A Multi‐granularity Parallel Model for Unified Remote Sensing Image Processing WebServices

This article made a detailed introduction to the design and implementation of image processing algorithm framework, with its focus on data-driven, image processing algorithm, UI and processing dialog optimization. Implemented by the C++ programming language, the final framework provides a new way and reference to the Customizable development of remote sensing image processing algorithm.The remote sensing image processing system is a set of technologies and methodoloies that can add enormous value in handling the problem of the earth sciences and related fields.Most practitioners are, primarily, professionals in fields such as hydrology, agriculture, engineering, forestry, geology and urban studies; giving them the writing remote sensing processing algorithm skills considerably enhances their expertise and capablities.

Remote-sensing image processing applications demand large-scale, collaborative processing and storage capabilities. Computational grid can provide abundant powerful computational resources with sharing and cooperation in grid environment. This paper gives an introduction to a collaborative processing mechanism based on workflow technology implemented in ImageGrid. The main idea is to encapsulate all parallel algorithms of remote-sensing image processing into grid services and recombine them with grid workflow to satisfy more and more complex remote-sensing image processing applications. In particular, we describe the architecture of the workflow-based grid system and the processing flow. We also give out the algorithm classification and an example of remote-sensing image processing grid application.

SIGNAL PROCESSING FOR REMOTE SENSING On the Normalized Hilbert Transform and Its Applications in Remote Sensing Steven R. Long and Norden E. Huang Statistical Pattern Recognition and Signal Processing in Remote Sensing Chi Hau Chen A Universal Neural Network-Based Infrasound Event Classifier Fredric M. Ham and Ranjan Acharyya Construction of Seismic Images by Ray Tracing Enders A. Robinson Multi-Dimensional Seismic Data Decomposition by Higher Order SVD and Unimodal ICA Nicholas Le Bihan, Valeriu Vrabie, and Jerome I. Mars Application of Factor Analysis in Seismic Profiling Zhenhai Wang and Chi Hau Chen Kalman Filtering for Weak Signal Detection in Remote Sensing Stacy L. Tantum, Yingyi Tan, and Leslie M. Collins Relating Time-Series of Meteorological and Remote Sensing Indices to Monitor Vegetation Moisture Dynamics J. Verbesselt, P. Joensson, S. Lhermitte, I. Jonckheere, J. van Aardt, and P. Coppin Use of a Prediction-Error Filter in Merging Highand Low-Resolution Images Sang-Ho Yun and Howard Zebker Blind Separation of Convolutive Mixtures for Canceling Active Sonar Reverberation Fengyu Cong, Chi Hau Chen, Shaoling Ji, Peng Jia, and Xizhi Shi Neural Network Retrievals of Atmospheric Temperature and Moisture Profiles from High-Resolution Infrared and Microwave Sounding Data William J. Blackwell Satellite-Based Precipitation Retrieval Using Neural Networks, Principal Component Analysis, and Image Sharpening Frederick W. Chen IMAGE PROCESSING FOR REMOTE SENSING Polarimetric SAR Techniques for Remote Sensing of the Ocean Surface Dale L. Schuler, Jong-Sen Lee, and Dayalan Kasilingam MRF-Based Remote-Sensing Image Classification with Automatic Model Parameter Estimation Sebastiano B. Serpico and Gabriele Moser Random Forest Classification of Remote Sensing Data Sveinn R. Joelsson, Jon Atli Benediktsson, and Johannes R. Sveinsson Supervised Image Classification of Multi-Spectral Images Based on Statistical Machine Learning Ryuei Nishii and Shinto Eguchi Unsupervised Change Detection in Multi-Temporal SAR Images Lorenzo Bruzzone and Francesca Bovolo Change-Detection Methods for Location of Mines in SAR Imagery Maria Tates, Nasser Nasrabadi, Heesung Kwon, and Carl White Vertex Component Analysis: A Geometric-Based Approach to Unmix Hyperspectral Data Jose M.B. Dias and Jose M.P. Nascimento Two ICA Approaches for SAR Image Enhancement Chi Hau Chen, Xianju Wang, and Salim Chitroub Long-Range Dependence Models for the Analysis and Discrimination of Sea-Surface Anomalies in Sea SAR Imagery Massimo Bertacca, Fabrizio Berizzi, and Enzo Dalle Mese Spatial Techniques for Image Classification Selim Aksoy Data Fusion for Remote-Sensing Applications Anne H.S. Solberg The Hermite Transform: An Efficient Tool for Noise Reduction and Image Fusion in Remote-Sensing Boris Escalante-Ramirez and Alejandra A. Lopez-Caloca Multi-Sensor Approach to Automated Classification of Sea Ice Image Data A.V. Bogdanov, S. Sandven, O.M. Johannessen, V.Yu. Alexandrov, and L.P. Bobylev Use of the Bradley-Terry Model to Assess Uncertainty in an Error Matrix from a Hierarchical Segmentation of an ASTER Image Alfred Stein, Gerrit Gort, and Arko Lucieer SAR Image Classification by Support Vector Machine Michifumi Yoshioka, Toru Fujinaka, and Sigeru Omatu Quality Assessment of Remote Sensing Multi-Band Optical Images Bruno Aiazzi, Luciano Alparone, Stefano Baronti, and Massimo Selva Index

Nowadays, the increase of quality and resolution of remote sensing image will lead to the data's capacity extremely increase; A single computer, even a very expensive minicomputer or a PC server, is difficult to meet the disk space, memory space and processing time requirements when faced with this situation. Cloud computing can give us a new resolution to these problems. In this study, we proposed a Hadoop-based remote sensing image processing system (HBRSIPS), on the one hand, it re-defines a storage structure to adapt the characters of remote sensing images, on the other hand, We construct Java program classes based on MapReduce architecture, remote sensing image processing algorithms can be easily applied to the cloud computing environment. The experiments show, the proposed system can save remote sensing image into Hadoop's distribute storage environment and algorithms based on cloud environment were faster than single process algorithm.

Frequent and accurate estimation of forest structure parameters, such as number of trees per hectare or total height, are mandatory for sustainable forest management. Unmanned aircraft system (UAS) equipped with inexpensive sensors can be used to monitor and measure forest structure. The detailed information provided by the UAS allows tree level forest inventory. However, tree identification depends on a variety of parameters defining the image processing and tree segmentation algorithms. The objective of our study was to identify parameter combinations that accurately delineated trees and their heights. We evaluated the impact of different tree segmentation and point cloud generation algorithms on forest inventory from imagery collected with a UAS over a mature Douglas-fir plantation forest. We processed the images with two commonly used commercial software packages, Agisoft PhotoScan and Pix4Dmapper, both implementing image processing algorithms called Structure from Motion. For each software we generated photogrammetric point clouds by varying the parameters defining the implementation. We segmented individual trees and heights using three tree algorithms: Variable Window Filter, Graph-Theoretical, and Watershed Segmentation. We assessed the impact of image processing algorithms on forest inventory by comparing the estimated trees with trees manually identified from the point clouds. We found that the type of tree segmentation and image processing algorithms have a significant effect in accurately identifying trees. For tree height estimation, we found strong evidence that image processing algorithms had significant effects, whereas tree segmentation algorithms did not significantly affect tree height estimation.These findings may be of interest to others that are using high-resolution spatial imagery to estimate forest inventory parameters.

Abstract: Deep learning algorithms have shown promising results for different image processing tasks, particularly in remote sensing & image recognition. Till now many studies have been carried out on image processing, which brings a new paradigm of innovative capabilities under the umbrella of intelligent remote sensing and computer vision. Accordingly, quantum processing algorithms have proved to efficiently solve some issues that are undetectable to classical algorithms and processors. Keeping that in mind, a Quantum Convolutional Neural Network (QCNN) architecture along with Hybrid Quantum filters would be utilized supported by cloud computing infrastructures and data centers to provide a broad range of complex AI services and high data availability. This research summaries the conventional techniques of Classical and Quantum Deep Learning and it’s research progress on realworld problems in remote sensing image processing as a comparative demonstration. Last but not least, we evaluate our system by training on Street View House Numbers datasets in order to highlight the feasibility and effectiveness of using Quantum Deep Learning approach in image recognition and other similar applications. Upcoming challenges and future research areas on this spectrum are also discussed.

A comparative study between two types of tool wear monitoring systems for milling processes is introduced in this paper. The suggested sensory fusion approach includes the implementation of an infrared camera, in addition to force, vibration, sound and acoustic emission sensors. The majority of the research work available in literature and industry focuses on using one dimensional signals, such as force, vibration. Two dimensional data, such as infrared and visual images, are limited in literature in relation to machining operations. This work compares between one dimensional and two dimensional data for the development of a tool condition monitoring system for milling processes. The paper presents a comparative study between the performance of signal and image processing algorithms using neural networks. Fourier Transformation and Wavelets analysis are used to process one dimensional and Two dimensional data respectively. The results indicate that two dimensional data obtained from infrared images has significant capability in comparison to one dimensional data for the detection of tool wear for the selected image and signal processing algorithms.

In this paper, based on remote sensing image processing and SWMM simulation, we evaluate the effectiveness of a coupled LIDs system for the watershed scale. The main content and contributions include: 1) the extraction and classification of historical and recent LULC data (in 1979, 1989, 1999, 2009, 2017) of the study area and an analysis of the characteristics of the LULC change; 2) a watershed-based SWMM applied to the simulation of the runoff and an analysis of the runoff change characteristics; 3) the proposal and design of three coupled LIDs scenarios which treat runoff change as evaluation metric to systematically discuss the effectiveness of LIDs in the watershed. The results show that the combination structure and scale can significantly affect the coupled LIDs effectiveness. A system with multiple LIDs is more effective than one with only single LIDs. With the increase of spatial scale, the effectiveness of the coupled LIDs gradually weakens. Our research enriches the application scale of LIDs and SWMM, and can be beneficial to the construction of the “Sponge City”, storm management and urban planning.

A numerical algorithm of principal component analysis (PCA) is proposed and its application in remote sensing image processing is introduced:(1) Multispectral image compression; (2) Multi-spectral image noise cancellation; (3) Information fusion of multi-spectral images and spot panchromatic images. The software experiments verify and evaluate the effectiveness and accuracy of the proposed algorithm.

Chapter 46 - Medical Image Processing Using GPU-Accelerated ITK Image Filters

One of the challenges in remote sensing image processing is subpixel detection where the target size is smaller than the ground sampling distance, therefore, embedded in a single pixel. Under such a circumstance, these targets can be only detected spectrally at the subpixel level, not spatially as ordinarily conducted by classical image processing techniques. This paper investigates a more challenging issue than subpixel detection, which is the estimation of target size at the subpixel level. More specifically, when a subpixel target is detected, we would like to know is the size of this particular target within the pixel?. The proposed approach is to estimate the abundance fraction of a subpixel target present in a pixel, then find what portion it contributes to the pixel that can be used to determine the size of the subpixel target by multiplying the ground sampling distance. In order to make our idea work, the subpixel target abundance fraction must be accurately estimated to truly reflect the portion of a subpixel target occupied within a pixel. So, a fully constrained linear unmixing method is required to reliably estimate the abundance fractions of a subpixel target for its size estimation. In this paper, a recently developed fully constrained least squares linear unmixing is used for this purpose. Experiments are conducted to demonstrate the utility of the proposed method in comparison with an unconstrained linear unmixing method, unconstrained least squares method, two partially constrained least square linear unmixing methods, sum-to-one constrained least squares, and nonnegativity constrained least squares.

Hyperspectral image (HSI) denoising is a fundamental task in remote sensing image processing, which is helpful for HSI subsequent applications, such as unmixing and classification. Thanks to the powerful representation ability of untrained deep neural networks, deep image prior (DIP)-based methods achieve tremendous successes in image processing (e.g., denoising and inpainting). However, DIP-based methods neglect the tensor low-rank prior of the underlying HSI which will be beneficial to capturing the global structure of the underlying HSI. To address this issue, we propose a novel model for HSI denoising, which can simultaneously take respective advantages of the tensor low-rank prior and the deep spatial-spectral prior. The tensor low-rank prior leads to a better global structure and the deep spatial-spectral prior is complementary to preserve better local details. On the one hand, we adopt low-rank tensor ring (TR) decomposition to characterize the tensor low-rank prior and capture the global structure of the underlying HSI. On the other hand, we employ untrained deep neural networks to flexibly represent the deep spatial-spectral prior and capture the local details of the underlying HSI. To solve the proposed model, we develop an efficient alternating minimization algorithm. Experimental results on simulated and real data validate the advantages of the proposed model in HSI denoising. Compared with state-of-the-art HSI denoising methods, the proposed method preserves better local details and the global structure of the underlying HSI.

This article gives a short history and the current state of research in the Institute of Engineering Cybernetics in the field of image processing. A developed integrated geographic information system is briefly described as well as some methods for interpretation of remote sensing images. Applications of these methods are also given.

Applications of Noncausal Gauss–Markov Random Field Models in Image and Video Processing

Reservoir microfacies analysis exploiting microscopic image processing and classification algorithms applied to carbonate and sandstone reservoirs