Research & Development
AI-based industrial vision systems
Our AI-based industrial vision systems combine Artificial Intelligence (AI) and Machine Vision (MV) technologies. Our AI-based vision systems enable various industrial equipment to autonomously “see” and analyse the product being manufactured and then take certain actions based on what they see. Industrial equipment using our machine vision systems can also operate autonomously without human intervention, effectively increasing human safety and quality control (QC) in manufacturing and logistics processes.
INDUSTRIAL AUTOMATION AND QUALITY CONTROL (QC)
Automation and monitoring
According to the current state of the science, there are several promising teachable object recognition algorithms that can recognize objects with complex shapes and patterns, but almost all of them have a training complexity that increases in proportion to the number of objects to be recognized.
On the other hand, well-trained systems are highly successful and can be used effectively in industry. The most prominent example being sophisticated systems operating in self-driving cars that can recognise cars, trucks, lorries, other various vehicles, road signs, road markings, etc.
Using some of the results of currently known generic object recognition databases such as PASCAL VOC, ImageNet, MS COCO or OPEN Images, we can build a custom database for our customers that can recognise previously learned objects, eg. metal products and parts, which stored in the database.
Our AI-based industrial vision systems combine the most efficient clustering algorithms and various convolutional neural network (CNN, R-CNN) algorithms, which enable our systems to efficiently and automatically recognise manufactured products as well as parts in warehouse stockpiles. Our systems are able to perform automated quality control (QC) on manufactured and stored parts and select defective ones.
Our AI-based industrial vision systems can automatically perform the following key steps on images of various parts and components:
(1) Object Classification
(2) Generic Object Detection
(3) Semantic Segmentation
(4) Object Instance Segmentation
(5) Object Instance Counting
The goal of our current research is to enable our AI-based industrial vision systems to be able to autonomously carry out stocktaking based on images only, suited for future warehousing applications.
Combination of image recognition and clustering algorithms
AI systems tailored to the needs of our clients
Automation of manufacturing and logistics processes
Machine Vision based automated quality control (MV-QC)
INDUSTRIAL SECURITY AND ACCESS CONTROL SYSTEMS
Industrial security systems
According to the current state of science, machine learning systems based on convolutional neural networks (CNNs) are an excellent way to identify information in images. However, the disadvantage of image analytics based on simple convolutional neural networks is that they can only make a statement about the whole image, and therefore cannot identify the various occupational safety and health (OSH) devices that are visible at the same time in a given image.
A solution for identifying multiple objects that need to be recognized in a given image could be the use of region-based convolutional neural networks. Based on current knowledge, our researchers hypothesize that the most promising and efficient solution for the detection of multiple objects in a given image can be achieved by further developing Region-based Convolutional Neural Networks (R-CNN) as well as its enhanced versions, Fast R-CNN algorithms.
The main advantage of these algorithms over classical CNN algorithms is their ability to split the original image into either a predefined number (typically close to 2000) or a dynamically defined number of distinct regions of interest (ROIs), which allow the unique identification of different image elements within the given image.
Use-cases related to our industrial security systems and biometric access control systems:
(1) Biometric access control based on facial recognition.
(2) Access control systems supported by machine learning based image recognition, which can verify whether the employee is equipped with the necessary protective equipment to enter the protected area.
(3) Facility Management (FM) Systems – Machine vision assisted security systems with virtual fencing and unauthorized intrusion detection.
CNN, R-CNN and Fast R-CNN based image analysis
Detection of an arbitrary number of objects in an image
Industrial biometric access control systems
Checking protective equipment with machine vision systems
Facility Management (FM) and security systems
Research and Development with Trilobita
Phases of our R&D projects
R&D Project planning
In the R&D project planning phase, we help our clients find the most optimal use of resources. We prepare the financial and technical design of the project and prepare the proposal for the selected funding scheme.
Applied research
In the applied research phase, we prepare the necessary research plans. We carry out and document the series of experiments based on our research methodology.
Evaluation of research results
We evaluate the results of the series of experiments using various data analysis methods and prepare the research summary document.
System Planning
Based on the research results, we design the systems for our customers. We use our own system design methodology and tools for the planning.
Development and testing
Our development methodology combines elements of classic waterfall and agile methodologies, flexibly adapting to the needs of the given client and project. The efficiency of our development and testing work is further enhanced by a number of our already tested, ready-to-use system modules.
Support
After the completion of our R&D projects, we always provide follow-up and support services to our customers for the solutions we have delivered. Our goal is to establish successful, long-term partnership with our clients.
We believe that every hour spent on design pays off many times over in the implementation and roll-out of our systems. Our ergonomically designed user interfaces provide our customers with a new user experience and ease of use.