Detect
Locate presence, absence, position or a relevant visual condition.
Computer vision / production
Seeing is not enough. It must become a decision.
We design computer vision systems that detect, count, measure and verify within industrial processes—connecting the result to the software that must act.
Explore an application↗01 / Applied vision
A camera is only the beginning.
The engineering lies in controlling capture, defining criteria, handling uncertainty and delivering an output the process can use. A solution may combine classical vision, machine learning, rules and human validation.
02 / Real product
From capture to model, inside one platform.
A product developed by Appkadia for a real industrial environment. It brings together image capture, dataset preparation, manual and AI-assisted annotation, training, testing and model execution. This is not a demo: it is operational software built around the process.
Real product screenshot. Identity and operational identifiers have been anonymised.
Pixel-level separation of the object and its components.
Reviewable classes and masks within a single image.
03 / Applications
Check what matters, at the right moment.
Count
Turn an image sequence into traceable quantities and events.
Measure
Extract dimensions or relationships when optics and environment allow it.
Verify
Compare the result against defined criteria and route it into the right workflow.
04 / Imaging technology
The camera is part of the algorithm too.
2D area-scan
Complete frames for presence, position, counting, reading and inspection of static or synchronised parts.
Line-scan
Builds an image one line at a time for continuous materials, webs, surfaces and high-speed processes.
3D and profiling
Height, volume, shape and deviations through laser, stereo or time of flight, depending on the required precision.
Spectral and thermal
Information beyond RGB to distinguish materials, temperature, moisture or defects the human eye cannot perceive.
05 / The difference
Vision ends inside the process.
We integrate the result with MES, IoT, alerts, histories and interfaces. Detection stops being an isolated proof of concept and becomes part of the system.
VISION · MES · IOT · AI