The spatial arrangement of various land cover types within a landscape is referred to as the Landscape pattern (LSP). An essential component of landscape ecology, LSP examination is significant for a variety of causes...
详细信息
The spatial arrangement of various land cover types within a landscape is referred to as the Landscape pattern (LSP). An essential component of landscape ecology, LSP examination is significant for a variety of causes, like species conservation, sustainable development, environmental monitoring, landscape planning, and management of accepted resources. The development of remotesensing (RS) images permits urban planners to additional systematically and economically. 0 identify the land use of a specified area on a slighter time scale. The objective of this study is to build up an artificial intelligence (AI)-based RS imageprocessing performance for LSP. This study, proposed a novel refined flamingo search-dynamic recurrent neural network (RFS-DRNN) to analyze the LSP. RS image data were gathered from landscape characteristics. The Discrete Wavelet Transform (DWT) utilizes pre-processed data to eliminate noise, though maintenance is important distinctiveness. Convolutional Neural Network (CNN) using extracted features from image data. RFS could be used to progress the constraint of a DRNN model that is used to analyze patterns in the landscape. It can be used to regulate an RNN's hyper parameters to enhance its ability to recognize and categorize landscape features. The results showed that the proposed method is effective at analyzing LSPs. The significance indicates that the proposed method has achieved superior performance in including accuracy [98.90%], precision [94.82%], recall [93.75%], and F1-score [95.29%]. The hierarchical land-cover mapping reveal process creates thorough LSP analysis possible by using satellite images and sophisticated algorithms. High training accuracy and decreasing training loss indicate effective model learning and generalization for landscape analysis. The execution times at the end highlight the important it is to maximize processing methods and computational capacity to build quick decisions while analyzing LSP.
We consider the security issues in a satellite remotesensing(SRS) and imageprocessing system. Specifically, we present a novel architecture of a secure SRS imagerecognition system by using a radio frequency(RF) fin...
We consider the security issues in a satellite remotesensing(SRS) and imageprocessing system. Specifically, we present a novel architecture of a secure SRS imagerecognition system by using a radio frequency(RF) fingerprint to identify whether a satellite user is authenticated or *** unauthenticated users, we address the method of adding perturbation into their received SRS images,
In order to overcome the low recall rate, peak signal-to-noise ratio and correct recognition rate of spatial pattern in traditional spatial patternrecognition methods, a landscape spatial patternrecognition method b...
详细信息
In order to overcome the low recall rate, peak signal-to-noise ratio and correct recognition rate of spatial pattern in traditional spatial patternrecognition methods, a landscape spatial patternrecognition method based on multi-source remotesensingimages was proposed. First, we obtain multi-source remotesensingimages of landscape architecture, use ORB algorithm to extract multi-source remotesensingimage feature points, and fuse multi-source remotesensingimages. Then, MSRCR algorithm is used to enhance the fused image, LOG edge detection operator is used to obtain the image edge, and MCR model is used to determine the landscape patch characteristics. Finally, the spatial patternrecognition model of landscape architecture is built, and the spatial patternrecognition results are obtained. The experimental results show that the maximum recall rate of this method is 97%, the maximum peak signal to noise ratio of image is 59.3 dB, and the correct recognition rate varies from 97% to 99%.
Why should we confine land cover classes to rigid and arbitrary definitions? Land cover mapping is a central task in remotesensingimageprocessing, but the rigorous class definitions can sometimes restrict the trans...
详细信息
Why should we confine land cover classes to rigid and arbitrary definitions? Land cover mapping is a central task in remotesensingimageprocessing, but the rigorous class definitions can sometimes restrict the transferability of annotations between datasets. Open vocabulary recognition, i.e. using natural language to define a specific object or pattern in an image, breaks free from predefined nomenclature and offers flexible recognition of diverse categories with amore general image understanding across datasets and labels. The open vocabulary framework opens doors to search for concepts of interest, beyond individual class boundaries. In this work, we propose to use Text As supervision for COntrastive Semantic Segmentation (TACOSS), and we design an open vocabulary semantic segmentation model that extends its capacities beyond that of a traditional model for land cover mapping: In addition to visual patternrecognition, TACOSS leverages the common sense knowledge captured by language models and is capable of interpreting the image at the pixel level, attributing semantics to each pixel and removing the constraints of a fixed set of land cover labels. By learning to match visual representations with text embeddings, TACOSS can transition smoothly from one set of labels to another and enables the interaction with remotesensingimages in natural language. Our approach combines a pretrained text encoder with a visual encoder and adopts supervised contrastive learning to align the visual and textual modalities. We explore several text encoders and label representation methods and compare their abilities to encode transferable land cover semantics. The model's capacity to predict a set of different land cover labels on an unseen dataset is also explored to illustrate the generalization capacities across domains of our approach. Overall, TACOSS is a general method and permits adapting between different sets of land cover labels with minimal computational overhead. Code
The processing technology of remotesensingimages has attracted more and more attention. Since remotesensingimage target detection technology has a wide range of applications in, terrain exploration and post-disast...
详细信息
The processing technology of remotesensingimages has attracted more and more attention. Since remotesensingimage target detection technology has a wide range of applications in, terrain exploration and post-disaster reconstruction, etc. remotesensingimage target detection refers to finding the target of interest in the remotesensingimage and giving the specific location, while remotesensingimage target recognition is the further classification of a certain target, which is a long-term concern in the field of remotesensingimageprocessing. Convolutional Neural Network CNN (Convolutional Neural Network) has achieved great success in the field of computer vision with its deep semantic features, and in recent years, it has been increasingly applied to remotesensingimage target detection and recognition tasks. Aiming at the task of remotesensingimage target detection, this paper proposes a new deep feature-based remotesensingimage target detection method. The depth feature extracted by CNN is used to extract the region of interest, and the target confirmation of the region of interest is carried out through multiple scales of CNN. This method does not require bounding box data for training, and improves the detection accuracy and reduces the false alarm rate.
Deep learning has revolutionized the remotesensingimageprocessing techniques over the past few years. Nevertheless, annotating high-quality samples is difficult and time-consuming, which limits the performance of d...
详细信息
Deep learning has revolutionized the remotesensingimageprocessing techniques over the past few years. Nevertheless, annotating high-quality samples is difficult and time-consuming, which limits the performance of deep neural networks because of insufficient supervision information. Aiming to solve this contradiction, we investigate the multimodal self-supervised learning (MultiSSL) paradigm for pre-training and classification of remotesensingimage. Specifically, the proposed self-supervised feature learning model consists of asymmetric encoder-decoder structure, in which deep unified encoder learns high-level key information characterizing multimodal remotesensing data and task-specific lightweight decoders are developed to reconstruct original data. To further enhance feature extraction capability, the cross-attention layers are utilized to exchange information contained in heterogeneous characteristics, thus learning more complementary information from multimodal remotesensing data. In fine-tuning stage, the pre-trained encoder and cross-attention layer serve as feature extractor, and leaned characteristics are combined with corresponding spectral information for land cover classification through a lightweight classifier. The self-supervised pre-training model can learn high-level key features from unlabeled samples, thereby utilizing the feature extraction capability of deep neural networks while reducing their dependence on annotated samples. Compared with existing classification paradigms, the proposed multimodal self-supervised pre-training and fine-tuning scheme achieves superior performance for remotesensingimage land cover classification.
Aiming at the difficulties in object detection and recognition in remotesensingimages caused by high background complexity, large scale variations of targets, and the presence of numerous small objects, an improved ...
详细信息
remotesensingimages play a crucial role in fields such as reconnaissance and early warning, intelligence analysis, etc. Due to factors such as climate, season, lighting, occlusion and even atmospheric scattering dur...
详细信息
remotesensingimages play a crucial role in fields such as reconnaissance and early warning, intelligence analysis, etc. Due to factors such as climate, season, lighting, occlusion and even atmospheric scattering during remotesensingimage acquisition, targets of the same model exhibit significant intra-class variability. This article applies deep learning technology to the field of military aircraft recognition in remotesensingimages and proposes a You Only Look Once Version 8 Small (YOLOv8s) remotesensingimage military aircraft recognition algorithm based on an attention mechanism-YOLOv8s-TDP (YOLOv8s+TripletAttention+dysample+PIoU). First, the TripletAttention attention module is used in the neck network, which captures cross-dimensional interactions and utilises a three-branch structure to calculate attention weights. This further enhances the network's ability to preserve details and restore colours in the process of image fusion. Secondly, an efficient dynamic upsampler, dysample, is used to achieve dynamic upsampling through point sampling, which improves the problems of detail loss, jagged edges, and image distortion that may occur with nearest neighbour interpolation. Finally, replacing the original model loss function with PIoU (Pixels Intersection over Union), IoU (Intersection over Union) is calculated at the pixel level to more accurately capture small overlapping areas, reduce missed detection rates, and improve accuracy. On the publicly available dataset The remotesensingimage Military Aircraft Target recognition Dataset(MAR20), our proposed YOLOv8s-TDP model achieved a Precision${\mathrm Precision} $ of 82.96%, Recall${\mathrm Recall} $ of 80.71%, mAP0.5$mA{P}_{0.5}$ of 87.11% and mAP0.5-0.95$mA{P}_{0.5 - 0.95}$ of 65.88%, outperforming the original YOLOv8s model, RT-DETR model, YOLOv5 series model, YOLOv7 series model, and YOLOv11 series model. Compared with the original YOLOv8s model, the YOLOv8s-TDP model improves Precision$Precision$ by 0
Unmanned aerial vehicle (UAV) remote-sensingimages present unique challenges to the object-detection task due to uneven object densities, low resolution, and drastic scale variations. Downsampling is an important com...
详细信息
Unmanned aerial vehicle (UAV) remote-sensingimages present unique challenges to the object-detection task due to uneven object densities, low resolution, and drastic scale variations. Downsampling is an important component of deep networks that expands the receptive field, reduces computational overhead, and aggregates features. However, object detectors using multi-layer downsampling result in varying degrees of texture feature loss for various scales in remote-sensingimages, degrading the performance of multi-scale object detection. To alleviate this problem, we propose a lightweight texture reconstructive downsampling module called TRD. TRD models part of the texture features lost as residual information during downsampling. After modeling, cascading downsampling and upsampling operators provide residual feedback to guide the reconstruction of the desired feature map for each downsampling stage. TRD structurally optimizes the feature-extraction capability of downsampling to provide sufficiently discriminative features for subsequent vision tasks. We replace the downsampling module of the existing backbone network with the TRD module and conduct a large number of experiments and ablation studies on a variety of remote-sensingimage datasets. Specifically, the proposed TRD module improves 3.1% AP over the baseline on the NWPU VHR-10 dataset. On the VisDrone-DET dataset, the TRD improves 3.2% AP over the baseline with little additional cost, especially the APS, APM, and APL by 3.1%, 8.8%, and 13.9%, respectively. The results show that TRD enriches the feature information after downsampling and effectively improves the multi-scale object-detection accuracy of UAV remote-sensingimages.
With the development of remotesensing technology, remotesensingimage is widely used in environmental monitoring, land use and other fields. However, traditional imagerecognition methods often face the problems of ...
详细信息
With the development of remotesensing technology, remotesensingimage is widely used in environmental monitoring, land use and other fields. However, traditional imagerecognition methods often face the problems of high consumption of computing resources and slow processing speed, especially in low-energy systems. Therefore, this paper aims to explore a low-energy multi-feature fusion support vector machine (SVM) model based on remotesensingimagerecognition. The characteristics of remotesensingimages and their application requirements in low energy consumption systems are analyzed, and a multi-feature extraction method combining spectral features, texture features and shape features is proposed. Aiming at the limitations of the traditional SVM model in processing high-dimensional data, an optimization algorithm is designed to reduce the computational complexity and improve the recognition accuracy of the model through dimensionality reduction and feature selection. Based on the model, a series of experiments were carried out on a low-energy hardware platform to test its performance in different scenarios. The experimental results show that the proposed multi-feature fusion SVM model has significantly improved recognition accuracy compared with the single feature method, and can effectively control the computing resource consumption, and can realize real-time recognition on low-energy systems.
暂无评论