Non-cooperative Target Recognition Research Articles

A Space Non-Cooperative Target Recognition Method for Multi-Satellite Cooperative Observation Systems

A Space Non-Cooperative Target Recognition Method for Multi-Satellite Cooperative Observation Systems

Image Segmentation Solutions for Improved Non-Cooperative Target Recognition

The number of failed satellites in space is increasing, and they need to be fueled through space docking to extend their life or clean them out of orbit. Visual measurement is widely used in the docking session because of its real-time performance and economy, but the collected images will have background interference, which directly leads to the degradation of the accuracy of visual measurement. Reduce the difficulty of subsequent image processing by solving these problems. For the non-cooperative target of the star-arrow docking ring and the solar sail panel, which are two significant targets, this paper is based on target detection, according to the image difference, binarization, morphological operations and other methods to complete the accurate segmentation of the two targets, which can be more accurately complete the target detection, and the image segmentation effect is obvious. The target image obtained by image segmentation does not contain background interference, which is convenient for feature extraction and matching of the image.

Contrastive Feature Disentangling for Partial Aspect Angles SAR Noncooperative Target Recognition

Deep learning algorithms have achieved state-of-the-art progress in synthetic aperture radar (SAR) automatic target recognition (ATR) tasks. They theoretically assume that training and test samples are independent and identically distributed (i.i.d.) for generalization, but it is intractable for practical ATR scenarios. In this paper, we propose a novel contrastive feature disentangling framework termed ConFeDent to learn features with improved generalization performance under a condition of a weaker distribution consistency. More specifically, ConFeDent aims to describe the semantic interactions between two arbitrary SAR training samples instead of treating them independently. It can implicitly disentangle features encoding the pose and identity knowledge from the whole samples with a semi-parametric geometric transformation model and a second-order energy model. In particular, except for the identity label, we use deductive-based geometry knowledge as supervision to teach the model to learn the concept of aspect angle variation. A progressively amortized inference scheme is constructed for efficient feature learning and recognition in an end-to-end manner. Finally, we further release a strengthened version, called ConFeDent+, which can explicitly utilize and learn more information from cross-category samples. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) benchmark demonstrate the effectiveness of our proposed models in the SAR ATR. In particular, we validate the algorithms in a more challenging scenario where the range of aspect angles for training and testing samples is permitted to be disparate. Our model can achieve much higher recognition accuracy than other SAR ATR algorithms.

Detection and Recognition of Spatial Non-Cooperative Objects Based on Improved YOLOX_L

In view of the intelligent requirements of spatial non-cooperative target detection and recognition tasks, this paper applies the deep learning method YOLOX_L to the task and draws on YOLOF (You Only Look One-Level Feature) and TOOD (Task-Aligned One-Stage Object Detection), which optimize and improve its detection accuracy to meet the needs of space Task Accuracy Requirements. We improve the FPN (Feature Pyramid Networks) structure and decoupled prediction network in YOLOX_L and perform a validation comparative analysis of the improved YOLOX_L on the VOC2007+2012 and spacecraft dataset. Our experiments conducted on the VOC2007+2012 benchmark show that the proposed method can help YOLOX_L achieve 88.86 mAP, which is higher than YOLOX_L, running at 50 FPS under the image size of 608 × 608. The spatial target detection method based on the improved YOLOX has a detection accuracy rate of 96.28% and a detection speed of 50 FPS on our spacecraft dataset, which prove that the method has certain practical significance and practical value.

Proportional Similarity-Based Openmax Classifier for Open Set Recognition in SAR Images

Most of the existing Non-Cooperative Target Recognition (NCTR) systems follow the “closed world” assumption, i.e., they only work with what was previously observed. Nevertheless, the real world is relatively “open” in the sense that the knowledge of the environment is incomplete. Therefore, unknown targets can feed the recognition system at any time while it is operational. Addressing this issue, the Openmax classifier has been recently proposed in the optical domain to make convolutional neural networks (CNN) able to reject unknown targets. There are some fundamental limitations in the Openmax classifier that can end up with two potential errors: (1) rejecting a known target and (2) classifying an unknown target. In this paper, we propose a new classifier to increase the robustness and accuracy. The proposed classifier, which is inspired by the limitations of the Openmax classifier, is based on proportional similarity between the test image and different training classes. We evaluate our method by radar images of man-made targets from the Moving and Stationary Target Acquisition and Recognition (MSTAR) dataset. Moreover, a more in-depth discussion on the Openmax hyper-parameters and a detailed description of the Openmax functioning are given.

Non-Cooperative SAR Automatic Target Recognition Based on Scattering Centers Models.

This article proposes an Automatic Target Recognition (ATR) algorithm to classify non-cooperative targets in Synthetic Aperture Radar (SAR) images. The scarcity or nonexistence of measured SAR data demands that classification algorithms rely only on synthetic data for training purposes. Based on a model represented by the set of scattering centers extracted from purely synthetic data, the proposed algorithm generates hypotheses for the set of scattering centers extracted from the target under test belonging to each class. A Goodness of Fit test is considered to verify each hypothesis, where the Likelihood Ratio Test is modified by a scattering center-weighting function common to both the model and target. Some algorithm variations are assessed for scattering center extraction and hypothesis generation and verification. The proposed solution is the first model-based classification algorithm to address the recently released Synthetic and Measured Paired Labeled Experiment (SAMPLE) dataset on a synthetic training data basis. As a result, an accuracy of in a 10-target test within a class experiment under Standard Operating Conditions (SOCs) was obtained. The algorithm was also pioneered in testing the SAMPLE dataset in Extend Operating Conditions (EOCs), assuming noise contamination and different target configurations. The proposed algorithm was shown to be robust for SNRs greater than dB.

Automatic Target Recognition Based on Alignments of Three-Dimensional Interferometric ISAR Images and CAD Models

Inverse synthetic aperture radar (ISAR) is capable of producing 2-D and 3-D images of non-cooperative targets. Compared with 2-D ISAR images, 3-D ISAR reconstructions can provide not only range and cross-range information, but also the information about the third dimension of the target, which is of great significance for automatic target recognition (ATR) and even more specifically to the case of non-cooperative target recognition. The alignment between the point-like 3-D ISAR reconstructions of targets and the targets’ models, such as computer aided design (CAD) models, becomes one of the essential issues in ATR that uses 3-D ISAR reconstructions. In this article, we propose an approach to address the alignment problem. The alignment problem can be decomposed into two steps, a coarse alignment and an accurate alignment. The coarse alignment can be accomplished by means of the principal component analysis (PCA), whereas the accurate alignment can be achieved by iterative closest point algorithm. In this article, the 180° ambiguity issue in PCA is discussed, and a corresponding robust solution is proposed. Moreover, the k-d tree is utilized to accelerate the searching for the pairs of correspondences, both in coarse and accurate alignment steps. In order to simulate real radar scenarios, target self-occlusion is also taken into consideration. The experimental results based on target's scattering point model simulation as well as the electromagnetic simulation in different radar scenarios verify the validity of the proposed method.

Non-Cooperative Target Recognition of Optical Remote Sensing Images based on Deep Learning Combined with Spatio-Temporal Reasoning

In order to solve the low utilization rate of spatio-temporal sequence information during non-cooperative target recognition of optical remote sensing images with deep learning method, this paper puts forward a new method for offshore non-cooperative target recognition and tracing based on spatio-temporal reasoning. In the first stage, this method uses YOLOv3 model through deep learning to recognize characteristics of large batch of offshore non-cooperative target images, and finish preliminary screening of suspicious target; in the second stage, it uses time weight function and Markov model to process and analyze temporal and spatial sequence information respectively during continuous tracing of the target and get the probability distribution of certain specific target in the two dimensions; in the third stage, it uses D-S evidence theory to process probability inforamtion in temporal and spatial dimensions and get target probability with higher reliability through fusion reasoning. The experimental verification shows the comprehensive recognition precision in the first stage is over 92%; subsequently the secondary recognition precision of the target can be improved for 35% through fusion reasoning of time weighted position distribution probability and Markov position transfer probability. Results show the reasoning elements of spatio-temporal sequences obvious improve precision of secondary discovery after the non-cooperative target is lost, and provide new thinking for non-cooperative target tracing with intelligent method.

Compressive Sensing-Based Bandwidth Stitching for Multichannel Microwave Radars.

The problem of obtaining high range resolution (HRR) profiles for non-cooperative target recognition by coherently combining data from narrowband radars was investigated using sparse reconstruction techniques. If the radars concerned operate within different frequency bands, then this process increases the overall effective bandwidth and consequently enhances resolution. The case of unknown range offsets occurring between the radars’ range profiles due to incorrect temporal and spatial synchronisation between the radars was considered, and the use of both pruned orthogonal matching pursuit and refined -norm regularisation solvers was explored to estimate the offsets between the radars’ channels so as to attain the necessary coherence for combining their data. The proposed techniques were demonstrated and compared using simulated radar data.

Non-cooperative UAV target recognition in low-altitude airspace based on motion model

Non-cooperative UAV target recognition in low-altitude airspace based on motion model

Moving target imaging in multistatic passive radar

This study presents the concept of multistatic passive images. Theoretical considerations are illustrated by the results obtained by imaging a small general aviation plane using DVB-T signals. These results are compared with simulated images derived from the same scenarios in real-life tests. The results described here show that it is possible to obtain ISAR images of aerial targets using multistatic passive radars based on DVB-T signals. Such images could be next applied for non-cooperative target recognition.

3D Fully Polarimetric Attributed Scattering Center Extraction Based on Sequential 2D SAR Images

Compared with the two-dimensional (2D) attributed scattering center model (ASCM), three dimensional (3D) fully polarimetric attributed scattering center model (FP-ASCM) is able to directly offer the target location, size, and orientation in 3D space, which plays an important role in noncooperative target recognition. Considering that 3D images of the target are hard to obtain, 3D fully polarimetric attributed scattering center feature extraction from sequential 2D SAR images is addressed, and a novel method is proposed based on sparse representation in a frequency-azimuth-elevation domain. The proposed method consists of two steps: scattering center associations from sequential images and 3D FP-ASCM parameter estimation. First, a new scattering center association method is proposed by using epipolar geometry and polarization characteristics. In this method, scattering centers from different images can be associated by minimizing their projection distance and polarimetric diversity. Second, 3D fully polarimetric attributed scattering center feature is extracted based on sparse representation in a frequency-azimuth-elevation domain. As 3D FP-ASCM includes eight target parameters and three radar parameters, the dimension of the parameterized dictionary is very large. Thus, parameter initialization and range-orientation uncoupling are adopted to reduce the dimension of the parameterized dictionary. The numerical results for electromagnetic computation data and measurement data show the efficiency of the proposed method.

Optimum Feature Selection for Efficient ISAR Image Classification

Objective: To identify optimum features from the initial feature set to classify ship ISAR images with good accuracy. The optimum feature selection brings down the computational complexity and also removes the redundant features in the feature vector. Methodology: Colour moments computed from R, G, B, RG, GB, RB component images are taken as one feature vector. The average wave energy values of horizontal, vertical and diagonal details of images decomposed up to five levels using wavelet transform are considered as second feature vector. The ship ISAR images measured with high resolution Radar are used as data set to test the classification performance of the two feature sets, which is explored uniquely in this work. Genetic algorithm and particle swarm optimization techniques are used to find the optimum feature sets. Probabilistic neural network is used to compare the classification accuracy achieved with full length feature vectors and optimum feature vectors of the two considered feature sets. Findings: Data set of two ships is taken. The data of each ship is divided in to two parts training data and test data. The feature vectors computed from training set are used to train the classifier and that computed from test set are used to test the classification accuracy. The number of variables (number of features in the feature vector) of optimization algorithms Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) is varied to find the optimum feature set to classify the ship ISAR image data. The function that computes the classification accuracy is used as the objective function of the optimization algorithms. From the initial feature vector of thirty colour moments a set of four features is identified that gives 100% accuracy for both the targets considered. From starting set of fifteen average energy levels, a set of four energy levels identified by PSO and the set of three energy levels are found to be more effective in classification of the ship ISAR images with classification accuracy of 66.6% for the first target and 75% for the second target. Both the optimization techniques performed in similar way. Other optimization algorithms can be explored to find feature combinations that classify ship ISAR images with better accuracy. Application: Classification based on ISAR images has application in automatic non-cooperative target recognition, development of monitoring systems for surveillance along transportation corridors etc. to avoid illegal immigration. Future Work: Different classifiers Self Organizing Maps, Support Vector Machines etc. can be used to study the possibility of obtaining optimum solution in smaller time. Keywords: Classification, Feature Selection, ISAR image, Optimum Feature Set, Optimization Techniques

Radar HRRP Target Recognition Based on t-SNE Segmentation and Discriminant Deep Belief Network

In radar high-resolution range profile (HRRP)-based target recognition, one of the most challenging tasks is the noncooperative target recognition with imbalanced training data set. This letter presents a novel recognition framework to deal with this problem. The framework is composed of two steps: first, the t-distributed stochastic neighbor embedding (t-SNE) and synthetic sampling are utilized for data preprocessing to provide a well segmented and balanced HRRP data set; second, a discriminant deep belief network (DDBN) is proposed to recognize HRRP data. Compared with the conventional recognition models, the proposed framework not only makes better use of data set inherent structure among HRRP samples for segmentation, but also utilizes high-level features for recognition. Moreover, the DDBN shares latent information of HRRP data globally, which can enhance the ability of modeling the aspect sectors with few HRRP data. The experiments illustrate the meaning of the t-SNE, and validate the effectiveness of the proposed recognition framework with imbalanced HRRP data.

Periodicity analysis on cat-eye reflected beam profiles of optical detectors

The cat-eye effect reflected beam profiles of most optical detectors have a certain characteristic of periodicity, which is caused by array arrangement of sensors at their optical focal planes. It is the first time to find and prove that the reflected beam profile becomes several periodic spots at the reflected propagation distance corresponding to half the imaging distance of a CCD camera. Furthermore, the spatial cycle of these spots is approximately constant, independent of the CCD camera’s imaging distance, which is related only to the focal length and pixel size of the CCD sensor. Thus, we can obtain the imaging distance and intrinsic parameters of the optical detector by analyzing its cat-eye reflected beam profiles. This conclusion can be applied in the field of non-cooperative cat-eye target recognition.

Classification of Birds and UAVs Based on Radar Polarimetry

This letter aims to show the potential of using polarimetric parameters to distinguish between large birds and unmanned aerial vehicles (UAVs) of comparable size in the context of a modern long-range air defense radar. Time is a critical resource in such systems, and techniques for robust noncooperative target recognition not relying on spatial resolution or long dwell times are highly desired. Furthermore, methods less dependent on target micromotion are, in many cases, required. Methods exploiting polarimetric features are shown to have potential in both cases. An experiment in S-band shows that a simple nearest-neighbor classifier can achieve good separation between UAVs and birds both with and without detectable micromotion based on a set of polarimetric parameters alone.

A TSR Visual Servoing System Based on a Novel Dynamic Template Matching Method.

The so-called Tethered Space Robot (TSR) is a novel active space debris removal system. To solve its problem of non-cooperative target recognition during short-distance rendezvous events, this paper presents a framework for a real-time visual servoing system using non-calibrated monocular-CMOS (Complementary Metal Oxide Semiconductor). When a small template is used for matching with a large scene, it always leads to mismatches, so a novel template matching algorithm to solve the problem is presented. Firstly, the novel matching algorithm uses a hollow annulus structure according to a FAST (Features from Accelerated Segment) algorithm and makes the method be rotation-invariant. Furthermore, the accumulative deviation can be decreased by the hollow structure. The matching function is composed of grey and gradient differences between template and object image, which help it reduce the effects of illumination and noises. Then, a dynamic template update strategy is designed to avoid tracking failures brought about by wrong matching or occlusion. Finally, the system synthesizes the least square integrated predictor, realizing tracking online in complex circumstances. The results of ground experiments show that the proposed algorithm can decrease the need for sophisticated computation and improves matching accuracy.

Automatic feature extraction from jet engine modulation signals based on an image processing method

This study presents an automatic method for extracting the jet engine features from the joint time-frequency (JTF) representation of jet engine modulation (JEM) signals. First, empirical mode decomposition with adaptive low-pass filtering was employed to extract the first harmonic component of the JEM signal. Then, a smoothed pseudo Wigner–Ville distribution (SPWVD) technique was used for acquiring the refined JTF representation. After converting the SPWVD result into an image with RGB colours, the green component was extracted as a representative of the JEM component. Finally, the peaks detected from the extracted green component can represent the jet engine features. The approach proposed in this study is significant because the overall procedures for extracting the jet engine features are not manual but automatically performed based on the image processing method. Application to measured JEM signals demonstrated that the automatic feature presented in this study improved the accuracy of JEM analysis and is expected to be efficient for real-time radar non-cooperative target recognition.

Automatic algorithm for estimating the jet engine blade number from the radar target signature of aircraft targets

Radar is a powerful tool for detecting and tracking airborne targets by day and night. Nowadays, noncooperative target recognition (NCTR) is considered an important subject. NCTR means technologies provide high-resolution target signatures and use them to make some decisions about the type of target that has been detected. Clearly, this function is performed with no cooperation from the target concerned, and it may even not be aware that its signature is being measured for recognition purposes [1]. Jet engine modulation (JEM) [2]–[4] generated by the engine rotation in an aircraft has been used for radar-based NCTR purposes, along with high-resolution range profile and inverse synthetic aperture radar (ISAR) [5]–[7]. Recognition by JEM uses micro-Doppler features induced by a rotating jet engine compressor. The micro-Doppler is characterized by a time-dependent Doppler shift that is different from that caused by the target bulk motion.

Non-cooperative target recognition by means of singular value decomposition applied to radar high resolution range profiles.

Radar high resolution range profiles are widely used among the target recognition community for the detection and identification of flying targets. In this paper, singular value decomposition is applied to extract the relevant information and to model each aircraft as a subspace. The identification algorithm is based on angle between subspaces and takes place in a transformed domain. In order to have a wide database of radar signatures and evaluate the performance, simulated range profiles are used as the recognition database while the test samples comprise data of actual range profiles collected in a measurement campaign. Thanks to the modeling of aircraft as subspaces only the valuable information of each target is used in the recognition process. Thus, one of the main advantages of using singular value decomposition, is that it helps to overcome the notable dissimilarities found in the shape and signal-to-noise ratio between actual and simulated profiles due to their difference in nature. Despite these differences, the recognition rates obtained with the algorithm are quite promising.