Research On Feature Analysis And Applications Of LADAR Imaging System

Due to the complex environmental adaptability, anti-interference ability, large information accessibility, and other characteristics, laser radar imaging (LADAR) system is widely used in battlefield reconnaissance and target early warning and tracking, and in vehicle aided, artificial vision, surveying and mapping. Feature prediction and performance evaluation of a LADAR system is essential for system development and algorithm validation. Since a simulated LADAR system have features like high reliability, non-destructibility, flexible control and no limitation to the weather and location, research on feature analysis and performance evaluation of a digital simulation is very important. Besides, a digital simulation can also reduce the downside of field experiments, at the same time save time and money.In order to analyze the features and evaluate the performance of a LADAR system, this thesis focuses on the design of a general LADAR simulation framework, which contains dynamic scene modeling and rendering, static scene modeling and atmospheric turbulence effect, Geiger-mode APD (GMAPD) imaging LADAR system modeling and performance evaluation. The thesis is organized as follows:The first section describes the design of a LADAR simulation framework. Imaging mode and detection principle of LADAR system are analyzed in detail, and the proven design modules and guidelines for LADAR simulation system are studied. According to the simulation guidelines, a general and systematic physically reasonable imaging LADAR simulation model combining "laser-target-background-atmosphere-LADAR imaging-point cloud reconstruction" is achieved based on the theory of LADAR imaging system.The second section depicts the methods of rending a dynamic laser radiation scene. Since simulating a LADAR scene in real time is difficult due to the enormous calculation complexity when calculating the speckle of reflection, we analyze the reflection characteristics of the target/background, transmission characteristics of laser in the space, and the technology of GPU-based rendering in detail. Considering the coherence, a method to calculate the laser radiation scene in real time is proposed by redefining laser radar expression based on a BRDF model and developing a GPU parallel processing algorithm. It achieves high performance when computing the energy distribution into the receiver. The feasibility of the proposed method is verified by comparing simulated and field data.The third section presents a new rendering model for static laser scenes. In order to improve the accuracy of laser propagation, basic principles of ray tracing are studied, and the feasibility of ray tracing technology applied to LADAR simulation is analyzed. To accurately describe multiple reflections of the light, semitransparent and shelter phenomena, ray tracing is introduced to modify a classical illumination model for simulating a LADAR system. To improve the accuracy of atmospheric effects in the LADAR simulator, exponential Weibull model is adopted to calculate atmospheric turbulence, achieving a physically-based simulation of a LADAR system integrated with quantitative atmospheric turbulence. Experimental results show that the turbulence can cause energy dispersion, leading to the detection of false alarm.The fourth section investigates the GMAPD-based imaging LADAR system model and evaluated its performance. By calculating energy distribution in a three-dimensional laser scene, Geiger-mode APD detection principle is analyzed in detail, and a GMAPD detection model is established. In order to improve the accuracy of the registration model in multi-carrier, a new method of point cloud registration is proposed based on rebuilding distance formula. Combining local and global 3D coordinate transformation model, the LADAR point cloud reconstruction in multi-carrier is implemented. The method has been verified by simulating the GMAPD imaging LADAR system in airborne platform. Finally, the LADAR imaging performance is evaluated with different system parameters in various environments.