Abstract:Scene-based nonuniformity correction technique for Infrared focal-plane array has been widely concerned as a key technology. However, the existed algorithms are now facing two major problems that is convergence speed and ghosting artifacts. The convergence speed of original constant statistics (CS) method has been demonstrated to be more rapidly than the neural network method but how to reduce ghosting artifacts efficiently is the largest challenge. To solve the ghosting problem, the conventional methods often set a threshold to wipe off the outliers, but the threshold is difficult to choose because it changes complexly for different scene. In this paper, a novel adaptive scene-based nonuniformity correction technique is presented that performs the nonuniformity correction based on CS method. Firstly, an analysis of statistical characteristic in every pixel is taken and the cause of ghosting artifacts is discussed that the underlying distribution does not satisfy the assumptions such as symmetry. For the Gaussian distribution can not describe the statistic property for every pixel's data, a model with mixture distribution is constructed and indicates the different distribution's influence to generate ghosting artifacts. Then, utilizing temporal statistics of infrared image sequences the proposed method applies an alpha-trimmed mean filter to estimate detector parameters instead of the conventional mean filter. The algorithm selects the parameter of the alpha-trimmed mean estimator optimally with minimizing the sample asymptotic variance estimate. Moreover, the alpha-trimmed mean filter is designed to detect the nonsymmetry points and trim out the outlier pixels such as edges or extreme distribution. Finally, the performance of the proposed algorithm is evaluated with infrared image sequences with simulated and real fixed-pattern noise. Compared with other nonuniformity correction techniques, the proposed method inherits the superiority of the CS method that converges rapidly but is more robust and gets little ghosting artifacts. The results of the simulated and the real infrared images experiments show a significantly more reliable ability to compensate for nonuniformity and reducing ghosting artifacts effectively.

An Improved Total Variation Model for Correcting the Stripe Nonuniformity in IRFPA Image

Single Image Stripe Nonuniformity Correction with Gradient-Constrained Optimization Model for Infrared Focal Plane Arrays

Fringe Non-Uniformity Correction Method Based on Gray Difference Estimation

Edge-Aware Unidirectional Total Variation Model For Stripe Non-Uniformity Correction

A Novel Infrared Focal Plane Non-Uniformity Correction Method Based on Co-Occurrence Filter and Adaptive Learning Rate

A Non-Uniformity Correction Algorithm of Infrared Image Sequences Based on Constant-Statistics

Efficient Single Image Stripe Nonuniformity Correction Method for Infrared Focal Plane Arrays

Local Spatial Correlation-Based Stripe Non-Uniformity Correction Algorithm for Single Infrared Images

A New Non-Uniformity Correction Method Based on Unidirectional Variational Model

Correction method for stripe nonuniformity.

A Unidirectional Total Variation and Second-Order Total Variation Model for Destriping of Remote Sensing Images

Parallel Algorithm of Real-Time Infrared Image Restoration Based on Total Variation Theory

Stripe Noise Removal of Remote Sensing Images by Total Variation Regularization and Group Sparsity Constraint

Stripe nonuniformity correction for infrared imaging system based on single image optimization

Total Variation-Based Inter-Frame Infrared Patch-Image Model for Small Target Detection

Removal of Stripe Noise with Spatially Adaptive Unidirectional Total Variation

Robust Approach for Nonuniformity Correction in Infrared Focal Plane Array.

An Improvement for Scene-Based Nonuniformity Correction of Infrared Image Sequences

Total Variation-Based Interframe Infrared Patch-Image Model for Small Target Detection

A New Nonuniformity Correction Algorithm for Infrared Line Scanners

A New Destriping Method Combining Moment Matching and Interpolation in Infrared Images