Laser Active Imaging System Research Articles

Time-domain signal extraction and image reconstruction through scattering medium based on fully convolutional neural network

In scattering environments, laser-active imaging systems are affected by medium absorption and multiple scattering, which reduces the system's ability to effectively detect objects hidden behind the scattering medium. Conventional penetrating scattering-medium imaging methods are less effective under dynamic and strong-scattering conditions. Current deep learning-based image reconstruction techniques only perform image reconstruction from the perspective of the image, without taking the characteristics of the difference between the distribution of the signal and noise in the initially acquired time-domain signals into consideration. In this study, we start from the underlying time-domain distribution characteristics between the signal and the noise and use the full convolutional neural network to learn the time-domain distribution discrepancy between the scattering noise and the echo signal. We directly establish the mapping between noise-containing and noise-free signals to complete the extraction of the target reflective echo signals and to realize target image reconstruction in the scattering medium. We experimentally prove that the proposed method can realize the reconstruction of the target intensity and distance images in a strong scattering environment. Finally, we experimentally demonstrate that the convolutional neural network (CNN) maintains its image reconstruction capability for scattered images outside the range of the training data distribution.

UAV Detection with Transfer Learning from Simulated Data of Laser Active Imaging

With the development of our society, unmanned aerial vehicles (UAVs) appear more frequently in people’s daily lives, which could become a threat to public security and privacy, especially at night. At the same time, laser active imaging is an important detection method for night vision. In this paper, we implement a UAV detection model for our laser active imaging system based on deep learning and a simulated dataset that we constructed. Firstly, the model is pre-trained on the largest available dataset. Then, it is transferred to a simulated dataset to learn about the UAV features. Finally, the trained model is tested on real laser active imaging data. The experimental results show that the performance of the proposed method is greatly improved compared to the model not trained on the simulated dataset, which verifies the transferability of features learned from the simulated data, the effectiveness of the proposed simulation method, and the feasibility of our solution for UAV detection in the laser active imaging domain. Furthermore, a comparative experiment with the previous method is carried out. The results show that our model can achieve high-precision, real-time detection at 104.1 frames per second (FPS).

凝视型激光主动成像系统性能验证

凝视型激光主动成像系统性能验证

Fast contour torque features based recognition in laser active imaging system

Recent years laser active imaging system are widely used as a useful detection artifice. While many denoising and enhancing algorithms have been proposed, target recognition in laser active imaging is still a new domain that few researches have been done. Classical recognition methods often break done due to the characteristic of laser active imaging. In this paper we present a novel target recognition method based on fast contour torque features (FCTF). The proposed fast contour torque features contain abundant information of the target, such as the size, position, darkness and shape regularly of the contours. Meanwhile the features are invariant to rotation, scaling and affine transform, and can be computed efficiently.We first extract feature regions by MSER algorithm, and transform them into circular areas, so the regions are invariant to affine transform. Then local invariant features of the feature regions were extracted by fast contour torque feature descriptor and we input into the trained SVM classifier for identify. Comprehensive experiments show that our approach achieves higher recognition rate in rotation and affine transformation as state-of-art method, and meet the real-time requirement in laser active imaging.

直接测距型激光主动成像系统发展现状

Laser active imaging system has been used in many fields due to its high precision of angular,range and velocity,so laser active imaging system with direct ranging are researched deeply. According to the operational mode,principle and systemic parameter of many laser active imaging systems from several overseas research institutes are introduced. The advantages and shortcomings of those operational modes are summarized and compared. Analyzing result shows that long operational range could be gained with single or small FPA combining optical-mechanical scanning. But the repeated- frequency of laser would be high and post data-processing is complicated with this method. The volume and power dissipation is also considerable. Compact laser active imaging system could be achieved easily with flash sensor and related repeated-frequency of laser is low.Although operational range is limited,it could be enhanced by using electronical steering method. In actural system,several factors,such as volume,weight and power dissipation of system,peak power and repeated frequency of laser,type of sensor and post data-processing capability,should be taken into account comprehensively to reach a comprise.

Quantitative assessment of laser-dazzling effects through wavelet-weighted multi-scale SSIM measurements

Abstract Laser active imaging systems are widespread tools used in region surveillance and threat identification. However, the photoelectric imaging detector in the imaging systems is easy to be disturbed and this leads to errors of the recognition and even the missing of the target. In this paper, a novel wavelet-weighted multi-scale structural similarity (WWMS-SSIM) algorithm is proposed. 2-D four-level wavelet decomposition is performed for the original and disturbed images. Each image can be partitioned into one low-frequency subband (LL) and a series of octave high-frequency subbands (HL, LH and HH). Luminance, contrast and structure comparison are computed in different subbands with different weighting factors. Based on the results of the above, we can construct a modified WWMS-SSIM. Cross-distorted image quality assessment experiments show that the WWMS-SSIM algorithm is more suitable for the subjective visual feeling comparing with NMSE and SSIM. In the laser-dazzling image quality assessment experiments, the WWMS-SSIM gives more reasonable evaluations to the images with different power and laser spot positions, which can be useful to give the guidance of the laser active imaging system defense and application.

基于特征点分布特性的激光干扰效果评估算法

Laser active imaging systems are usually used in region surveillance and target identification. However, the photoelectric imaging detector in the imaging systems is easy to be disturbed and this leads to errors of the recognition and even the missing of the target. Assessment of laser-dazzling effects in view of the invalidation of the target feature must be better understood. A new feature-point similarity (FPSIM) assessment algorithm is proposed. The feature accelerated segmentation testing (FAST) algorithm is used to extract feature points of the original image and disturbed image. The target area is obtained via feature-point matching, and the feature-point maintenance as well as stabilization is computed in the target area. The location of the feature points in the original image is obtained, and the local luminance and contrast distortion are compared in the same place of the two images. The normalized FPSIM is obtained via product of the feature-point maintenance, stabilization, luminance distortion and contrast distortion. The luminance imaging experiment is performed for the target by utilizing the laser active imaging system. In the experiment, the disturbed images of different disturbing powers, different intense backgrounds and different spot positions are obtained. The proposed FPSIM algorithm is used to evaluate the newly obtained laser-dazzling images, and the results show that the FPSIM reflects the varieties of the feature points in target recognition objectively. Compared with normalized mean square error (NMSE) and structural simila rity (SSIM), the FPSIM gives a more reasonable evaluation result for different laser-dazzling images. The evaluation results are more suitable for the subjective visual feeling, and FPSIM can also give the guidance of the laser active imaging system defense and application.

基于光斑与图像特征的动态激光干扰效果评估

Laser active imaging systems are usually used in important region surveillance and dangerous target identification. However, the photoelectric imaging detector in the imaging systems is easy to be disturbed from opponents and this leads to errors of the recognition. In practice, the laser power and the spot position change momentarily and this leads to different laser-dazzling effects in every frame. Therefore, dynamic assessments of laser-dazzling effects must be better understood. A new no-reference dynamic feature metric (DFM) assessment algorithm based on the character of the laser spot and image feature is proposed. The features from accelerated segment test (FAST) algorithm is used to extract feature points of the target image and disturbed image. The target area is obtained via feature-point matching, and the characteristic of the laser spot reflects the image quality through the saturated pixel number and spot position in the target area. After that, the distribution of the feature points, edges and edge definition reflect the varieties of the image feature. The normalized DFM is obtained via product of the five factors above. The luminance imaging experiment is performed for the targets by utilizing the laser active imaging system. In the experiment, the disturbed images of different disturbing powers and different spot positions are obtained. The proposed DFM algorithm is used to evaluate the newly obtained laser-dazzling images, and the results show that the DFM gives a more reasonable evaluation result for different laser-dazzling images. The evaluation results reflect the invalidation of the target identification algorithm objectively.

结合激光功率和光斑位置的多帧动态干扰效果评估

Laser active imaging systems are usually used in important region surveillance and dangerous target identification. However, the imaging system is easy to be disturbed and this leads to the distortion of the images. The positions of the target, imaging system, interference and the laser power change momentarily, so the dazzling effect is different in every frame. Therefore, how to evaluate the influence of the laser on imaging systems must be better understood. A new no-reference image quality assessment algorithm based on the dynamic features of consistent multi-frame images is proposed, which is called feature-point dynamic (FPD). The feature-point matching correlation and area diversification of one frame are calculated in the target region. The difference of correlation, area distribution, structure, frequency and salience of multi-frame images are also compared in the target region. The normalized results are obtained via product of these factors above. The luminance imaging experiment is performed for the targets by utilizing the laser active imaging system. The disturbed images of different disturbing powers and different spot positions are obtained. The proposed FPD algorithm is used to evaluate the laser-dazzling images, and the results show that the FPD gives a more reasonable evaluation for multi-frame laser-dazzling images in the period of time. The evaluation results reflect the invalidation of the target identification algorithm objectively.

基于方向滤波器组的激光主动成像图像复原

Since images from laser active imaging system on moving flat are blurred and polluted by kinds of noise, state of the art image deblurring techniques which are sensitive to image noise cannot achieve satisfying results. A new deblurring method for laser active imaging system based on directional filters is proposed. The key observation is applying a group of directional filters to the blurry and noisy image to reduce the noise level and estimate the blur kernel. For each directional filter, compute the Radon transform of the blur kernel along the orthogonal direction of the filter. Then the two-dimensional (2D) blur kernel can be reconstructed by inverse Radon transform. The last step is to restore the sharp image due to the blur kernel. An active imaging system is built up for experiment and the results indicate that compared with state of the art blind image deblurring techniques, the proposed method not only drastically improves the image quality in term of subjective visual, but also remarkably improves the PSNR.

结合光斑与目标特征的激光干扰效果评估

The global features of laser-dazzling images, the local features of targets and the distribution characteristics of laser spots are analyzed.A new algorithm composed of Weighted Feature Similarity (WFSIM) is proposed for the assessment of laser-dazzling effects. The method firstly uses the global features including luminance, contrast, edge definition and local feature-point maintainability to obtain the assessment model. Then, it computes the weighting factors, such as the saturated pixel number, signal to noise ratio and spot visual impotence based on the size, intensity and position of the laser spot. Based on the above results, a modified assessment algorithm can be constructed and a normalized evaluation can be implemented. An experiment platform of laser active imaging and recognition system is built for laser-dazzling images under different laser intensities, different background intensities and different spot positions. The proposed WFSIM algorithm is used to evaluate obtained laser-dazzling images. In terms of experiment results, the WFSIM algorithm shows good feasibility and rationality as comparing with Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Structural Similarity (SSIM) image quality assessment methods. The evaluation results is consistent with the subjective human visual perception.

基于双隐含层BP算法的激光主动成像识别系统

An experiment platform for laser active imaging and recognition was established based on the traditional laser active imaging system to investigate the target recognition after laser active imaging. The working mechanism of the platform was introduced and the Hu moment feature based BP neural network algorithm with double hidden layers and an experimental process were given. The target feature vector was consisted of seven invariant Hu moments. The BP neural network algorithm with double hidden layers including 136 weight coefficients was trained by 240 original sample libraries. The trained BP neural network algorithm was used to research a distance moving target in the dark condition, a model of 43 submachine gun, and a clear infrared laser active image was obtained. Experiment results show that statistical recognition probability is 68.87% for 2740 frames of images at 450 m and 72.11% for 2420 frames of images at 550 m. The corresponding recognition probabilities from rotation transformation are 80.05% and 84%, respectively, which is better than the results by affine transformation.

Noise Analysis and Filtering for Laser Active Imaging System

In order to deal with the contradiction between suppressing speckle noise and reserving details in laser active imaging recognition system, a denoising method based on contour curvature is proposed. Due to the contour curvature, the pixels in the image are divided into different classes, which contain different amount of information. The filter parameters are different for each class. Firstly, the origin image is smoothed using wavelet soft thresholding, then the contours are extracted by Morphological edge detection operator. Due to the difference of contour curvature, the pixels are labeled with point of strong signal, point of weak signal or point of no signal. Pixels with different labels are filtered by Lee filter of different step width. Experiment result indicates that compared with classical Lee filter, the proposed method performs better in filtering and keeping edge information.

Research on rocket laser scattering characteristic simulation software

The modeling and simulation of target laser scattering characteristics is one of the most important areas of research on rocket optical characteristics, and is also the foundation of the design and evaluation of laser active detecting systems. This paper develops rocket target laser scattering characteristic simulation software, which has the primary functions of calculating the rocket target’s laser radar cross section (LRCS) and imaging simulation. This software supports the whole simulation process, including target modeling, experimental test of the simulation, result analysis, and so on. Finally, the corresponding laser active imaging experiment is performed, and the imaging system specifications and performance are validated. This simulation platform provides a good simulation infrastructure for research on laser active imaging systems in shooting range, and helps to meet the application demand in the rocket imaging area.

Intensity image denoising for laser active imaging system using nonsubsampled contourlet transform and SURE approach

This paper presents an algorithm based on nonsubsampled contourlet transform (NSCT) and Stein's unbiased risk estimate with a linear expansion of thresholds (SURE-LET) approach for intensity image denoising. First, we analyzed the multiplicative noise model of intensity image and make the non-logarithmic transform on the noisy signal. Then, as a multiscale geometric representation tool with multi-directivity and shift-invariance, NSCT was performed to capture the geometric information of images. Finally, SURE-LET strategy was modified to minimize the estimation of the mean square error between the clean image and the denoised one in the NSCT domain. Experiments on real intensity images show that the algorithm has excellent denoising performance in terms of the peak signal-to-noise ratio (PSNR), the computation time and the visual quality.

Influence of Atmospheric Propagation on Performance of Laser Active Imaging System

Atmospheric propagation has serious influence on the performance of a good designed laser active imaging system. Atmospheric attenuation and turbulence are two main effects on laser atmospheric propagation. Imaging SNR (Signal-Noise-Ratio) and resolution are two key indexes to describe the performance of a laser active imaging system. Establishing the relation between system performance index and atmospheric propagation effect is significant. The paper analyzed the relation between imaging performance and atmospheric attenuation and turbulence through simulation. And also the experiments were done under different weather to validate the conclusion of simulation.

基于序贯检验的激光主动成像目标探测方法

基于序贯检验的激光主动成像目标探测方法

提高激光主动成像系统性能的匹配设计

提高激光主动成像系统性能的匹配设计