Skilltellect are specialists in the development of computer vision systems for a variety of applications related to industrial automation, particularly for use on conveyor and technological lines.
FEATURES AND ADVANTAGES: Our own software based on trained neural networks and developed computer vision algorithms Use of easily available and inexpensive cameras. All calculations are performed on local servers (computers) installed at the customer's premises or on Skilltellect clouds/servers.
SOLVED TASKS
Quality control: culling, defect inspection
Skilltellect computer vision allows to perform single-type production tasks, such as quality control of manufactured products: accurately and quickly rejecting substandard products, defects and deviations from specified parameters. Unlike humans, computer vision system does not get tired and can work without breaks.
Sorting and product type identification
Our vision system accurately identifies the product type according to the required parameters (shape, color, size, etc.) even among chaotically arranged items.
Counting items
One of the main advantages of a vision system is its ability to perform routine tasks quickly, uninterruptedly and without error, such as the ability to determine the number of items on a conveyor line or in a container.
Determining the position and coordinates of the item on the conveyor line and working area
The Skilltellect inspection system due to the software developed by our team, is able to accurately determine the position and coordinates of the item, which allows the customer's actuator to perform any manipulations with the object.
Laying out on the line from the bulk
The vision system allows to distinguish chaotically lying objects and receive coordinates for their capture by a robot or other actuating mechanism. Then the object is captured and transferred to the conveyor line according to the given coordinates.
Reading labels and markings (optical character recognition)
We adapt computer vision for a number of tasks related to reading labels and markings. In particular, the system checks the presence of markings, readability of symbols. It can also check the correctness of the specified date, etc.
Measurement of object dimensions
Measurement of the parameters of various objects, especially complex shapes, is largely more efficient with an automated system and provides the required level of accuracy, according to specified criteria.
Computer vision for training cells and study rooms
The Skilltellect computer vision systems are able to act as a visual learning equipment for students to solve theoretical and practical tasks related to detection, recognition, measurement and visualization of objects from the image - depending on the needs of the course.
Container filling control
Automation of routine processes facilitate qualitative visual control of container filling, including assess of appearance (e.g. water bottles filling control or candy box stacking control).
Computer vision systems for industrial robots and cobots
Our computer vision systems are often used on projects involving industrial robot manipulators and cobots. It performs the same tasks as described above including robots as the actuators.
The Skilltellect team has trained its own neural network models and developed optimal machine vision algorithms that eliminate the need for precise object positioning, recognize and provide coordinates for complex and non-typical items such as:
- transparent objects
- shiny objects
- objects of complex shape
- small-sized objects (less than 10 mm)
FEATURES OF OUR SOLUTIONS: Use of our own laser scanner (the scanner is fully developed by Skilltellect). Working with 2D and 3D cameras including selection of optics for the task, based on the customer's needs. Working with permanently fixed cameras and cameras installed directly on the robot. Our system has an API that makes it easy to communicate with it. Output of object coordinates in the coordinates of the robot. Handle unstable object sizes (or positions). Cloud and local solution.
