The principle of the image fusion is to integrate complementary information heterogeneous images obtain a fused that more in line with visual effect human eyes. However, most decomposition methods cannot distinguish textures and edges an image, which easy produce halo artifacts around edges. In this paper, we proposed novel strategy (co-occurrence analysis shearlet transform, CAST) preprocess input depending on co-occurrence statistic generate base layer detail components. order improve...

Dynamic vision sensor (DVS) is a new type of image sensor, which has application prospects in the fields automobiles and robots. sensors are very different from traditional terms pixel principle output data. Background activity (BA) data will affect quality, but there currently no unified indicator to evaluate quality event streams. This paper proposes method eliminate background activity, performance index for evaluating filter performance: noise real (NIR) (RIN). The lower value, better...

Optical remote sensing images are widely used in the fields of feature recognition, scene semantic segmentation, and others. However, quality is degraded due to influence various noises, which seriously affects practical use images. As have more complex texture features than ordinary images, this will lead previous denoising algorithm failing achieve desired result. Therefore, we propose a novel image network (RSIDNet) based on deep learning approach, mainly consists multi-scale extraction...

Recently, optical remote-sensing images have been widely applied in fields such as environmental monitoring and land cover classification. However, due to limitations imaging equipment other factors, low-resolution that are unfavorable for image analysis often obtained. Although existing super-resolution algorithms can enhance resolution, these not specifically designed the characteristics of cannot effectively recover high-resolution images. Therefore, this paper proposes a novel algorithm...

Object detection in remote sensing plays a crucial role various ground identification tasks. However, due to the limited feature information contained within small targets, which are more susceptible being buried by complex backgrounds, especially extreme environments (e.g., low-light, motion-blur scenes). Meanwhile, event cameras offer unique paradigm with high temporal resolution and wide dynamic range for object detection. These advantages enable without intensity of light, perform better...

Object detection is a crucial task in the field of remote sensing. Currently, frame-based algorithms have demonstrated impressive performance. However, research on sensing applying event cameras has not yet been conducted. Meanwhile, there are still three issues to address: 1) Remote targets often disrupted by complex backgrounds, resulting poor performance, especially extremely challenging environments (e.g., low-light, motion blur, and occlusion scenarios). 2) Mainstream deep learning...

ABSTRACT Access control is a critical security measure to ensure that sensitive information and resources are accessed only by authorized users. However, attribute‐based access in the big data environment faces challenges such as large number of entity attributes, poor availability, difficulty manual labeling. In this paper, we focus on problem mining optimizing attributes unstructured propose method for textual based multi‐model collaboration. First, utilize unsupervised methods extract...

Due to the influence of atmospheric scattering, quality remote sensing images is degraded, which severely limits utility images. In this article, a novel dehazing algorithm for single image proposed based on low-rank and sparse prior (LSP). According an scattering model, dark channel hazy decomposed into two parts: direct attenuation with sparseness veil low rank. The obtained from overall decomposition rather than patches image; therefore, pixel changes local blocks have little prior....

The advantages of an event camera, such as low power consumption, large dynamic range, and data redundancy, enable it to shine in extreme environments where traditional image sensors are not competent, especially high-speed moving target capture lighting conditions. Optical flow reflects the target's movement information, detailed can be obtained using camera's optical information. However, existing neural network methods for prediction cameras has problems extensive computation high energy...

The event camera efficiently detects scene radiance changes and produces an asynchronous stream with low latency, high dynamic range (HDR), temporal resolution, power consumption. However, the large output data caused by imaging mechanism makes increase in spatial resolution of limited. In this paper, we propose a novel super-resolution (SR) network (EFSR-Net) based on deep learning approach to address problems poor visualization cameras. model is capable reconstructing high-resolution (HR)...

In this paper, we propose an unstructured road-free space detection method that integrates distance imaging information in the Transformer framework. The proposed network is FNS-Swin, which innovatively supplements features through data fusion to reduce dependence of networks applied computer vision on RGB image volume while improving accuracy segmentation algorithms. Regarding framework, adopted Swin Transformer’s sliding window structure. We reset number layers model better adapt learning...

Autofocus methods are conventionally based on capturing the same scene from a series of positions focal plane. As result, it has been difficult to apply this technique scanning remote sensing cameras where scenes change continuously. In order realize autofocus in cameras, novel method is investigated paper. Instead introducing additional mechanisms or optics, overlapped pixels adjacent CCD sensors plane employed. Two images, corresponding ground, can be captured at different times. Further,...

Edge computing is a new paradigm, which provides storage, computing, and network resources between the traditional cloud data center terminal devices. In this paper, we concentrate on application-driven task offloading problem in edge by considering strong dependencies of sub-tasks for multiple users. Our objective to joint optimize total delay energy generated applications, while guaranteeing quality services First, formulate tasks jointly delays consumption. Based that, propose novel...

The dynamic vision sensor (DVS) measures asynchronously change of brightness per pixel, then outputs an asynchronous and discrete stream spatiotemporal event information that encodes the time, location, sign changes. has outstanding properties compared to sensors traditional cameras, with very high range, temporal resolution, low power consumption, does not suffer from motion blur. Hence, have considerable potential for computer in scenarios are challenging cameras. However, visualization is...

With the vigorous development of industries such as self-driving, edge intelligence, and industrial Internet Things (IoT), amount type data generated are unprecedentedly large, users’ demand for high-quality services continues to increase. Edge computing has emerged a new paradigm, providing storage, computing, networking resources between traditional cloud centers end devices with solid timeliness. Therefore, resource allocation problem in online task offloading process is main area...

NA

Conventionally, high dynamic-range (HDR) imaging is based on taking two or more pictures of the same scene with different exposure. However, due to a high-speed relative motion between camera and scene, it hard for this technique be applied push-broom remote sensing cameras. For sake HDR in applications, present paper proposes an innovative method which can generate images without redundant image sensors optical components. Specifically, adopts area array CMOS (complementary metal oxide...

To address the challenge of no-reference image quality assessment (NR-IQA) for authentically and synthetically distorted images, we propose a novel network called Combining Convolution Self-Attention Image Quality Assessment (Conv-Former). Our model uses multi-stage transformer architecture similar to that ResNet-50 represent appropriate perceptual mechanisms in (IQA) build an accurate IQA model. We employ adaptive learnable position embedding handle images with arbitrary resolution. new...

The image decomposition strategy that extracts salient features from the source is crucial for fusion. To this end, we proposed a novel saliency-based infrared and visible In particular, latent low-rank representation (LatLRR) rolling guidance filter (RGF) are together employed to process images, which called DLatLRR_RGF. method, images first decomposed components base based on LatLRR, filtered by RGF. Then, final can be calculated difference between processed components. fusion rule...

Event cameras are the emerging bio-mimetic sensors with microsecond-level responsiveness in recent years, also known as dynamic vision sensors. Due to inherent sensitivity of event camera hardware light sources and interference from various external factors, types noises inevitably present camera's output results. This noise can degrade perception events performance algorithms for processing streams. Moreover, since is form address-event representation, efficient denoising methods...

Dynamic vision sensor is a kind of bioinspired sensor. It has the characteristics fast response, large dynamic range, and asynchronous output event stream. These make it have advantages that traditional image sensors do not in field tracking. The form stream, object information needs to be provided by relevant cluster. This article proposes method based on correlation index obtain object’s position, contour, other compatible with tracking methods. Experiments show this can position moving...

Event cameras, as bio-inspired visual sensors, offer significant advantages in their high dynamic range and temporal resolution for tasks. These capabilities enable efficient reliable motion estimation even the most complex scenes. However, these come with certain trade-offs. For instance, current event-based vision sensors have low spatial resolution, process of event representation can result varying degrees data redundancy incompleteness. Additionally, due to inherent characteristics...

To mitigate the influence of satellite platform vibrations on space camera imaging quality, a novel approach is proposed to detect vibration parameters based correlation rolling-shutter CMOS. In meantime, restoration method address image degradation CMOS caused by such proposed. The parameter detection utilizes time-sharing and row-by-row principle obtain relative offset comparing two frames images from continuous imaging. Then, camera's are derived fitting curve offset. According detected...