What We Do: improve line speed & inspection accuracy
We deliver improvements in line speed and inspection accuracy - learn more. Our physical deliverable is a turn-key visual inspection system, customized for the client's needs, integrated and tested on-site with performance guarantees. The core functions are image acquisition with cameras, optics and lighting, image processing hardware & software and interactions with surrounding systems. We also deliver peripheral functionality including user interfaces, data logging & analysis and supporting automation.
1. Image Acquisition: Cameras, Optics & Lighting
We work with top industrial camera vendors such as Cognex and Basler to select the most appropriate components for the client's application. These range from hardened cameras for harsh environments to miniaturized devices for embedded vision applications. Optics refers to the selection and configuration of lenses to achieve optimum image contrast, resolution and depth of field. Lighting can make or break a visual inspection system. Correct selection of wavelength and positioning of illumination devices can save considerable time and cost when processing images downstream. When selecting components, our objective is to make the appropriate trade-offs between cost and performance to match the client's needs.
2. Image Processing Hardware & Software
Once an image is captured, image processing refers to the extraction of useful information from the image, for example component dimensions and presence of defects. This is the brain of the system which assesses incoming images and takes decisions based on their content. To minimize cost and lead-time, we lever image processing tools such as Cognex VisionPro & ViDi, Matrox MIL and MATLAB. Where highly customized solutions are required, we work with lower-level libraries including OpenCV, TensorFlow and PyTorch.
Artificial Intelligence, specifically Machine Learning and Deep Learning, enables the introduction of human-like judgement to automated visual inspection. We apply Deep Learning techniques where appropriate to complement traditional Computer Vision techniques.
3. Interactions with Surrounding Systems
Visual inspection systems usually take action based on the image processing output, for example directing defective parts into reject bins. Visual inspection systems therefore communicate with surrounding infrastructure such as Programmable Logic Controllers and industrial robots. When a new visual inspection system is installed, clients understandably want to maintain their existing systems. We work with all industrial communication protocols in order to minimize the impact on existing client systems.
4. User Interfaces
Visual inspection system user interfaces provide a means of configuring parameters such as decision thresholds and product types, and visualizing image processing decisions and defects. We build custom user interfaces to provide the functionality desired by the client. To accelerate development, we lever external frameworks including Qt, JavaFX and Microsoft WPF.
5. Logging & Analytics
Logging captures visual inspection data, which can be required for regulatory compliance. We provide both on-premises and cloud-based logging solutions. Analytics is the assessment of visual inspection data to identify trends and drive improvements. We use external analytics tools such as Power BI and Tableau, levering Machine Learning frameworks where necessary for more complex analysis.
6. Supporting Automation
Visual inspections systems work with mechanical and electrical automation to manipulate products. To minimize cost when installing a new vision system, we lever the existing automation equipment as far as possible. Where new automation equipment is required, we use off-the-shelf components for conveyors, sensors and actuators. Depending on the scale of the supporting automation, we may work with a general automation integrator to help us best serve our clients' needs.
Cut Operational Costs
Contact us to cut operational costs through automated visual inspection