Bachelor Thesis
Development of a Lightweight Miniature Camera with On-Board Image Processing
Vision is an important source of sensory input for autonomous mobile robotics. In most current robotic systems, standard video cameras transmit captured images to a PC for image processing, e.g. through FireWire, USB, or possibly wireless. Continuously increasing autonomy and increasing number of cooperating robots for a given task often requires rudimentary on-board vision processing, as transmission bandwidth and latency in such systems get hard to handle. However, a miniature and lightweight yet powerful stand-alone image-processing-system is not currently available.
For this thesis the student will design, build, and program a stand-alone vision system composed of a tiny CMOS camera module as is typically found in mobile telephones (around 5x5mm), an SRAM module for buffering captured images, and a microprocessor to perform simple vision algorithms in an embedded system (such as blob-tracking). Particular attention should be given to size and weight constraints, as the vision system shall be employed on miniature flying quadro-copters. We have designed a similar system some years ago which shall serve as a reference; however, current technology has advanced dramatically, which requires a re-design with significantly improved features.
- Investigate functionality of miniature CMOS camera modules
- Evaluation of CMOS camera modules, SRAM chips and possibly microcontrollers
- Design system composed of CMOS camera, SRAM and microcontroller
- Circuit-Board design and assembly of such a system
- Basic software development for camera system (in C for microcontroller) and host PC
This thesis requires familiarity with electronic components (e.g. microcontrollers, SRAM), experience in circuit board design (Eagle), and programming skills in C (possibly Java/PC).
This work is part of a new student-project started at Fachbereich NST that investigates in swarms of miniaturized autonomous flying vehicles. The miniature camera shall ultimately be used on-board of flying quadro-copters to detect salient objects (as part of an international UAV-competition).