We propose a double image encryption by using random binary encoding and gyrator transform. Two secret images are first regarded as the real part imaginary of complex function. Chaotic map is used for obtaining matrix. The function exchanged under control data. An iterative structure composed method designed employed enhancing security algorithm. parameters in chaotic transform serve keys this scheme. Some numerical simulations have been made, to demonstrate performance

Abstract Computer vision-based deep learning models are of great significance in industrial defect quality detection. Unlike natural objects, defects products typically quite small and exhibit highly uneven scales, resulting suboptimal performance conventional object detectors when encountered with complex Hence, this paper introduces an efficient progressive aggregation enhanced network (EPAE-Net) the goal strengthening detection scenarios. Firstly, a global context feature enhancement...

In order to preserve the spatial and spectral information of original panchromatic multispectral images, this article designs a loss function suitable for pan-sharpening four-layer convolutional neural network that could adequately extract features from source images. The major advantage study is designed does not need reference fused image, then proposed method make simulation data training. This big difference most existing methods. Moreover, takes into account characteristics remote...

The majority of existing deep learning pan-sharpening methods often use simulated degraded reference data due to the missing real fusion labels which affects performance. normally used convolutional neural network (CNN) can only extract local detail information well may cause loss important global contextual characteristics with long-range dependencies in fusion. To address these issues and fuse spatial spectral high quality from original panchromatic (PAN) multispectral (MS) images, this...

Clouds in optical remote sensing images cause spectral information change or loss, that affects image analysis and application. Therefore, cloud detection is of great significance. However, there are some shortcomings current methods, such as the insufficient extendibility due to using multiple bands, intense relying on manually determined thresholds, limited accuracy, especially for thin clouds complex scenes caused by low-level manual features. Combining above requirements efficiency...

In this article, a multiple-image encryption method based on the optical interference principle and phase-only mask (POM) multiplexing is proposed. During process, each secret image encoded into two analytically obtained POMs one computer-generated random POM, in which no iterative computation required. The taken from different images are then synthesized by POM further complex ciphertext images. silhouette problem that exists earlier principle-based approaches totally resolved proposal....

Cloud detection is of great significance for optical remote sensing images. Most existing cloud approaches often rely on many thresholds among multiple bands with a wide spectrum range, and normally just can be applied to specific satellite data. difficult in some sensor data limited number available spectral bands. To tackle this challenge, we propose method based random forest (RFCD) only the most common RGB NIR The RFCD normalizes different by calculating top-of-atmosphere (TOA)...

Clouds hinder the surface observation by optical remote sensing sensors. It is of great significance to detect clouds and non-clouds in images. Compared with traditional cloud detection methods, deep learning methods usually achieve promising results. Moreover, large-scale, high-quality labeled datasets can effectively improve accuracy generalization models. However, this incurs a deal label effort cost. In paper, we proposes method based on semi-supervised learning(SSL) active learning(AL)...

We present a hybrid configuration of joint transform correlator (JTC) and fractional (JFTC) for encryption purpose. The original input is encoded in the power spectrum distribution JFTC. In our experimental arrangement, an additional random phase mask (master key) holographically generated beforehand by Mach–Zehnder interferometer with JTC as object arm. order JFTC, together master key, can remarkably strengthen safety level encryption. Different from many previous digital-holography-based...

Road extraction from the high-resolution remote sensing image is significant for land planning, vehicle navigation, etc. The existing road methods normally need many preprocessing and subsequent optimization steps. Therefore, an automatic centerline method based on self-supervised learning framework proposed. This proposed does not to manually select training samples other steps, such as nonroad area removing. First, positive sample selection combining spectral shape features extract sample....

We propose a multiple-image encryption scheme, based on polarized light encoding and the interference principle of phase-only masks (POMs), in Fresnel-transform (FrT) domain. In this each secret image is converted into an intensity by encoding, where random key pixilated polarizer with angles are employed as keys. The encrypted images produced different convolved together then inverse Fresnel-transformed. Phase amplitude truncations used to generate asymmetric decryption phase-truncated FrT...

As the earth's only natural satellite, Moon has unique advantages as an earth observation platform because of its long distance and vast expanse lunar surface. To observe from Moon, looking vector direction variation regularity need to be investigated. In this paper, we first establish a Moon-based location algorithm reveal variations vector. algorithm, assuming sensor points earth, expressed azimuth elevation angles is calculated, intersection point between ellipsoid solved using relevant...

The spectral fidelity and the spatial richness enhancement are two primary objectives for deep learning pan-sharpening algorithms. Consequently, two-stream fusion architecture is utilized to focus on individual features of panchromatic (PAN) multi-spectral (MS) images. However, most existing networks have a high degree extracted feature redundancy large workload. Based these, we propose dual spatial-spectral network (DSSN) implement divided-objective fusion, with one stream fusing...

Remote-sensing image registration is an important prerequisite for many remote-sensing applications. The accuracy, efficiency, and automatic degree of have direct impacts on follow-up With the improving resolution, size data amount are constantly increasing. Meanwhile, with development application, performance also put forward increasing requirement. Therefore, this paper proposes fast method based angle matching edge point features (EPFs). First, original transformed by Haar wavelet to get...

Moon-based platform is a potential that can realize the observations of large-scale geoscience phenomenon. Unlike existing earth observation platforms, moon-based equipped on natural celestial body. Its position calculated by lunar and libration derived from planetary ephemeris. However, limited to astrometric model accuracy observational data, no ephemeris provide absolutely accurate this will lead platform's error. This letter investigates impacts error geolocation for sensor. We first...

Pansharpening is crucial for obtaining high-resolution multispectral images. Existing deep learning-based pansharpening networks rely on supervised learning with external reference labels. Due to the lack of actual fusion results labeling, simulated degraded data used original image as result label. This process steps are cumbersome, which also leads problem scale degradation, and relationship between before after degradation cannot represent real relationship. To address these limitations,...

It has already been confirmed that the traffic in high-speed terrestrial network presents self-similarity, but there is little research on self-similarity of satellite network.Considering time-varying topology and link status, this paper analyzes aggregation propagation self-similar between nodes network.Furthermore, a sort special node called ground gateway modeled, based which characteristics output input from passes into are analyzed.Theoretically analyses demonstrate after nodes, still...