AI vision system for robot guidance
K-ROBOT
Modular, Simple and Performing
K-ROBOT is the robot-guided vision system that uses software and cameras to locate the position of components to be processed and then guide the robot in its movements.
Thanks to the high precision of the coordination between robot and machine vision system, it can be applied for different needs: picking from untidy warehouse (bin picking), pick and place, palletising and depalletising, picking from flatbed, assembly, start-up, riveting or welding, etc.
The wide variety of software packages and cameras that can be integrated allow K-ROBOT to adapt to all kinds of production requirements. Moreover, it can be integrated directly on the robot flange of the main brands, or fixed inside the robot cell or island, using all degrees of freedom to perform gripping and assembly.
Easily integrated with any PLC via established communication protocols.
The scalable architecture allows the realisation of the simplest applications up to the most complex vision systems, using a single programming language. K-ROBOT uses our K-VISION vision platform. K-ROBOT utilizza la nostra piattaforma di visione K-VISION.
Localisation
K-ROBOT uses 2D and 3D reconstruction sensors and a powerful 2D and 3D localisation algorithm that does not require contrasting backgrounds with respect to objects and enables localisation even of reflective objects.
Gripping Accuracy
Gripping accuracy is guaranteed by the different types of calibration: fast and immediate or very accurate.
The system is capable of detecting and picking up small objects and has a built-in collision avoidance function, as well as the ability to recognise and pick up different objects randomly arranged together.
Fast learning
K-ROBOT’s algorithm adopts innovative machine and deep learning based on neural networks (Convolution Neural Network) to increase the robustness of the application under different environmental conditions and points of view.
The recognition and localisation of an object takes less than 0.1 second, allowing robots to pick up an object without having to wait for point clouds to be calculated.
The algorithm adapts to each situation, fully customisable to the application.
Ease of use
Fully configurable via an easy-to-use operator panel, K-Robot allows total control of the vision system and a complete view of the analysis in progress, as well as real-time results.
The system is eligible for Industry 4.0 and 5.0 funding.
