|Thesis abstract: |
First image acquisitions introduced in the first Decades of the XIX Century were based upon chemical reactions, were very slow and the results were not so satisfactory in terms of stability of the process and image quality. Moreover, photographic (or lithographic) techniques were not as widespread as we could think nowadays. But something new was created, one of the greatest technological revolutions of all time, if not equal to the birth of electronics and computers. Imaging technology has made great strides since it¿s born. Once the process became very stable, the photograph started to be more and more widespread. Introduction of motion in the acquisition, color, instant printing photographs, and other techniques succeed one another. Once the electronics met the photography (or better, the image acquisition) the world changed again and the humans started to communicate with images.
Meanwhile, imaging acquisition techniques became so accurate that they could be used in research applications, for example in microscopy. The birth of robotics led another great development in imaging research field because new autonomous applications (as for example robotics movement in an unknown area) could be developed. First time-of-flight cameras were developed, based on the distance-to-time conversion. Assuming a photon travelling between the object and the camera, it is possible to measure the distance of the objects pixel by pixel within the image. This kind of systems (based on active illumination) is faster and simpler compared to the others; moreover, the lower required computing power allows the implementation of a high number of new applications, for example medical aids, autonomous improved surveillance of sensible areas and road safety. While for the former the depth range is very short (up to two meters), for the other two applications it can be much higher, this opening many big issues. Moreover the high speed (up to 100 fps) and the great change in the ambient light conditions (from night-time to very bright daytime) require an accurate analysis, study and development of a complete camera solution. The development of a complete camera starts with the analysis and the improvement of existing range-meter algorithms, continues defining the performances required for a specific application and the requirements for the sensor used in the application, and concludes with the analysis of the performances of the new camera and the study of the performances in the application field. The measurement algorithms are presented and deeply analyzed, with comparison between methods and analysis of optics, geometrical dependence, light illumination, different light and background sources (a crucial problem for this kind of applications). The activity of this Thesis work is focused within an international European Project, which aims to develop a complete application-oriented 3D camera. All the implementation chain is presented, starting from hardware, through firmware and software, towards characterizations and real-scenario measurements. A compact camera (12×6.5×5 cm3) with external high-performance LASER-based illumination (10 W peak power with hundreds of nanoseconds on-time and repetition rate up to 1.7 MHz) is presented. The camera provides C-type lens mount, is USB-bus-powered and fully PC controlled. The system is able to acquire 3D movies with a frame rate up to 100 fps, while simultaneously acquiring also 2D movies up to 32 kfps. The illuminator is directly controlled by the camera, in order to guarantee proper synchronization. The camera was tested in indoor controlled conditions, providing very good performances. i.e. accuracy better than 0.2 m and precision better than 0.1 m.