Computer vision earns its value when it stays stable in real conditions, with varying light, angle and quality. From plate recognition in CarnoWash to object detection in satellite imagery with YOLO, data quality, field testing and integration with the operational system decide the result.
01Real conditions, not the lab
A model that is accurate on clean data may drop under weak light or a poor angle. Testing in the real environment is needed from the start.
02Plate recognition in car wash operations
In CarnoWash, automatic plate capture makes vehicle intake faster and less error-prone, and connects to service tracking, cost and invoicing.
03Satellite imagery and YOLO
In the satellite image processing project, object detection with YOLO and an orderly data pipeline make analysing large volumes of imagery possible. Details are in the projects section.
Frequently asked questions
The variety of real conditions such as light, angle and image quality, and data preparation.
OpenCV, YOLO and Python, with containerized deployment for operational use.