SoftSystem srl
PROCESSING
Traditional processing in vision systems consists of designing a software that It can measure distances, recognize patterns or defects, select a found object, read codes and recognize characters OCRThanks to these characteristics, they are found in various applications such as product quality control, traceability, and robot-guided handling.
We are official partners in the development of industrial vision systems Cognex, but we are also able to use other types of devices such as Falcon.
THE software of vision I knowThey are not designed to meet customer requirements. Vision systems work by analyzing every pixel of the captured image. If a black and white image is taken, the system will work on a grayscale, and will therefore look for a defect or read a string only if these elements are highlighted. For example, I need to read a dark alphanumeric code with the naked eye against a light background. By playing with exposure and lighting, I need to make the code appear as dark as possible and the background as light as possible. This will make it easier for the system to read the resulting code.
During the design phase, software is developed that can use all the tools necessary for the project's implementation.
To ensure the presence/absence of an object, or if you want to analyze the continuity of some products that vary in shape, such as a section of sealant (see the page Sealant Inspection Control), then it is recommended to use the blob tool.
The blob tool divides the pixels in the analysis region into blobs and background, based on the grayscale threshold value, which can be set during setup. This threshold value separates the pixels so that those below the threshold are classified as background and those above the threshold are classified as blobs.
When checking, pixels are examined to determine whether they are connected to adjacent pixels. A pixel in the blob is said to be connected to neighboring pixels if the blob pixel touches another pixel of similar value, always above the threshold, on both sides, above, below, or diagonally. The same applies to pixels determined to be background.
For further information, please consult the page Cognex, our partner in vision systems management.
In the example image, you can see the search for a dark blob in a particular point of the image. The blob must recognize the outline of the gear. Thanks to the settings, we were able to exclude the other dark pixels present in the gear's fill and highlight only the darkest ones.
Another tool useful for finding a reference on the image, or in analyzing the presence/absence of a particular, is the search for the pattern toolIt consists in teaching, during the planning phase, a specific pattern, that is, an object present in the image, which is evaluated for its presence every time an acquisition is performed. This too tool It works with grayscale and in particular the pattern is catalogued based on its contours in the contrast between light pixels and dark pixels, following thresholds that can be set during the settings phase.
In the research phase of the pattern, new images are analyzed where it might be present, and the results are sorted in decreasing order based on similarity to a specified minimum level. All results that meet certain rotation, scale, and match parameters are considered valid.
A recognition pattern can be seen in the photo. During the design phase, the software was trained to detect the gear in the logo. SoftSystem, the result obtained is also able to highlight the position of the pattern in space,
The management software of the artificial vision It can also measure distances or lengths of objects in the image. Areas, heights, and lengths can be measured based on the client's request. The results will be measured in pixels and must be converted, through simple calculations, into the international metric system.
In addition to measurement, you can also search for specific shapes such as lines, calipers, circles, ellipses, etc.
The software It is able to show the found shape on video and catalogue it based on the teaching given during the setup phase.
Here too, by setting specific parameters, it will be possible to classify the results obtained as good or not compliant with the research.
A good example would be finding the outline of a glass tube. If the cutting process is not done correctly, the resulting shape will not match the one you specified during setup.
Whenever we deal with images, even the best setup possible isn't always enough to provide the analysis the client desires. This requires using specific image filters to accurately assess the specific detail being sought.
Image filtering is therefore used to highlight edges, rather than colors, or the contrast between dark and light pixels, so as to be able to work at maximum performance in a production line.
The software Of vision allows the customer to recognize any string he wants, read barcodes, QR Code, and many other types of encrypted codes. The acquired image must highlight what needs to be read or deciphered, ensuring that it is also legible, within the range of the camera's resolution.
By placing the character search region at the desired position, the system is able to read the string SoftSystem, even recognizing the gear as an O, after it is specifically taught to him.
In the world of artificial vision, there is also the possibility of using color cameras, capable of recognizing the color of an object and verifying, once trained, that the objects maintain this type of standard in production.
For more information on Traditional Processing in Vision Systems in SoftSystem, fill in the form contact below.
Fill the form below to request a contact.
One of our officers will answer you as soon as possible.
Thank you.