Research On High Resolution And Three-dimensional Image Data Fusion Algorithms Based On Compound Eye Model

Nowadays, with the development of science and technology, more and morepeople pay attention to three-dimensional imaging which is based on the TOF (time offlight) principle. At present, these three-dimensional cameras are mainly used inpilotless automobile systems, robot navigation and mapping systems, space roboticsystems and so on. In view of the development of these three-dimensional cameras inthe field of high technology, we need to explore and develop their performances andtechniques continually. Only in this way can we use them more efficiently andproficiently in the future, and also use them to achieve a multiplier effect in thefuture.In this paper, firstly, we introduced SR-3000camera’s parameters, hardwarecomponent modules, imaging principle, and some problems in SR-3000cameras weuse. Secondly, we briefly described a2D/3D imaging system, and adjusted the systemby a number of optimizations. Thirdly, we used the traditional camera calibrationmethod for the SR-3000camera. But different to it, based on the observation andanalysis of the acquired data from the camera, we proposed a new method to acquiresome parameters of the camera, and then we made some comparisons and analysis ofthe parameters from the two methods. Fourthly, we provided a way of coordinatetransformation and visualization improvement of the SR-3000camera data, which isconvenient for us to understand the images and extract the features. At last, wedetailed two methods of data fusion for the acquired2-dimensional and3-dimensionalimage data of the system: nearest neighbor method and triangulation tessellationmethod, and then analyzed the experimental results of these two methods.