Research And Application Of Flight Visual Simulation

Information visualization has become a development trend of application with the development of simulation, virtual reality, communication and the corresponding technologies. Virtual visualization of flight simulation has been a hot topic with time going.In this thesis, on the basis of reading a lot of references about the flight simulation and virtual visualization, the visualization modeling and dynamical driving of cockpit instruments,terrain creation of large area, visualization design of scene and aircraft location and gesture’s keyboard control are emphases in the research of flight simulation of three-dimensional visualization system.Firstly, on the basis of studying the common modeling techniques as well as model database optimization methods of Creator, include Levels of Detail technique, Degree of Freedom technique and some other technique, and the cockpit instrumentation and the large area terrains has been modeled.Secondly, by the use of visual simulation tools Vega Prime software, a three-channel virtual scene for the flight simulation of three-dimensional visualization system is created,and the collision detection and special effects as well as the house’s status change before and after the collision are added to enhance the realism of the simulation system to protect the real thing.To study on the integer-level edge detection operator, and introduce classical edge detection operators. The simulation of classical edge detection operators are proceed based on the principle, comparing the simulation results, then choose Canny edge detection algorithm.In allusion to the limitation of Canny algorithm that parameters is fixed, a new method of scale-adaptive Gaussian filter is proposed as the improved Canny algorithm, and the parameter can be changed with the image. Verify the feasibility of the improved algorithm for each different type of image information, based on the principle that the filtering scale factor is increasing proportional to the filtering frequency. Verify the superiority of improved Canny algorithm, by comparing the detection result of classical and improved Canny.The image magnification is divided into internal image interpolation and edge image interpolation. This thesis introduces the traditional linear interpolation and non-linear interpolation algorithm. Analyzing their characters and aim at the smoothness of internal image, the bilinear interpolation algorithm is selected to magnify the internal image.Traditional sub-pixel edge detection algorithm using integer gray values as calculated template should not be used because the location and gray of sub-pixel are changed at the same time, so this thesis selects edge interpolation algorithm based on error back-propagation neural network. Low-resolution unit is defined, the image information is classified. Using sub-pixel edge points as training data to resize the weights detecting by the back-propagation learning algorithm adaptively, the edge function is achieved. The quality evaluation system of image segmentation is established, analyze the quality of the image from subjective visual perception and objective data. Simulate the image magnification algorithm proposed in this thesis and calculate the average gradient value to validate the correction of the new algorithm.Based on analying visual imaging system, this thesis uses Vega Prime as the developed tool, and the new algorithm has been used in flight simulatin successfully.