一种基于场景的红外图像非均匀性校正算法

提出了一种基于场景的红外图像非均匀性校正算法。该算法结合了两点定标校正算法和基于场景的改进的恒定统计算法,将两点校正算法的校正系数作为恒定统计算法的系数初值,并引入阈值进行运动状态检测,对运动场景和非运动场景分别进行系数更新。实验表明,该算法可以实现对红外图像非均匀性的校正,对于本文实验中的视频图像,在100帧时算法收敛,其收敛时间优于其他传统基于场景的非均匀性校正算法,并一定程度上抑制了"鬼影"现象。     

基于自适应滤波的红外焦平面阵列非均匀性校正算法

 红外探测器响应漂移特性会降低红外焦平面阵列（IRFPA）非均匀性校正的精度。针对该问题提出了一种基于场景的IRFPA非均匀性校正算法。该算法利用所获得的序列成像场景信息,采用一种基于快速自适应滤波器的最优化递归估计方法来获得非均匀性校正参数,并利用当前的成像信息来更新校正参数,以此降低探测器响应漂移特性对非均匀性校正的影响。算法仿真实验显示,对非线性参数为26.12%的同一图像,使用该算法、两点校正算法和卡尔曼滤波校正算法校正1 h后,可分别将非线性参数降至1.856%,3.122%和1.893%,说明该算法可获得稳定而较好的非均匀性校正效果。     

基于场景的红外焦平面阵列非均匀性评价方法

针对红外遥感图像非均匀性的定量评价问题,研究了基于场景的红外焦平面阵列响应非均匀性的定量评价方法。利用真实红外遥感图像序列量值分布的特点,针对小序列图像提出了基于序列排序和小波变换的标准图像构建算法;结合非均匀性评价指标,提出了利用构建标准图像的响应非均匀性对红外图像的非均匀性进行定量估计;分别利用黑体定标实验和统计方法对标准图像构建算法进行了验证,同时进行了不同校正算法的剩余非均匀性评价实验。结果表明,128frame序列构建得到的标准图像的非均匀性分别比同温黑体图像小0.92%,比统计标准图像大0.59%。该方法可用于定量估计红外遥感图像的非均匀性。     

小波变换的红外焦平面阵列非均匀性校正算法

红外焦平面阵列(IRFPA)的非均匀性是其应用中必须解决的技术难题之一。基于小波理论,提出了一种基于成像场景的红外焦平面非均匀性校正算法。该算法选择合适的小波函数对红外成像序列进行小波分解,而后对分解的信号计算出相应的统计量,从而得出红外焦平面非均匀性校正的偏置和增益校正系数,以此最终实现非均匀性校正。对真实红外序列图像的处理效果验证了该算法可较好地实现非均匀性校正。此外该算法对慢变化量具有较好的自适应性,可较好地抑制一般基于场景统计的非均匀性校正算法中出现的"人工虚影"的现象。     

一种适用于红外搜索跟踪系统的实时非均匀性校正算法

针对目前红外搜索系统实用性较强的两点非均匀性校正存在难以实时跟踪图像非均匀的不足点,提出一种新的基于两点非均匀性校正和基于场景的实时联合校正算法。该算法利用两点校正提供基础校正系数,并充分利用红外搜索系统大数据量的特点,对实时数据量进行统计、分析,进而找出系统非均匀性随时间的漂移量,解决只采用两点校正算法带来的红外图像退化的问题。多次试验证明,采用联合非均匀校正算法的相对非均匀度由两点校正的5%降到了2%左右,并具有时间稳定性,获得了较好的校正效果。     

基于时域矩匹配的红外图像自适应非均匀性校正EI北大核心CSCD

非制冷红外焦平面阵列的响应非均匀性通常表现出与条带噪声相类似的特性。基于场景的校正算法是提高图像质量,补偿响应特性漂移的有效措施。在深入研究矩匹配理论的基础之上,提出了一种新的时域矩匹配非均匀性校正方法。利用相邻帧矩匹配后的图像对场景的变化列进行估计,并在时间域对校正参数进行自适应更新。利用真实的红外图像序列验证了该算法在收敛速度和去鬼影方面的优越性和有效性。     

基于时域矩匹配的红外图像自适应非均匀性校正

非制冷红外焦平面阵列的响应非均匀性通常表现出与条带噪声相类似的特性。基于场景的校正算法是提高图像质量,补偿响应特性漂移的有效措施。在深入研究矩匹配理论的基础之上,提出了一种新的时域矩匹配非均匀性校正方法。利用相邻帧矩匹配后的图像对场景的变化列进行估计,并在时间域对校正参数进行自适应更新。利用真实的红外图像序列验证了该算法在收敛速度和去鬼影方面的优越性和有效性。     

基于特征分解的红外焦平面非均匀性校正算法

针对目前红外焦平面自适应场景校正算法工程应用的局限,提出了一种基于红外焦平面非均匀性特征分解的场景校正算法。分析了红外焦平面非性匀性构成因素,把其中的高频部分分解成盲点、斑块、行列非均匀性等,把缓慢变化的低频部分分解成梯度渐变非均匀性;分别对各类非均匀性采用不同的校正算法;合并校正结果,得到校正后的图像。实验结果表明,该算法校正精度高、收敛速度快、抑止目标退化能力强,适合工程应用。     

一种新的红外焦平面阵列非均匀性代数校正算法

 针对传统红外图像非均匀性代数校正算法收敛速度慢、运动估计精度不高的缺点,提出一种多尺度光流帧间运动估计的非均匀代数校正算法。通过时域低通滤波,采用多尺度光流估计下一帧图像,将所得图像对进行代数校正。该方法性能在自主研发的热像仪中得到验证,和传统的红外图像非均匀性代数校正算法相比,算法收敛速度从60帧提高到25帧,提高一倍左右,运动估计精度从校正后图像方差76.539减少到32.482。     

基于场景的红外焦平面阵列非均匀性校正算法综述

红外焦平面阵列的非均匀性噪声是制约红外成像质量的主要因素。基于场景的非均匀校正算法通常利用图像序列并依赖帧间运动对焦平面阵列的非均匀性进行校正。介绍和分析了全局非均匀性校正,Kalman滤波器法,自适应滤波法,轨迹跟踪法,基于场景运动分析的校正算法,基于小波变换实现低通滤波的校正算法,高通滤波与神经网络相结合的算法,基于小波去噪、序列图像配准和正交最小二乘拟合的校正算法,改进的神经网络算法以及代数算法等,对算法进行了比较。     

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.     

An enhanced non-uniformity correction algorithm for IRFPA based on neural network

Influenced by detector materials’ non-uniformity, growth and etching techniques, etc., every detector’s responsivity of infrared focal plane arrays (IRFPA) is different, which results in non-uniformity of IRFPA. And non-uniformity of IRFPA generates fixed pattern noises (FPN) that are superposed on infrared image. And it may degrade the infrared image quality, which greatly limits the application of IRFPA. Non-uniformity correction (NUC) is an important technique for IRFPA. The traditional non-uniformity correction algorithm based on neural network and its modified algorithms are analyzed in this paper. And a new improved non-uniformity correction algorithm based on neural network is proposed in this paper. In this algorithm, the desired image is estimated by using three successive images in an infrared sequence. And blurring effect caused by motion is avoided by applying implicit motion detection and edge detection. So the estimation image is closer to real image than the estimation image estimated by other algorithms, which results in fast convergence speed of correction parameters. A comparison is made to these algorithms in this paper. And experimental results show that the algorithm proposed in this paper can correct the non-uniformity of IRFPA effectively and it prevails over other algorithms based on neural network.     

Fast iterative adaptive nonuniformity correction with gradient minimization for infrared focal plane arrays

A fast scene-based nonuniformity correction algorithm is proposed for fixed-pattern noise removal in infrared focal plane array imagery. Based on minimization of L0 gradient of the estimated irradiance, the correction function is optimized through correction parameters estimation via iterative optimization strategy. When applied to different real IR data, the proposed method provides enhanced results with good visual effect, making a good balance between nonuniformity correction and details preservation. Comparing with other excellent approaches, this algorithm can accurately estimate the irradiance rapidly with fewer ghosting artifacts.     

Single image non-uniformity correction using compressive sensing

A non-uniformity correction (NUC) method for an infrared focal plane array imaging system was proposed. The algorithm, based on compressive sensing (CS) of single image, overcame the disadvantages of “ghost artifacts” and bulk calculating costs in traditional NUC algorithms. A point-sampling matrix was designed to validate the measurements of CS on the time domain. The measurements were corrected using the midway infrared equalization algorithm, and the missing pixels were solved with the regularized orthogonal matching pursuit algorithm. Experimental results showed that the proposed method can reconstruct the entire image with only 25% pixels. A small difference was found between the correction results using 100% pixels and the reconstruction results using 40% pixels. Evaluation of the proposed method on the basis of the root-mean-square error, peak signal-to-noise ratio, and roughness index (ρ) proved the method to be robust and highly applicable.     

Nonuniformity correction of infrared images based on infrared radiation and working time of thermal imager

The spatial fixed-pattern noise (FPN) reduces the quality of the infrared image seriously, even makes infrared images inappropriate for some applications. In order to lower the FPN, some critical nonuniformity correction (NUC) algorithms such as NUC based on linear model, scene-based NUC and so on have been developed. But the algorithms have some drawbacks: restricted application in small dynamic range of objects temperature, low performance under the drift with the working time and complex calculations. In these cases, we develop a bivariate and quadratic model (bivariate is radiation and working time) of the FPA and the NUC technique based on the model. The proposed method is a true reflection of the infrared response and is a good solution for hardware implementation. It overcomes the drawbacks of the critical algorithm mentioned above. The last simulations and experiments show that the proposed algorithm exhibits a superior correction effect in both large objects temperature range and long working time of the thermal imager.     

适应积分时间调整的红外图像非均匀性校正方法

针对应用常规红外图像非均匀性校正方法在变积分时间时,图像灰度值会发生改变的现象,提出了一种适应积分时间调整的红外图像非均匀性校正方法.该方法将不同积分时间、不同温度的黑体定标数据和对应的理论红外辐射量整合为一个整体数据库,借助神经网络损失函数和误差反向传递机制,对模型中的校正系数进行学习.训练得到的校正网络能在红外相机积分时间实时调整过程中,保证图像均匀地稳定输出,对后端红外图像处理有着重要意义,并验证训练该网络不需要大量定标数据.而针对红外探测器响应漂移的现象,则提出了在线修正校正系数的方法以有效应对.     

Non-uniformity correction of infrared focal plane array in point target surveillance systems

We discuss the influence of non-uniformity and non-uniformity correction on point target detection in infrared surveillance system, and propose a non-uniformity correction approach which is based on signal intensity and sensor characteristics. Theoretical models are used to derive the combined effect of background clutter, sensor random noise, target, non-uniformity and correction error on the signal-to-noise-and-clutter ratio. From our analysis, it can be noted that background clutter intensity is successively modulated by sensor non-uniformity and non-uniformity correction, while sensor random noise is modulated by the non-uniformity correction process only. Furthermore, background clutter and sensor random noise are the key factors that affect the performance of a surveillance system, when it is used to detect point targets. The method presented in this paper takes all of the above into account, moreover, it considers the difference between scanning and staring focal plane array. The experimental results demonstrate the effectiveness of the proposed method.     

基于衬底温度和贝叶斯估计的红外非均匀性校正（英）

分析了红外焦平面阵列(IRFPA)基于定标的非均匀性校正法(NUC)和基于场景的NUC算法各自的优势和问题,在此基础上提出了联合非均匀性校正方法。根据上电时刻焦平面衬底的温度值,从FLASH中提取事先存储的对应温度区间的增益和偏置校正参数,初步消除探测器的非均匀性。通过分析初步校正后图像残余非均匀性噪声的特性,提出了一种自适应非均匀性校正算法NSCT,对经过NSCT分解后的子带图像,利用贝叶斯阈值逐点进行信号方差和噪声方差估计,计算出残余非均匀性噪声后并加以去除。实验结果表明,该算法能有效提高校正精度,并具有更强的环境适应性。