To solve the question of monocular pose measurement of the non-projected-axisymmetric targets, a Contour Image Length matching method is proposed,firstly,the contour of the targets is simplified as many triangles based on nose to solve the information redundancy and visual occlusion;secondly,the simplified target is projected to the image plane to get virtual image length，the actual image length is extracted simultaneously, then the pitch, yaw and roll angle can be obtained by length matching between virtual image and actual image. The aircraft flight calibration test shows that the precision of the pitch angle, yaw angle and roll angle was 0.9 degrees,1.2 degrees, and 1.5 degrees respectively, which indicated the result is slightly lower than intersection, but can save cost and improve efficiency. Finally, the key factors concerning method error are analyzed; the method can lay an important foundation for monocular pose measurement in range.
With the continuous research of polarization technology, how to apply polarization enhancement to time-varying scene has become a difficult problem at present. This problem is reduced to the fast and accurate acquisition of atmospheric polarization transmission characteristics.This method can obtain polarization transmission parameters by multi-scale particle modeling, laser polarization state measurement and scene polarization information extraction. It can provide a solution for the design of polarization enhancement system by combining the characteristics of various methods. Firstly, the multi-scale particle air transmission is modeled, the composition and transmission law of particles are analyzed, the transmission model is established, and the haze particles are classified by microimaging. Then, the polarization data of the atmosphere are measured by laser transmission, and three kinds of color laser are selected, and the influence of the atmosphere on the polarization state is measured in conjunction with the polarizer. The three colors cover the visible light range, and can get the typical polarization characteristics of atmospheric transmission. Besides, we use camera polarization imaging to extract the polarization state of atmospheric transmission. Through the combination imaging of multi cameras and multi polarizers, polarization transmission data of polarized state parameters and distance are obtained. Finally, based on the combination of the three methods, we analyze the results of atmospheric polarization transmission. The atmospheric propagation model provides the change rule. The laser polarization state can get the typical fast changing transmission data, and the atmospheric background imaging can get the steady state transmission result. Through multi information fusion, the laser polarization state based data and the particle transmission model are combined to obtain the wide band transmission law, and the image extraction results are corrected to realize the atmospheric transmission analysis. Using the technical route proposed in this paper, a prototype of atmospheric polarization transmission analysis system is designed. As the key component of the polarization enhancement imaging device, the integrated design of hardware and software is carried out. It can be applied to the use of polarization imaging in time-varying scenes, which has a guiding significance for improving the effect of polarization imaging.
The photoelectric theodolite is an important scheme to realize the tracking, detection, quantitative measurement and performance evaluation of weapon systems in ordnance test range. With the improvement of stability requirements for target tracking in complex environment, infrared scene simulation with high sense of reality and complex interference has become an indispensable technical way to evaluate the track performance of photoelectric theodolite. And the tail flame is the most important infrared radiation source of the weapon system. The dynamic tail flame with high reality is a key element for the photoelectric theodolite infrared scene simulation and imaging tracking test. In this paper, an infrared simulation method for the full-path tracking of tail flame by photoelectric theodolite is proposed aiming at the faint boundary, irregular, multi-regulated points. In this work, real tail images are employed. Simultaneously, infrared texture conversion technology is used to generate DDS texture for a particle system map. Thus, dynamic real-time tail flame simulation results with high fidelity from the theodolite perspective can be gained in the tracking process.
In this work, a texture mapping scheme with massive data is proposed based on quadtree for multi-level scene modeling. A simplified LOD model is established for the multi-resolution texture data which enhance the fidelity of multilevel target simulation in scene. By predictable appreciation according to viewpoint displacement in real-time simulation, the correspongding nodes and level details of quadtree are confirmed when the dynamical scene is loading. This mean effectively decreases the schedule data in real-time simulation. A validation experiment has been performed to certificate the availabitity and improvement.
In this paper pose measurement refers to flying pose measurement of rigid body including the pitch angle, yaw angel and roll angle. Pose measurement is of vital importance for such items as weapons settings, fault analysis and optimation design. Pose measurement based on optical images has many merits such as intuitive and non-contacted, which is a main method to measure pose currently. According to the parameters used and principle of the algorithms, the existing methods of pose measurement based on optical images are classified systematically and comprehensively for the first time as following: the methods of one station un-using camera’s inner parameters are divided into the feature length ratio method and the direct linear transformation(DLT )method, otherwise they are divided into the perspective n points(PNP)problem and the optical and radar integration method, the axes from planes intersection using two stations extensible to multistation, and model matching applied to one or more stations, and then they are comparatively analyzed .At last combined with practical applications such as one or more stations, have or no model and inner parameters used or unused, some selection and improvement of key points are given practically.