Technologies

METRICAM TECHNOLOGIES

UNLIMITED POSSIBILITIES TO ADAPT TO ANY SITUATION

At Metricam we develop several core technologies which can adapt to the majority of business challenges. Whether it is about detection, recognition, classification, quality control or about any manipulations and movements in 3D space, we probably have solution. Here are our our core technologies in a glance:

DEEP LEARNING

Deep Neural Networks provide modern world with almost limitless possibilities to solve problems. We explore more specifically the class of image classifications. Our work is based on TensorFlow as a core technology. Tensorflow is feeding with data the next level of our applications which determines the desired system behavior. Our recent work is an application for gender and age recognition. Based on network output data, specific content is provided to the end user.

And this is just one example. Another extension is towards data collection for statistical purposes, and many more. Are you working in the area of marketing? - then this could be the right solution for you. Or if you are in the area of process automation, perhaps you search will lead you to the same place.

FLEXIBLE 3D
FOR PRODUCTION LINES

Traditional way of designing production lines with fixed positions and cycles is now obsolete. We believe it has to be flexible, reconfigurable, ant it does not have to be precisely build. How is that possible?

In all camera based automation systems, there exist three parts that move and operate in their own coordinate system - the production objects, the camera, and the robot manipulator.

Classic automation system solve this problem by trying to put everything in a one coordinate system. That inevitably makes the system very sensitive to geometric variations. All objects, and materials must be placed in specific predefined places.

Our tecchnology makes the difference. Using markers or visually detected features, we are able to combine these three coordinate systems and make things work together. Now it is easy to imagine that working environment starts looking in a more natural way - a camera is observing the working space, objects are fed somewhere in a suitable place, object manipulation is done and the object are placed where output bin or track is detected to be,

Imagine all the frexibility you have now at your disposal!

OBJECT DETECTION

Objects detection and recognition are well-know subjects in the field of industrial automation.

For objects with known geometry, it is also possible to estimate distance and pose in the scene. This enables robots to work with the objects without the need of fixed positions which leads to great flexibility in organizing the working space. Objects can be seen from different angles and distances. In other words, automation systems can now be really versatile and reconfigurable. This feature could be combined with our system for flexible 3D robot steering, or it could be integrated with your own system for robot steering as long as it can interpret distance and rotation data in real time. Typical applications are listed below.

Assembly Control - two or more details are detected and their orientation in 3D-space is checked against predefined specification. This could form an intermediate assembly check-point or final check of production modules.
Feature Control - Multiple parameters are defined, like shape and dimensions. These can be checked again specification data. Depending on the particular configuration, often these can be checked without specific mechanics to ensure proper orientation of the objecs. In most cases our RobotEye system handles these situations.
Sorting - based on dimension, shape, color or other properties, objects can be sorted in different categories for further processing.