES.
The AMADYNE Image Processing give users the opportunity to clearly identify components to be put into the production process by the help of printed texts (OCR) or 2D-Codes.
2D-Code.
Under Difficult Conditions.
2D-Decoding is No Problem.
E.G. DATAMATRIX.
2D-decoding such as Datamatrix or QR-Code enable users to distinctly identify the production material.
INTRINSIC REDUNDANCY.
The 2D-code’s intrinsic redundancy supports high reliability also under difficult conditions.
OCR-Recognition. A Classic Method.
Any Text in the Camera’s Range is Clearly Interpreted by the System.
ANY LETTER CAN BE RECOGNIZED.
Our AI-based OCR-System is capable to recognize any cipher, letter or number with the standard main camera as soon these are within its range.
COMFORTABLE FORMAT REQUIREMENTS
A plausibility-checked recognition can thus be realized and subsequently used in the documentation of the production.