About the project
| Project type: | internal project |
| Duration: | 2 years (ongoing) |
Camera sensor technology is now established as the primary sensor technology for monitoring, control and regulation applications in industrial and everyday environments. The importance of small systems in particular, e.g. in mobile applications, is constantly increasing with the rapid technological progress of processor and computer architectures.
However, the processing of data from complex sensors such as camera sensors (but also data from radar sensors, laser scanners, etc.) is very complex and computationally intensive, depending on the requirements and tasks, and poses major challenges for the processor in small embedded systems. This is of course particularly true for stereo camera systems.
For this reason, current embedded product solutions are either still larger (PC-based or PC-like) microprocessor systems or powerful, smaller systems using special processors (DSP, FPGA) are used. Manual code optimization and the complex use of the processor-specific instruction set are also common practice. In many cases, however, this limits the portability, scalability and further development of systems in an unfavorable way, which makes the product development process inefficient and expensive.
This research project is concerned with the development methodology of small embedded image processing systems, in particular for stereo image processing, and primarily aims to implement at least one of the two concepts
Use of model-based development methods and code generation
Adaptation of hardware and SW algorithms to the task at hand
to the task.
On completion of the project, the findings and equipment will be available to define follow-up projects based on this. The focus can be on advancing the design methodology, but also on developing new image processing algorithms and their applications.
