PD-SegNet: Semantic Segmentation of Small Agricultural Targets in Complex Environments

Vehicle detection is an important technology in au-tonomous driving, for which high detection accuracy and real-time performance are often required. The YOLOv5-GE vehicle detection algorithm is proposed to address the situation that the YOLOv5 vehicle detection model has false detection and missed detection for small and dense targets in complex environments. The global attention mechanism is added to the backbone net-work of the YOLOOv5 model, which is composed of two inde-pendent submodules of channel attention and convolutional spa-tial attention, which prevents the loss of information to a certain extent and amplifies the interaction of global dimensions. Second-ly, the training process is optimized using the Focal-EloU loss function to replace the GloU loss function, which improves the accuracy of vehicle detection. Finally, the proposed YOLOv5-GE algorithm and the YOLOv5 algorithm are subjected to a con-trolled experiment on the KITTI dataset. The experimental re-sults show that the YOLOv5-GE algorithm achieves an average accuracy of 86% while maintaining real-time performance, which is 2.5% higher than that of the YOLOv5 algorithm, and can im-prove the detection accuracy of small and dense targets in com-plex environments.

In the task of using robots to capture end-of-life cars, the position of the vehicle door frame needs to be grasped. A fast and accurate positioning of the vehicle door frame is the key to realizing the automatic car grasping process. Traditional methods for locating and grasping scrap cars rely heavily on manual operations and suffer from low grasping efficiency and poor accuracy. Therefore, this paper proposes a binocular vision robot vehicle door frame spatial localization method based on the improved YOLOv4. This method includes a lightweight and efficient feature fusion target detection network in complex environments, and the target detection results are combined with an enhanced SURF feature–matching method to locate the vehicle door frame position. To simplify the network structure, MobileNetv3 is used instead of the backbone network CSPDarknet53, and deep separable convolution is used in the network. To increase the sensitivity of the network to vehicle door frame targets in complex environments, an improved convolutional block attention module is added to the pyramid attention with simple network backbones. Moreover, adaptive spatial feature fusion is introduced into the network to fully use the features at different scales for more effective feature fusion. Compared with YOLOv4, the number of network parameters is reduced by 73.8%, the mAP is improved by 1.35%, and the detection speed is increased by 28.7%. The experimental results demonstrate that the positioning accuracy of the system is 0.745 mm, which meets the positioning measurement error of less than 1 cm required for the vehicle door frame. The paper also compares our findings with other network models. The results show that the method achieves a good balance between detection speed and detection accuracy, satisfying the task of identifying vehicle door frames in complex environments with good detection results.

MXene-based thermal camouflage materials have gained increasing attention due to their low emissivity, however, the poor anti-oxidation restricts their potential applications under complex environments. Various modification methods and strategies, e.g., the addition of antioxidant molecules and fillers have been developed to overcome this, but the realization of long-term, reliable thermal camouflage using MXene network (coating) with excellent comprehensive performance remains a great challenge. Here, a MXene-based hybrid network comodified with hyaluronic acid (HA) and hyperbranched polysiloxane (HSi) molecules is designed and fabricated. Notably, the presence of appreciated HA molecules restricts the oxidation of MXene sheets without altering infrared stealth performance, superior to other water-soluble polymers; while the HSi molecules can act as efficient cross-linking agents to generate strong interactions between MXene sheets and HA molecules. The optimized MXene/HA/HSi composites exhibit excellent mechanical flexibility (folded into crane structure), good water/solvent resistance, and long-term stable thermal camouflage capability (with low infrared emissivity of ≈0.29). The long-term thermal camouflage reliability (≈8 months) under various outdoor weathers and the scalable coating capability of the MXene-coated textile enable them to disguise the IR signal of various targets in complex environments, indicating the great promise of achieved material for thermal camouflage, IR stealth, and counter surveillance.

To meet the urgent demand of low-observable moving target detection in complex environments, a novel method of Frequency Diverse Array (FDA) radar signal processing method based on Space-Rang-Doppler Focusing (SRDF) is proposed in this paper. The current development status of the FDA radar, the design of the array structure, beamforming, and joint estimation of distance and angle are systematically reviewed. The extra degrees of freedom provided by FDA radar are fully utilizsed, which include the Degrees Of Freedom (DOFs) of the transmitted waveform, the location of array elements, correlation of beam azimuth and distance, and the long dwell time, which are also the DOFs in joint spatial (angle), distance, and frequency (Doppler) dimensions. Simulation results show that the proposed method has the potential of improving target detection and parameter estimation for weak moving targets in complex environments and has broad application prospects in clutter and interference suppression, moving target refinement, etc..

High-resolution millimeter wave radar sensor plays an important role in the process of detecting and classifying targets in complex environments. To process the radar data, clustering is often used to group the measured data. However, due to the different numbers and densities of available target points in different scenarios, data with high sparsity characteristics can cause incomplete target detection in the process of classifying targets, which is very challenging. Therefore, this study introduces a target detection method based on inter-frame DBSCAN clustering. In order to improve the accuracy of target detection, we employ a modified DBSCAN algorithm to cluster the inter-frame data by fusing Doppler features. The method uses a multi-frame merging process to improve the single-frame clustering accuracy, and uses frame sequence features to solve the multi-target noise problem. The effectiveness of the proposed method is verified using real-world dataset, as the clustering accuracy is improved compared to other methods, indicating the advantages of the proposed method.

The rapid progress of microwave imaging technology has made conventional camouflage materials with fixed absorption performance ineffective. As the imaging band expands, camouflage materials capable of broadband operation, especially in the S and the C band, dynamic modulation are required to hide targets in complex environments. Here, we propose a dynamically modulated camouflage metasurface employing deep-subwavelength slots to enhance multiband modulation capability. By varying the vertical displacement of the structure, reflectivity can be modulated from below -10dB to near 0dB over 2.7-19.1GHz (150.4% relative bandwidth), while maintaining insensitivity to the incidence angle and polarization. An equivalent interface-impedance model is established to reveal the mechanism of slot-enhanced low-frequency resonance. The laser processing parameters are optimized to reduce the slot width to 25μm (λ max/4440), enabling broadband dynamic modulation, as experimentally verified. Benefiting from its broadband dynamic modulation performance extending to the S and the C band, the proposed camouflage metasurface demonstrates potential for countering emerging microwave imaging technologies.

The traditional threat assessment (TA) methods are confronted with the problems that most of them only focus on the static threat of a single target and the threshold of threat degree fusion is hard to set. For this reason, a novel DBN (dynamic Bayesian network) based TA algorithm is proposed. In the proposed algorithm, firstly, DBN is constructed with various factors, i.e., terrain, weather, time, relative strength, distance and velocity vector, for the TA of group targets. Then, the fast approximate inference is implemented according to Markov property. Finally, the probabilities of threat degrees are integrated into the continuous threat index and the discrete threat degree. Simulation results show that the proposed algorithm can be used to reliably and dynamically evaluate the threat of group targets in complex environment.

Revolutionary entrapment model of uniformly distributed swarm robots in morphogenetic formation

Target detection and tracking represent key challenges facing miniature fixed-wing unmanned aerial vehicles (UAVs), particularly at high cruising speeds. Therefore, this paper proposes a vision-based target detection and tracking algorithm that systematically couples two mainstream methods, namely, you only look once (YOLO) and kernel correlation filter (KCF) algorithms. This combination enables small fixed-wing UAVs to achieve reliable target detection and rapid target tracking. A customized vision-guidance module is constructed to implement this algorithm, and a dual-thread execution mechanism is developed to ensure that the computational resources are used effectively. A miniature fixed-wing UAV experimental platform is also constructed and evaluated. Flight experiments are performed, and the results demonstrate that the developed algorithm can achieve satisfactory detection and tracking accuracy for stationary and moving ground targets in complex environments.

To enhance the recognition of camouflaged targets in complex environments, this work employs polarization hyperspectral imaging to systematically examine the spectral response characteristics of four representative camouflage materials across multiple polarization channels. The contribution of each polarization channel (S1, S2, DOLP, and AOLP) is quantified relative to the non-polarized baseline using four spectral dissimilarity metrics: spectral angle (SA), root-mean-square error, spectral information divergence, and the red-edge spectral difference index. The results demonstrate that polarization information effectively enhances the spectral distinction between the camouflage materials and the grassland background, with the angle of linear polarization (AOLP) and the polarization difference channels (S1 and S2) exhibiting the most prominent effects. Based on these findings, channel fusion strategies centered on AOLP, S1, and S2 are proposed for different materials, and quantitative thresholds are established to improve the accuracy of practical camouflage target identification. This work provides a theoretical basis and methodological reference for the real-world application of polarization hyperspectral imaging in camouflage target detection.

PE-Transformer: Path enhanced transformer for improving underwater object detection

Accurate bud detection is a prerequisite for automatic tea picking and yield statistics; however, current research suffers from missed detection due to the variety of singleness and false detection under complex backgrounds. Traditional target detection models are mainly based on CNN, but CNN can only achieve the extraction of local feature information, which is a lack of advantages for the accurate identification of targets in complex environments, and Transformer can be a good solution to the problem. Therefore, based on a multi-variety tea bud dataset, this study proposes RT-DETR-Tea, an improved object detection model under the real-time detection Transformer (RT-DETR) framework. This model uses cascaded group attention to replace the multi-head self-attention (MHSA) mechanism in the attention-based intra-scale feature interaction (AIFI) module, effectively optimizing deep features and enriching the semantic information of features. The original cross-scale feature-fusion module (CCFM) mechanism is improved to establish the gather-and-distribute-Tea (GD-Tea) mechanism for multi-level feature fusion, which can effectively fuse low-level and high-level semantic information and large and small tea bud features in natural environments. The submodule of DilatedReparamBlock in UniRepLKNet was employed to improve RepC3 to achieve an efficient fusion of tea bud feature information and ensure the accuracy of the detection head. Ablation experiments show that the precision and mean average precision of the proposed RT-DETR-Tea model are 96.1% and 79.7%, respectively, which are increased by 5.2% and 2.4% compared to those of the original model, indicating the model’s effectiveness. The model also shows good detection performance on the newly constructed tea bud dataset. Compared with other detection algorithms, the improved RT-DETR-Tea model demonstrates superior tea bud detection performance, providing effective technical support for smart tea garden management and production.

针对图像序列中运动目标检测、跟踪的难点问题，提出了一种实时运动目标检测与跟踪算法。该算法基于自适应背景建模，获取运动目标背景模型和前景图像，从而实现运动目标检测；通过建立运动目标的位置、大小、形状以及颜色分布模型，构造运动目标全局匹配函数，结合目标活力特征，实现多运动目标连续匹配和跟踪。实验结果表明，相对于传统的运动目标跟踪方法，本文方法明显减少了运算时间，增强了环境适应性，实现了复杂场景下运动目标的准确检测和稳定跟踪，对非刚性目标的形变、旋转具有较强的鲁棒性。In this paper, we propose a real-time mobile target detection and tracking algorithm for challenges of mobile target detection and tracking in sequential images. This algorithm based on the adaptive background modeling obtains background model and front-view images of mobile targets, which is the way to achieve target detection. Continuous matching and tracking of multiple mobile targets are realized through constructing position, size, shape and color distribution of the mobile targets, defining a global matching function for those targets, and associating their vitality characteristics. It is demonstrated by experiments that the algorithm presented in this paper, compared to the traditional methods of mobile target tracking, significantly reduces the computation time, improves adaptive feature to environments, achieves accurate detection and robust tracking of mobile targets in complex environments, and shows strong robustness to deformation and rotation of non-rigid targets.

Ultrawideband (UWB) radar technology has emerged as one of the preferred choices for through-wall detection due to its high range resolution and good penetration. The resolution is a result of high bandwidth of UWB radar and helpful for better separation of multiple targets in complex environment. Detection of human targets through a wall is interesting in many applications. One significant characteristic of human is the periodic motion, such as breathing and limb movement. In this letter, we apply the UWB radar system in through-wall human detection and present the methods based on fast Fourier transform and S transform to detect and identify the human's life characteristic. In particular, we can extract the center frequencies of life signals and locate the position of human targets from experimental data with high accuracy. Compared with other research studies in through-wall detection, this letter is concentrated in the processing and identifying of the life signal under strong clutter. It has a high signal-to-noise ratio and simpler to implement in complex environment detection. We can use the method to search and locate the survivor trapped under the building debris during earthquake, explosion, or fire.

Objective: In the task of camouflage target detection, there is a problem that the target is highly integrated with the complex environment background, which is difficult to identify and leads to false detection and missed detection. A target detection algorithm CM-YOLOv5s is proposed for camouflage characteristics. Method: The algorithm uses YOLOv5s as the basic framework. First, a coordinated attention mechanism is embedded in the backbone feature extraction network, which enhances the network’s ability to extract camouflaged target features, weakens the attention to the surrounding background, and effectively improves the algorithm’s anti-background. Interference ability; secondly, the Mixup data enhancement strategy is used to simulate overlapping occlusion scenarios, which further strengthens the network model’s learning ability for complex samples. Results: The training and verification were carried out on the self-made Military Camouflage Target Dataset (MCTD), and the precision, recall, and average mAP of the improved CM-YOLOv5s algorithm reached 95.9%, 87.1%, and 94.1, respectively. %, compared with the original YOLOv5s model, the average accuracy rate is improved by 3.8 percentage points. Conclusion: The improved algorithm has better detection effect, and realizes accurate identification and rapid positioning of military camouflage targets in complex environments.