MAE-ResTransFPN: a transformer-based framework for radar target detection in clutter-rich environments

A novel target detection algorithm for synthetic aperture radar (SAR) images based on an improved visual attention method is proposed in this paper. With the development of SAR technology, target detection algorithms are confronted with many difficulties such as a complicated environment and scarcity of target information. Visual attention of the human visual system can make humans easily focus on key points in a complex picture, and the visual attention algorithm has been used in many fields. However, existing algorithms based on visual attention models cannot obtain satisfactory results for SAR image target detection under complex environmental conditions. After analysing the existing visual attention models, we combine the pyramid model of visual attention with singular value decomposition to simulate the human retina, which can make the visual attention model more suitable to the characteristics of SAR images. We introduce variance weighted information entropy into the model to optimize the detection results. The results obtained by the existing visual attention algorithm for target detection in SAR images yield a large number of false alarms and misses. However, the proposed algorithm can improve both the efficiency and accuracy of target detection in a complicated environment and under weak-target conditions. The experimental results validate the performance of our method.

Both visible and infrared images are important sources of intelligence information on the battlefield, and air‐to‐ground reconnaissance by UAV is an important means to obtain intelligence. However, there are great challenges in ground target detection and tracking, especially in complex battlefield environments. Aiming at the problem of insufficient accuracy of target detection by a single type of sensor in the battlefield environment at this stage, a target detection method by fusion of visible and infrared images is proposed in this paper, which is called ReconnaissanceFusion‐YOLO (RF‐YOLO), and with the help of infrared imagery, it can effectively improve the accuracy of target detection in the case of insufficient light. The performance of target detection in the battlefield is significantly improved by introducing two key innovative modules: dual feature fusion (DFF) module and feature fusion corrector (FFC) module. The DFF module enhances multi‐channel feature fusion through a novel concatenation and channel‐wise attention mechanism, while the FFC module performs feature correction between parallel streams using spatial and channel‐wise weights, addressing noise and uncertainty in different modalities. These modules are integrated on top of a dual‐stream YOLO architecture, allowing for effective fusion of visible and infrared information. RF‐YOLO was trained and evaluated using the FLIR data set, containing 5142 pairs of strictly aligned visible and infrared images. Results demonstrate that RF‐YOLO significantly outperforms benchmark networks in terms of robustness requirements. Specifically, the large model of RF‐YOLO achieves an mAP of 0.831, which is a significant improvement compared to the YOLOv5l inf benchmark's 0.739. This represents an improvement of over 12% in detection accuracy. Additionally, RF‐YOLO offers a Nano version that balances accuracy and speed. The Nano version achieves an mAP of 0.765, while maintaining a model size of only 11.5 MB, making it suitable for deployment on UAV edge computing devices with limited resources. To validate the practical applicability of our approach, this paper successfully implements target detection and tracking on a real UAV's edge computing device using the ROS system and SiameseRPN, combined with the proposed RF‐YOLO. Real‐world flight tests were conducted on an internal playground, demonstrating the effectiveness of our method in actual UAV applications. The system achieved a processing rate of approximately 10 fps at 640 × 640 resolution on an NVIDIA TX2 edge computing device, showcasing its real‐time performance capability in practical scenarios. This study contributes to enhancing UAV‐based battlefield reconnaissance capabilities by improving the accuracy and robustness of target detection and tracking in complex environments. The proposed RF‐YOLO method, along with its successful implementation on a real UAV platform, provides a promising solution for advanced military intelligence gathering and decision‐making support.

Camouflage target detection is one of the key challenges in the field of deception detection, especially with the continuous advancement of camouflage technologies. The high fusion of target and environmental background makes conventional single-modal object detection algorithms perform significantly worse in complex environments. To address this issue, this paper proposes a novel camouflage target intelligent detection algorithm based on visible-infrared feature-level fusion, called YOLO-DenseFuse. First, a YOLO-Dense dual-stream feature extraction channel for visible and infrared images is designed to efficiently extract and reuse information from both infrared and visible light bands. Then, to efficiently utilize the positional information in infrared images, residual connections are employed to extract target location information. Finally, a camouflage target multi-spectral dataset is constructed for performance validation of multi-modal image fusion methods. Experimental results show that the proposed method achieves a detection accuracy of 91.91% and a recall rate of 93.33%, outperforming the YOLOv7 algorithm. Compared with other mainstream object detection algorithms, the proposed method demonstrates significant advantages in detection accuracy and other evaluation metrics. This approach has significant application prospects and practical significance for camouflage target detection in complex field battle environments, contributing to enhanced battlefield situational awareness.

In recent years radar target detection environment is more and more complex. Traditional CFAR (Constant False Alarm Rate) is a technology in which the radar system discriminates the output signal and noise of the receiver to determine whether the target signal exists under the condition that the False Alarm probability is kept Constant. In order to improve the radar target detection performance, a radar target detection method based on NN (Neural Network) is proposed. In this paper, the radar signal received by a single RBFNN is used for network training, and the probability of detection target is studied by combining the binary detection theory. Simulation results show that the proposed algorithm can effectively improve the radar target detection probability.KeywordsRadarConstant false alarm rateNeural networkTarget detection

Among various types of sensors, millimeter-wave radar is widely used, especially in the field of intelligent transportation, but often limited by low signal-to-noise ratio at the same time. As a result, sensor fusion is currently the most common way for target detection. In contrast to this, we focus on the problem of single radar target detection. Considering that automotive radar usually contains a plot generation module (which provides a limited number of candidate targets), the track-before-detect method can meet real-time requirements in ADAS. Successive track cancellation is one of the effective strategies for enhancing the performance of track-before-detect algorithms. We are inspired by this and derive a graph-based radar target detection algorithm. Since the data to be processed between frames have structural similarity, we can measure the correlation between plots better without a large amount of calculation by taking advantage of the dynamically maintained graph. Furthermore, we define a simple state transition process of each plot, which enhances robustness for temporary target signal loss. We verify the effectiveness of the proposed algorithm with both the simulation experiment and real data test in the last part of the paper.

Unmanned aerial vehicles (UAVs) become increasingly available devices with extensive usage as environmental monitoring systems. With the benefit of higher mobility, UAVs are applied to fuel various significant uses in computer vision (CV), providing more effectiveness and accessibility than surveillance cameras with permanent camera view, angle, and scale. Nevertheless, owing to camera motion and composite environments, it is problematic to identify the UAV; conventional models frequently miss UAV detection and make false alarms. Drone-equipped cameras monitor objects at changing altitudes, leading to essential scale variants. The model increases targeted accuracy and decreases false positives using real-time data and machine learning (ML) methods. Its enormous applications range from military operations to urban planning and wildlife monitoring. Therefore, this study develops a novel long-term target detection and tracking model for unmanned aerial vehicles using a deep fusion-based convolutional transform learning (LTTDT–UAVDFCTL) model. The LTTDT–UAVDFCTL model presents a new model to improve the robustness and accuracy of target tracking and detection in scale-variant environments. At first, the presented LTTDT–UAVDFCTL technique performs image pre-processing by utilizing the median median-enhanced wiener filter (MEWF) technique to improve clarity and reduce noise. For object detection (OD), the highly accurate YOLOv8 technique is utilized, followed by feature extraction through a backbone deep fusion-based convolutional transform learning of VGG16, CapsNet, and EfficientNetB7 to capture both spatial and hierarchical features across varying scales. Moreover, the graph convolutional neural network (GCN) technique is employed for long-term target detection and tracking models. Finally, the hybrid nonlinear whale optimization algorithm with sine cosine (SCWOA) is implemented for the optimum choice of the hyperparameters involved in the GCN technique. The experimental study of the LTTDT–UAVDFCTL approach is performed under the VisDrone dataset. The performance validation of the LTTDT–UAVDFCTL approach portrayed a superior mAP value of 80.13% over existing models.

Radar detection of maritime targets plays an important role in marine environment monitoring. For civil maritime detection in the areas of inshore coastal, pulse-compression radar is universally used owing to its low cost. The complex sea clutter in the practical application will greatly affect the received radar echoes. Due to the inability to accurately describe the differences in characteristics between sea clutter and maritime targets, the detection performance of methods based on mathematical derivation is not satisfactory in actual deployment. Recently, neural-based methods have made strides in many pattern recognition tasks, such as computer vision and natural language processing. The sophisticated deep neural models can be applied to different downstream tasks due to their powerful learning ability. Inspired by this idea, we propose a maritime radar target detection method in sea clutter based on deep learning. To better model the sequence correlation of radar echoes, we propose a Self-Adaption Local Augmented Long Short-Term Memory (SALA-LSTM) structure. The proposed SALA-LSTM integrates adaptive convolution into vanilla LSTM cells, which not only maintains the inherent overall sequence modeling ability of vanilla LSTM, but also strengthens its ability to perceive the correlation on a small scale in the local scope. Based on SALA-LSTM and other neural structures, we propose a radar target detection network. A measured dataset containing different typical scenarios is utilized to evaluate the detection probability and false alarm rate. The detection performance of our proposed network is superior to that of the existing methods.

Prohibited item detection is crucial for the safety of public places. Deep learning, one of the mainstream methods in prohibited item detection tasks, has shown superior performance far beyond traditional prohibited item detection methods. However, most neural network architectures in deep learning still lack sufficient local feature representation ability for overlapping and small targets, and ignore the problem of semantic conflicts caused by direct feature fusion. In this paper, we propose MFPIDet, a novel prohibited item detection neural network architecture based on improved YOLOV7 to achieve reliable prohibited item detection in complex environments. Specifically, a multi-scale attention module (MAM) backbone is proposed to filter the redundant information of target regions and further applied to enhance the local feature representation ability of overlapping objects. Here, to reduce the redundant information of target regions, a squeeze-excitation (SE) block is used to filter the background. Then, aiming at enhancing the feature expression ability of overlapping objects, a multi-scale feature extraction module (MFEM) is designed for local feature representation. In addition, to obtain richer context information, We design an adaptive fusion feature pyramid network (AF-FPN) to combine the adaptive context information fusion module (ACIFM) with the feature fusion module (FFM) to improve the neck structure of YOLOV7. The proposed method is validated on the PIDray dataset, and the tested results showed that our method obtained the highest mAP (68.7%), which is improved by 3.5% than YOLOV7 methods. Our approach provides a new design pattern for prohibited item detection in complex environments and shows the development potential of deep learning in related fields.

Automatic target recognition and detection in synthetic aperture radar (SAR) images is playing a significant role in military and civilian fields. Traditional methods are weak in migration capabilities and cumbersome in manual design. Existing deep learning-based methods can achieve high recognition and detection accuracy, but most of them have a very large number of parameters and slow recognition and computational efficiency, which makes it difficult for them to migrate to edge computing equipments, and computational efficiency is crucial in rapid maritime rescue and emergency military deployment. Therefore, we propose a lightweight feature extraction block, lightweight split concat block (LSCB), for SAR automatic target recognition and detection. Specifically, LSCB consists of several operators, such as split, depthwise convolution (DC), multiple kernel sizes block (MKSB), and channel-shuffle (CS). MKSB is a novel channel attention mechanism designed by us, which can make feature maps have target features of different scales and adjust their receptive field sizes adaptively. In addition, to assess the performance of LSCB, we use LSCB to build LSCNet-R and LSCNet-D for SAR target recognition and detection, respectively, and conduct extensive experiments on MSTAR, SSDD, and SAR-Ship-Dataset. The experimental results show that our LSCNet-R and LSCNet-D built with LSCB achieve higher recognition and detection accuracy and computational efficiency with few parameters compared with existing excellent methods. This further illustrates that the proposed LSCB can extract fine and stable features with few parameters.

Target heat-map network: An end-to-end deep network for target detection in remote sensing images

Hypersonic vehicle-borne radar target detection will suffer from time-varying plasma sheath which is formed during hypersonic flight. In this study, effect of time-varying plasma sheath on hypersonic vehicle-borne radar target detection is investigated. First, the radar echo model coupling time-varying plasma sheath effect is proposed using an analytical computation method. The regulation of electron density variation is based on flow parameters. Coherent integration result with the present of time-varying plasma sheath effect is derived. The degradations of hypersonic vehicle-borne radar target detection under time-varying plasma sheath are studied through real hypersonic vehicle flight cases. Simulation results shows that high electron density of time-varying plasma sheath will cause significant attenuation, which will lead to decrease of detection probability. In some cases, the signal-noise ratio (SNR) threshold that target can be detected increases more than 40 dB compared with situation without time-varying plasma sheath. Moreover, time-varying plasma sheath will cause time-varying amplitude and phase in pulse and between different pulses, causing intra-pulse and inter-pulse modulation phenomena. These phenomena lead to the disappearance of peak value in range-Doppler plane which will degrade the coherent accumulation as well as target detection performance. By using plasma sheath electron density diagnosis method on vehicle, compensation factor according to the electron density can be established to compensate plasma sheath effect in radar echoes. Simulations results based on real hypersonic vehicle flight flow parameters confirm the conclusions above. The work in this paper will benefit the design and application of hypersonic vehicle-borne radar in the future.

Constant false alarm rate (CFAR), a target detection method commonly used in the radar systems, has an inconsistent performance against various environments. For improving the radar detectability, this paper proposes a novel scheme of radar target detection using neural network-based adaptive selection CFAR. The proposed method employs cell-averaging, ordered-statistic, greatest-of and smallest-of CFAR thresholds as the basis of references. The pattern of those threshold values combined with the cell under test signal value will be identified and classified by the neural network to compute the raw threshold. Then, the final threshold is selected depending on the nearest value between raw and four referenced CFARs. The performance of the proposed method is examined against three possible cases of the radar systems including homogeneous background, multiple targets and clutter boundary. The result of this research shows that the proposed method outperforms the classical CFARs due to the adaptive selection algorithm can select properly among referenced CFARs against the given cases particularly in the homogeneous and multiple target environments.

Hyperspectral Image (HSI) is used widely in many areas, especially in the remote sensing field. Compared with the traditional remote sensing HSI, the large-scale and high-resolution HSI (LHHSI) which has big data and large size is high-resolution both in spatial domain and spectral domain. However, traditional methods of automatic target detection do not apply to LHHSI. Therefore, this paper proposes a novel framework of automatic target detection for LHHSI based on spatial-spectral interest point (SSIP). It contains five key steps. Firstly, bands selection of LHHSI is used to reduce spectral dimension of LHHSIs. Second, we extract candidate SSIPs from the LHHSIs. Third, we need to determine whether there exist potential target regions by using spectral curves of many selected key SSIPs. And next, the image which contains the potential target regions is divided into image blocks by using quad-tree segmentation, and then every image block is represented by a vector with BoW model based on the selected SSIPs. Finally, these image blocks are classified with SVM. During the classification, if the result is what we need, the quad-tree segmentation of the current block will be ended. The experimental results show that the proposed algorithm has a better performance than traditional algorithms.

Traditional radar target detection algorithms are mostly based on statistical theory. They have weak generalization capabilities for complex sea clutter environments and diverse target characteristics, and their detection performance would significantly be reduced. In this paper, the range-azimuth information obtained by scanning radar is converted into PPI image, and a Radar-YOLONet is proposed and used for radar image target detection. The model includes CSPDarknet53, SPP, PANet and Power Non-Maximum Suppression layers. The prediction frame coordinates, target category and corresponding confidence are directly given through the feature extraction network. The network structure strengthens the receptive field and attention distribution structure, and further improves the efficiency of network training. The verification using the constructed X-band navigation radar PPI image dataset shows that compared with the traditional constant false alarm rate detector and the two-stage Faster-RCNN algorithm, the proposed method has significantly improved detection performance under certain false alarm probability conditions. It is more suitable for various environment and target characteristics.

A fundamental task of radar is the detection of targets in noise and clutter. It is very difficult to detect weak targets in heavy sea clutter. In this paper we apply the discrete wavelet transform (WT), along with multilayer feedforward neural networks, to the detection of radar weak targets. The wavelet transform employs the Haar scaling function, which is well matched to the signals from targets. The hard threshold filter is adopted to remove sea clutter. The neural networks are trained with the back-propagation rule, which are used to detect weak targets. The simulated results show that the proposed method is very effective for radar detection of weak targets.