Abstract
In this master’s thesis, a test framework was developed that is used for testing imageprocessing hardware-software-applications. During the development process, a specialfocus was placed on the Connectivity Control Unit 3, CCU3 for short. The framework
enables the automated testing of image and text content, which is crucial for ensuring
the functionality and validation of software updates in safety-critical systems.
The structure of the test framework is based on a modular architecture that allows flexible and efficient integration of various test methods. A central feature of the framework
is the use of gRPC for communication between the test components and the application
under test. This enables a robust, scalable and easily expandable infrastructure for data
exchange and the execution of test cases.
As part of the implementation, various approaches for image and text-based testing
were analyzed, including SIFT and template matching. Also, EasyOCR was integrated
to ensure precise text recognition in the test processes. These methods were tested on
sample images to evaluate their effectiveness and accuracy.
A particular advantage of the developed test framework is its ability to being usable
even with closed interfaces. By working exclusively with image streams, the production
software can be tested despite security restrictions. This is particularly relevant as conventional test tools based on open interfaces can’t be used in the production version.
The test framework was successfully implemented for the CCU3 and shows great potential for use in other projects of the partner company. The framework can contribute to
quality assurance, particularly when testing dashboards and other image-processing systems, provided that the specified minimum requirements are met. Future developments
aim to further increase modularity and flexibility in order to make the test framework
widely applicable. Through continuous adjustments and optimizations, the test framework can contribute to ensuring the quality and reliability of series products in the long
term.
| Date of Award | 2024 |
|---|---|
| Original language | German (Austria) |
| Supervisor | Herwig Mayr (Supervisor) & Martin Reitsberger (Supervisor) |