Research On Technologies Of Infrared Image Quality Improvement

Infrared imaging system has been widely used in military and civil commercial fields due to its own passive imaging characteristics and the ability to work all day.However,infrared focal plane array has inherent non-uniformity due to the limitations of manufacturing level and device material,which seriously affects the quality of infrared image.At the same time,due to the mechanism of infrared imaging,the infrared images with poor contrast and vague details will have a great impact on subsequent target detection and recognition.Therefore,in view of the research of infrared image quality improvement technology,this paper studies two technical solutions which includes infrared image non-uniformity correction and infrared image enhancement,and propose improved algorithms.Nonuniformity correction technology based on neural network can adjust the gain parameters and bias parameters simultaneously according to the scene changes.It has the superiorities of high precision and high adaptability.In this paper,the traditional neural network nonuniformity correction is deeply studied,and a correction algorithm based on trilateral filtering and neural network is proposed.Aiming at the key problems of the traditional neural network algorithm,the algorithm is improved from two aspects,namely,the expectation graph and the updating of learning rate.First,the fast trilateral filter is used to replace the neighborhood filter in the traditional algorithm to accurately acquire the desired image.Then,combining the neighborhood difference and the inter frame movement,the learning rate can be updated adaptively with changing scenes,which can be effective.In short,the algorithm in this paper can improve problems such as ghosting artifacts and blurring,and achieving better non-uniformity correction effect.Although nonuniformity correction can remove streak noise from infrared images,it does not solve the problem of poor contrast and blurred details of infrared images.Therefore,image enhancement is a key approach to enhance the quality of infrared images.After studying many current infrared enhancement algorithms,an image enhancement algorithm based on layered difference representation and S curves is gived.The proposed algorithm first uses guided filtering to separate the infrared image into details and background information,and then processes the background information in a layered difference representation to increase the image contrast and improves the visual effect of the image.In the meantime,the detail information is enhanced by the S curve principle.Finally,the background information and the details information are fused linearly.The image contrast is improved while the details information is saved,and the quality of the infrared image is significantly improved.Through subjective experiments and objective experiments,it is indicated that the advised algorithms can effectively correct infrared non-uniformity and enhance the visual effect of infrared images,in order to improve the quality of infrared images.