Search

Poyang Lake is the largest freshwater lake in China and plays a significant ecological role. Deep-learning-based video surveillance can effectively monitor bird species on the lake, contributing to the local biodiversity preservation. To address the challenges of multi-scale object detection against complex backgrounds, such as a high density and severe occlusion, we propose a new model known as the YOLOv8-bird model. First, we use Receptive-Field Attention convolution, which improves the model’s ability to capture and utilize image information. Second, we redesign a feature fusion network, termed the DyASF-P2, which enhances the network’s ability to capture small object features and reduces the target information loss. Third, a lightweight detection head is designed to effectively reduce the model’s size without sacrificing the precision. Last, the Inner-ShapeIoU loss function is proposed to address the multi-scale bird localization challenge. Experimental results on the PYL-5-2023 dataset demonstrate that the YOLOv8-bird model achieves precision, recall, mAP@0.5, and mAP@0.5:0.95 scores of 94.6%, 89.4%, 94.8%, and 70.4%, respectively. Additionally, the model outperforms other mainstream object detection models in terms of accuracy. These results indicate that the proposed YOLOv8-bird model is well-suited for bird detection and counting tasks, which enable it to support biodiversity monitoring in the complex environment of Poyang Lake.

Edge detection for PolSAR images has demonstrated its importance in various applications such as segmentation and classification. Although there are many edge detectors which have demonstrated an impressive ability to achieve accurate edge detection results, these methods only focus on edge detection in a single-date PolSAR image. However, a single-date PolSAR image cannot fully characterize the changes in scattering mechanisms of land cover in different growth cycles, resulting in some omissions of the true edges. In this paper, we propose a novel edge detection method for multi-temporal PolSAR images based on the SIRV model and an SDAN-based 3D Gaussian-like kernel. The spherically invariant random vector (SIRV) and span-driven adaptive neighborhood (SDAN) improve the estimation accuracy of the average covariance matrix (ACM) in terms of data representation and spatial support, respectively. We propose an SDAN-based 2D Gaussian kernel to accurately extract the edge strength of single-date PolSAR images. Then, we design a 1D convolution kernel in the temporal dimension to smooth fluctuations in the edge strength of multi-temporal PolSAR images. The SDAN-based 2D Gaussian kernels in the X- and Y-directions are combined with the 1D convolution kernel in the Z-direction to form an SDAN-based 3D Gaussian-like kernel. In addition, we design an adaptive hysteresis threshold method to optimize the edge map. The performance of our proposed method is presented and analyzed on two real multi-temporal PolSAR datasets, and the results demonstrate that the proposed edge detector achieves a better performance than other edge detectors, particularly for crop regions with time-varying scattering mechanisms.

SAR image registration is the basis for applications such as change detection, image fusion, and three-dimensional reconstruction. Although CNN-based SAR image registration methods have achieved competitive results, they are insensitive to small displacement errors in matched point pairs and do not provide a comprehensive description of keypoint information in complex scenes. In addition, existing keypoint detectors are unable to obtain a uniform distribution of keypoints in SAR images with complex scenes. In this paper, we propose a texture constraint-based phase congruency (TCPC) keypoint detector that uses a rotation-invariant local binary pattern operator (RI-LBP) to remove keypoints that may be located at overlay or shadow locations. Then, we propose a Siamese dense capsule network (SD-CapsNet) to extract more accurate feature descriptors. Then, we define and verify that the feature descriptors in capsule form contain intensity, texture, orientation, and structure information that is useful for SAR image registration. In addition, we define a novel distance metric for the feature descriptors in capsule form and feed it into the Hard L2 loss function for model training. Experimental results for six pairs of SAR images demonstrate that, compared to other state-of-the-art methods, our proposed method achieves more robust results in complex scenes, with the number of correctly matched keypoint pairs (NCM) at least 2 to 3 times higher than the comparison methods, a root mean square error (RMSE) at most 0.27 lower than the compared methods.

Change detection (CD) is an important research topic in remote sensing, which has been applied in many fields. In the paper, we focus on the post-processing of difference images (DIs), i.e., how to further improve the quality of a DI after the initial DI is obtained. The importance of DIs for CD problems cannot be overstated, however few methods have been investigated so far for re-processing DIs after their acquisition. In order to improve the DI quality, we propose a global and local graph-based DI-enhancement method (GLGDE) specifically for CD problems; this is a plug-and-play method that can be applied to both homogeneous and heterogeneous CD. GLGDE first segments the multi-temporal images and DIs into superpixels with the same boundaries and then constructs two graphs for the DI with superpixels as vertices: one is the global feature graph that characterizes the association between the similarity relationships of connected vertices in the multi-temporal images and their changing states in a DI, the other is the local spatial graph that exploits the change information and contextual information of the DI. Based on these two graphs, a DI-enhancement model is built, which constrains the enhanced DI to be smooth on both graphs. Therefore, the proposed GLGDE can not only smooth the DI but also correct the it. By solving the minimization model, we can obtain an improved DI. The experimental results and comparisons on different CD tasks with six real datasets demonstrate the effectiveness of the proposed method.

The traditional sensorless control system of permanent magnet synchronous motor (PMSM) has the problems of low estimation accuracy and poor anti-interference ability. Moreover, the position estimation performance is subjected to position harmonic ripples caused by inverter nonlinearities and flux spatial harmonics. To optimize the dynamic performance of the PMSM sensorless control system, this paper proposes a sensorless control scheme that combines integral backstepping control with enhanced linear extended state observer (ELESO). The ELESO consists of two linear extended state observers (LESOs), which estimate the internal and external disturbances of the system, to improve the estimation accuracy of rotor position. Then, the integral backstepping controller processes the estimated rotor position and speed information to obtain d and q-axis voltages. The sensorless control scheme is implemented in the Matlab/Simulink and verified by experiments. The simulation and experiment show that the scheme can effectively suppress load interference and improve control accuracy.

Although the deep neural network (DNN) has shown a powerful ability in hyperspectral image (HSI) classification, its learning requires a large number of labeled training samples; otherwise, it is prone to over-fitting and has a poor classification performance. However, this requirement is impractical for HSIs due to the difficulty in obtaining class labels. To make DNNs suitable for HSI classification with few labeled samples, we propose a graph-based deep multitask few-shot learning (GDMFSL) framework that learns the intrinsic relationships among all samples (labeled and unlabeled) of HSIs with the assistance of graph information to alleviate the over-fitting caused by few labeled training samples. Firstly, a semi-supervised graph is constructed to generate graph information. Secondly, a deep multitask network (DMN) is designed, which contains two subnetworks (tasks): a classifier subnetwork for learning class information from labeled samples and a Siamese subnetwork for learning sample relationships from the semi-supervised graph. To effectively learn graph information, a loss function suitable for the Siamese subnetwork is designed that shortens (and expands) the distance between the target sample and its nearest (and farthest) neighbors. Finally, since the number of training samples of the two subnetworks is severely imbalanced, a multitask few-shot learning strategy is designed to make two subnetworks converge simultaneously. Experimental results on the Indian Pines, University of Pavia and Salinas datasets demonstrate that GDMFSL achieves a better classification performance relative to existing competitors in few-shot settings. In particular, when only five labels per class are involved in training, the classification accuracy of GDMFSL on the three datasets reaches 87.58%, 86.42% and 98.85%, respectively.

The controller design of hypersonic flight vehicles is a challenging task, especially when its flexible states are immeasurable. Unfortunately, the flexible states are difficult to measure directly. In this paper, an angular-accelerometer-based method for the estimation of flexible states is proposed. By adding a pitch angel angular accelerometer and designing an Extended Kalman Filter-based online estimation method, the flexible states could be obtained in real time. Then, based on the estimated flexible states, a stable inversion-based controller-design method was utilized, and a robust tracking controller was designed for hypersonic flight vehicles. The proposed method provides an effective means of estimating flexible states and conducting the observer-based controller design of hypersonic flight vehicles. Finally, a numeral simulation is given to show the effectiveness of the proposed control method.

Population based search techniques have been developed and applied to wide applications for their good performance, such as the optimization of the unmanned aerial vehicle (UAV) path planning problems. However, the search for optimal solutions for an optimization problem is usually expensive. For example, the UAV problem is a large-scale optimization problem with many constraints, which makes it hard to get exact solutions. Especially, it will be time-consuming when multiple UAV problems are waiting to be optimized at the same time. Evolutionary multi-task optimization (EMTO) studies the problem of utilizing the population-based characteristics of evolutionary computation techniques to optimize multiple optimization problems simultaneously, for the purpose of further improving the overall performance of resolving all these problems. EMTO has great potential in solving real-world problems more efficiently. Therefore, in this paper, we develop a novel EMTO algorithm using a classical PSO algorithm, in which the developed knowledge transfer strategy achieves knowledge transfer between task by synthesizing the transferred knowledges from a selected set of component tasks during the updating of the velocities of population. Two knowledge transfer strategies are developed along with two versions of the proposed algorithm. The proposed algorithm is compared with the multifactorial PSO algorithm, the SREMTO algorithm, the popular multifactorial evolutionary algorithm and a classical PSO algorithm on nine popular single-objective MTO problems and six five-task MTO problems, which demonstrates its superiority.

Deep neural networks have achieved significant development and wide applications for their amazing performance. However, their complex structure, high computation and storage resource limit their applications in mobile or embedding devices such as sensor platforms. Neural network pruning is an efficient way to design a lightweight model from a well-trained complex deep neural network. In this paper, we propose an evolutionary multi-objective one-shot filter pruning method for designing a lightweight convolutional neural network. Firstly, unlike some famous iterative pruning methods, a one-shot pruning framework only needs to perform filter pruning and model fine-tuning once. Moreover, we built a constraint multi-objective filter pruning problem in which two objectives represent the filter pruning ratio and the accuracy of the pruned convolutional neural network, respectively. A non-dominated sorting-based evolutionary multi-objective algorithm was used to solve the filter pruning problem, and it provides a set of Pareto solutions which consists of a series of different trade-off pruned models. Finally, some models are uniformly selected from the set of Pareto solutions to be fine-tuned as the output of our method. The effectiveness of our method was demonstrated in experimental studies on four designed models, LeNet and AlexNet. Our method can prune over 85%, 82%, 75%, 65%, 91% and 68% filters with little accuracy loss on four designed models, LeNet and AlexNet, respectively.

This Special Issue is focused on breakthrough developments in the field of Wireless Systems and Networks in the IoT. The selected contributions report current scientific progress in a wide range of topics covering clock error compensation in sensor networks, backscatter communication networks, Radio-Frequency Identification (RFID)-based inventory management, resource allocation in Long-Term Evolution (LTE)/LTE-A, (Long Range Wide-Area Network (LoRaWAN) modeling and key generation for the IoT.