Research And Implementation Of Infrared Array Image Restoration Technology For Shipborne Platform

Rotation and fuzzy infrared array images take certain effects to command center operators and subsequent image processing of dim target detection and tracking,due to Shipborne infrared tracking system is disturbanced by sea waves and winds.With more and more image data and more complex image processing algorithms,higher performance of image processing platforms are requires.In this paper,based on the problems of above three aspects,research is made on infrared image racemic,infrared image deblurring and high performance Signal processing board and implementation of restoration algorithm.The main research work is as follows:1)?The infrared image rotation angle is derived for shipborne platform.This paper analyzes the reason of infrared tracking system rotation,and defines the rotation angle of the infrared image images from space geometry.Based on the five angles of yaw angle,pitch angle,roll angle,azimuth angle and elevation angle,and physics knowledge about the rotational motion of a rigid body with a fixed point,derive infrared image rotation angle using quaternion combined with spatial analytical geometry knowledge.Finally,based on the rotation angle,racemic and restore the stable images.This alleviates the degree of shaking of the infrared images and achieves the goal of electronic image stabilization.2)?The motion length and direction angle of linear motion degradation function is derived.Analyse and define the geomethical relationship of infrared array image sequences.Based on Quaternions and Spatial analytic geometry,get the angle difference of LOS and direction angle of motion lengt.According to the integral time of infrared array detector,instantaneous field of view and image frame frequency,motion length is derived.The advantage of the derivation method is in any direction,but not limited to a specific direction of the movement.3)?Take the linear motion degradation function as the initial value of Expectation-Maximization restoration algorithm,after finite iteration of the E step and M step,get more precise linear motion degradation function.This method improves definition of the restored infrared image,and is verified by experiments.4)?In view of the ringing effect produced in the image restoration,put forward the circular boundary method to suppress boundary ringing effect caused by boundary truncation;The improved constrained Self-adaptive Image Restoration Algorithm is used to restore the image and restrain ripple caused by edge ringing effect.The experiments results show that the the proposed method effectively suppresses the ringing effect and highlights the image detail,and the image restoration effect is clearer.5)?Around two aspects of image data fast transmission and complex image processing algorithms,developed a high performance Signal processing board which take TMS320C6678 DSP chip as the main processing chip,and Migrate image racemic algorithm and image restoration algorithm to high performance computing DSP.The migration which considered fully the calculation model of the parallel technology,load balancing among the eight cores and the inter core communication and synchronization,improves the parallelism and computation efficiency.According to process of image restoration algorithms and computation load of algorithms,design a parallel hybrid model which include data Flow Model and Master/Slave Model.Experiments verify the feasibility of the design of the parallel model,basically achieve the equalization of the cores load,and lay a good foundation for the further optimization of the program and Engineering practice.