Software integration · Visual recognition
Connecting software, recognition and parking equipment
In a parking project, we worked on the connections between the operator application, kiosk, number plate recognition and local equipment. Our contribution includes integrating existing components, handling events and verifying specific changes in the installed system.
The context
The project includes an administration application, a payment kiosk and a local component that communicates with cameras, the card terminal and access equipment. Number plate recognition supplies data to the application workflow. Operators need consistent session and payment states even when information comes from different components.
The problem
A recognition result alone does not make the complete workflow function. The application must receive the event, relate it to the session state and communicate with equipment through available interfaces. Repeated events, an unavailable component and differences between the local and cloud applications need explicit handling.
Our contribution
- We worked on the kiosk and administration application and the component connecting the software workflow to cameras, the payment terminal and access equipment.
- We integrated an existing recognition engine: receiving captures, processing the result and forwarding the event to the application, with handling for repeated reads.
- We strengthened restrictions on changes to selected operational settings, with checks in both the interface and the data layer.
- We verified changes to the kiosk and installed versions, with documented rollback steps and checks after updates.
How the work was checked
The adapter code documents the path from capture to the event received by the application. Access checks covered permitted and rejected roles; data probes were rolled back after testing. For a kiosk intervention, we checked normal operation and recovery after a controlled disruption. These checks concern the components and interventions described, rather than measuring recognition accuracy under every condition.
What this means for similar projects
For a similar project, start with the existing equipment, available interfaces and the complete path of an event. Define separately what makes a recognition result acceptable, who reviews exceptions and which actions the application may take. The same separation is useful when integrating an AI result into a business process.
Need to connect an application with recognition or equipment?
Describe the workflow, existing systems and available APIs or equipment documentation. We can assess the interfaces, exception handling and a first testable integration before estimating the work.
