Nonuniformity Correction Algorithm Research Articles

An evaluation method based on absolute difference to validate the performance of SBNUC algorithms

Scene-based non-uniformity correction (SBNUC) algorithms are an important part of infrared image processing; however, SBNUC algorithms usually cause two defects: (1) ghosting artifacts and (2) over-correction. In this paper, we use the absolute difference based on guided image filter (AD-GF) method to validate the performance of SBNUC algorithms. We obtain a self-separation source using the improved guided image filter to process the input image, and use the self-separation source to obtain the space-high-frequency parts of the input image and the corrected image. Finally, we use the absolute difference between the two space-high-frequency parts as the evaluation result. Based on experimental results, the AD-GF method has better robustness and can validate the performance of SBNUC algorithms even if ghosting artifacts or over-correction occur. Also the AD-GF method can measure how SBNUC algorithms perform in the time domain, it’s an effective evaluation method for SBNUC algorithm.

Guided filter and adaptive learning rate based non-uniformity correction algorithm for infrared focal plane array

Imaging non-uniformity of infrared focal plane array (IRFPA) behaves as fixed-pattern noise superimposed on the image, which affects the imaging quality of infrared system seriously. In scene-based non-uniformity correction methods, the drawbacks of ghosting artifacts and image blurring affect the sensitivity of the IRFPA imaging system seriously and decrease the image quality visibly. This paper proposes an improved neural network non-uniformity correction method with adaptive learning rate. On the one hand, using guided filter, the proposed algorithm decreases the effect of ghosting artifacts. On the other hand, due to the inappropriate learning rate is the main reason of image blurring, the proposed algorithm utilizes an adaptive learning rate with a temporal domain factor to eliminate the effect of image blurring. In short, the proposed algorithm combines the merits of the guided filter and the adaptive learning rate. Several real and simulated infrared image sequences are utilized to verify the performance of the proposed algorithm. The experiment results indicate that the proposed algorithm can not only reduce the non-uniformity with less ghosting artifacts but also overcome the problems of image blurring in static areas.

A registration based nonuniformity correction algorithm for infrared line scanner

A scene-based algorithm is developed for nonuniformity correction in focal plane of line scanning infrared imaging systems (LSIR) based on registration. By utilizing the 2D shift between consecutive frames, an implicit scheme is proposed to determine correction coefficients. All nonuniform biases are corrected to the same designated value, without estimating and removing biases explicitly, permitting quick computation for high-quality nonuniformity correction. Firstly, scene motion is estimated by image registration and consecutive frames exhibiting required 2D subpixel shift are collected. Secondly, we retrieve the difference matrix of adjacent biases by utilizing the 2D shift between consecutive frames. Thirdly, we perform specified elementary transformations and corresponding cumulative sums to the difference matrix to obtain a bias compensator. This bias compensator converts nonuniform biases to a designated detector’s bias. Finally, based on the different bias compensators obtained from several frame pairs, we calculate an averaged bias compensator for nonuniformity correction with less error. Quantitative comparisons with other nonuniformity correction methods demonstrate that the proposed algorithm achieves better fixed-pattern noise reduction with low computational complexity.

Nonuniformity correction for an infrared focal plane array based on diamond search block matching.

In scene-based nonuniformity correction algorithms, artificial ghosting and image blurring degrade the correction quality severely. In this paper, an improved algorithm based on the diamond search block matching algorithm and the adaptive learning rate is proposed. First, accurate transform pairs between two adjacent frames are estimated by the diamond search block matching algorithm. Then, based on the error between the corresponding transform pairs, the gradient descent algorithm is applied to update correction parameters. During the process of gradient descent, the local standard deviation and a threshold are utilized to control the learning rate to avoid the accumulation of matching error. Finally, the nonuniformity correction would be realized by a linear model with updated correction parameters. The performance of the proposed algorithm is thoroughly studied with four real infrared image sequences. Experimental results indicate that the proposed algorithm can reduce the nonuniformity with less ghosting artifacts in moving areas and can also overcome the problem of image blurring in static areas.

Iteratively reweighted unidirectional variational model for stripe non-uniformity correction

In this paper, we propose an adaptive unidirectional variational nonuniformity correction algorithm for fixed-pattern noise removal. The proposed algorithm is based on a unidirectional variational sparse model that makes use of unidirectional characteristics of stripe nonuniformity noise. The iteratively reweighted least squares (IRLS) technique is introduced to optimize the proposed correction model, which makes the proposed algorithm easy to implement with existing conjugate gradient method without introducing additional variables and parameters. Moreover, we derive a formula to automatically update the regularization parameter from the images. Comparative experimental results on real infrared images indicate that the proposed method can remove the stripe nonuniformity noise effectively while maintaining more useful image details.

基于衬底温度的红外焦平面联合非均匀性校正

基于衬底温度的红外焦平面联合非均匀性校正

基于自适应扩散模型的单帧红外条纹非均匀性校正算法

基于自适应扩散模型的单帧红外条纹非均匀性校正算法

A novel nonuniformity correction algorithm based on speeded up robust features extraction

The traditional image-registration based non-uniformity correction algorithm assumes that the displacements of inter-frame images are horizontal or vertical which ignores rotation, affine transformation, etc. This restriction results in low registration accuracy and then depresses the quality of nonuniformity correction. In this paper, we bring forward a novel nonuniformity correction algorithm based on speeded up robust features extraction (SURF). The feature points are extracted by hessian operator in the algorithm and are described in use of the 64-dimensional descriptor. Then the algorithm gets all the matching pairs by matching threshold, and all matching pairs can be used to estimate transform converting relations between the two consecutive images. The converting relations combined with the steepest gradient descent method can adaptively correct the gain and offset calibration parameters, which can get the accurate registration including rotation and affine transformation, and then reduce excessive noise and ghost artifacts in the process of correction. The experiments also show that our algorithm can solve the registration accuracy problem in complex conditions, and improve the correction results effectively.

Static scene statistical algorithm for nonuniformity correction in focal-plane arrays

A static scene statistical nonuniformity correction ( S 3 NUC ) method was developed based on the higher-order moments of a linear statistical model of a photodetection process. S 3 NUC relieves the requirement for calibrated targets or a moving scene for NUC by utilizing two data sets of different intensities but requires low scene intensity levels. The first-, second-, and third-order moments of the two data sets are used to estimate the gain and bias values for the detectors in a focal-plane array (FPA). These gain and bias values may then be used to correct the nonuniformities between detectors or to initialize other continuous calibration methods. S 3 NUC was successfully applied to simulated data as well as measured data at visible wavelengths.

Adaptive Image-Registration-Based Nonuniformity Correction Algorithm With Ghost Artifacts Eliminating for Infrared Focal Plane Arrays

In this paper, we present a new adaptive image registration nonuniformity correction method with the function of eliminating ghost artifacts. This method assumes that the irradiation of objects stays unchanged during the adjacent two frames' time interval and then corrects the corresponding pixels of two frames by the result of image registration. With regard to real-time continuous image sequence, we first calculate the displacement vectors based on their row and column projections. Then, by bidirectional image registration, we can get the overlapped area matrices of two frames accurately. Meanwhile, a variance threshold is set to judge the scene classification, and then, displacement revising is selectively added to decrease stripe noise. Considering that moving targets exist, the error function for updating is also accordingly adapted to eliminate the ghost artifacts. By using the least-mean-square-error iteration algorithm, the gain and offset coefficients of the overlapped area can be updated in real time with the image sequence adaptively. The advantages of this algorithm lie in its high efficiency of image processing, low computational complexity, and few ghost artifacts. A clean infrared image sequence and two real infrared datasets with nonuniformity are used to compare the ability and correction performance of different algorithms. The results show that our algorithm performs better than other compared algorithms and obtains almost no ghost artifacts.

Scene-based correction of fixed pattern noise in hyperspectral image data using temporal reordering

Hyperspectral image data suffer from pixel-to-pixel response nonuniformity that degrades the imagery in the form of columnated striping noise. This nonuniformity, or fixed pattern noise (FPN), is typically compensated for through flat-field calibration procedures. FPN is a particularly challenging problem because the detector responsivities drift relative to one another in time, requiring that the sensor be periodically recalibrated. Both the rate and severity of the drift depend on a host of factors that result in varying levels of residual calibration error being present within the data at all times. Scene-based nonuniformity correction (SBNUC) algorithms estimate and remove FPN by exploiting content within the scene data and are often necessary to acceptably remove sensor artifacts for subpixel target detection applications. We present results from two SBNUC techniques that reduce residual FPN and improve target signal-to-clutter ratio. We make the observation that temporally reordering the data in conjunction with the use of spatial ratios or differentials results in algorithms that require a low number of temporal data samples to reliably correct for FPN with minimal introduction of image artifacts. Additionally, application of the algorithms within the principal components domain can further improve their correction ability.

A scene-based nonuniformity correction algorithm based on fuzzy logic

Scene-based nonuniformity correction algorithms based on the LMS adaptive filter are quite efficient to reduce the fixed pattern noise in infrared images. They are famous for their low cost of computation and storage recourses. Unfortunately, ghosting artifacts can be easily introduced in edge areas when the inter-frame motion slows. In this paper, a gated scene-based nonuniformity correction algorithm is proposed. A novel low-pass filter based on the fuzzy logic is proposed to estimate the true scene radiation as the desired signal in the LMS adaptive filter. The fuzzy logic can also evaluate the probability that a pixel and its locals belong to edge areas. Then the update of the correction parameters for the pixels in edge areas can be gated. The experiment results show that our method is reliable and the ghosting artifacts are reduced.

A combined temporal and spatial deghosting technique in scene based nonuniformity correction

The least mean square error based nonuniformity correction algorithm is a kind of classical method to reduce the fixed pattern noise in infrared focal plane array. It is well-known for its low cost of computation and storage resources. However, it suffers from the drawback that ghosting artifacts can be easily generated in the edge areas when the inter-frame motion slows. In this paper, a combined temporal and spatial deghosting technique is proposed. Both spatial correlation detection and temporal motion detection are used to gate the update of correction parameters. The experimental results demonstrate that the deghosting performance of the proposed method is superior to other deghosting methods.

Non-uniformity Correction Algorithm Based on Polynomial Fit Estimation

Non-uniformity Correction Algorithm Based on Polynomial Fit Estimation

A Non-Uniformity Correction Algorithm Based on Single Infrared Image

To overcome the shortcoming that most of the recent Non-uniformity Correction (NUC) algorithms which need a large number of underlying images, a non-uniformity correction method based on single infrared image is proposed. In the proposed method, the moment matching algorithm is used to correct the non-uniformity image to get the pilot image. Then, the banding occurring in the pilot image is removed by the local weighted-histogram specification algorithm to get the final result. Experiment results show the effectiveness of the proposed non-uniformity correction algorithm.

改进的神经网络非均匀校正算法的研究与实现

改进的神经网络非均匀校正算法的研究与实现

Thermal imager fixed pattern noise prediction using a characterization of the infrared detector

Cooled infrared detectors are typically characterized by well-known electro-optical parameters: responsivity, noise equivalent temperature difference, shot noise, 1 / f noise, and so on. Particularly important for staring arrays is also the residual fixed pattern noise (FPN) that can be obtained after the application of the nonuniformity correction (NUC) algorithm. A direct measure of this parameter is usually hard to define because the residual FPN strongly depends, other than on the detector, on the choice of the NUC algorithm and the operative scenario. We introduce three measurable parameters: instability, nonlinearity, and a residual after a polynomial fitting of the detector response curve, and we demonstrate how they are related to the residual FPN after the application of an NUC (the relationship with three common correction algorithms is discussed). A comparison with experimental data is also presented and discussed.

A scene based nonuniformity correction algorithm for line scanning infrared image

In this paper, a fast scene based nonuniformity correction algorithm using Landweber iteration is proposed for line scanning infrared imaging systems (LSIR). The method introduces a novel framework of nonuniformity correction for LSIR by optimization. More specifically, first a “desired” image is obtained by an 1D Guassian filter from the corrected image; then a weighted mean square error optimization function is established in each line to minimize the mean square error between the corrected value and “desired” image. Correction parameters update adaptively by Landweber iteration, and then update the desired image. A stopping rule of the framework is also proposed. The quantitative comparisons with other state-of-the-art methods demonstrate that the proposed algorithm has low complexity and is much more robust on fixed-pattern noise reduction in the static scene.

Empirical frequency domain model for fixed-pattern noise in infrared focal plane arrays

In this paper, a new empirical model for the spatial structure of the fixed-pattern noise (FPN) observed in infrared (IR) focal-plane arrays (FPA) is presented. The model was conceived after analyzing, in the spatial frequency domain, FPN calibration data from different IR cameras and technologies. The analysis showed that the spatial patterns of the FPN are retained in the phase spectrum, while the noise intensity is determined by the magnitude spectrum. Thus, unlike traditional representations, the proposed model abstracts the FPN structure using one matrix for its magnitude spectrum and another matrix for its phase spectrum. Three applications of the model are addressed here. First, an algorithm is provided for generating random samples of the FPN with the same spatial pattern of the actual FPN. Second, the model is used to assess the performance of non-uniformity correction (NUC) algorithms in the presence of spatially correlated and uncorrelated FPN. Third, the model is used to improve the NUC capability of a method that requires, as a reference, a proper FPN sample.

Non-Uniformity Correction Algorithm Research Based on One FPGA Processor

The two-point non-uniformity correction based on Least mean square (LMS) algorithm which was proposed by Suxia Xing need one FPGA processor, one DSP processor and four SRAM chips. This will increase the cost and volume of the whole infrared system. In the LMS algorithm, this paper improves the estimation of the ideal signal, so that it can reduce the influence of blind elements and improve the effect of non-uniformity correction. Besides, this paper realizes the two-point non-uniformity correction based on LMS algorithm with only one FPGA processor and two SRAM chips in the final, this will reduce the volume, weight, cost of the whole infrared system. Comparing with the two-point non-uniformity correction based on one FPGA processor, this paper did not need blackbody with uniform radiation in system. Above all, the method of this paper is useful and practical.