Paddy fields area has a strategic function as the main food provider for the majority of Indonesia's population. Based on the fixed area of paddy fields from 2013 to 2019, there have been three significant changes in the national paddy fields area. Changes in paddy fields into other land use occurred massively, especially in urban and semi-urban areas. The utilization of remote sensing can be one of the methods in detecting changes in paddy fields area temporally. The purpose of this study was to calculate changes in the area and its distribution of paddy fields in Klaten Regency from 2016 to 2020 using multi-sensor remote sensing data. The method used in this study was detecting paddy fields area temporally in 2016, 2018, and 2020 adopting machine learning approach Random Forest classification using Sentinel-1 and Sentine1-2 imagery. The results of the analysis of paddy fields area with Random Forest method using Sentine1-2 indicated an average test accuracy value of 95% compared to Sentinel-1 analysis with an average test accuracy value of 92%. The results of the analysis of the changes in paddy fields area in Klaten Regency using Sentinel-1 showed that in 2016 there were 30,480 hectares of paddy fields area and in 2020 it decreased to 29,588 hectares. Furthermore, the results of the analysis paddy fields area using Sentine1-2 in 2016 was 31,683 hectares and decreased to 31,196 hectares in 2020. These results demonstrated an under-estimate in the calculation of the results of paddy fields area analyzed by Sentinel-1 and Sentine1-2 compared to the officially fixed area of paddy fields released by ATR/BPN.
ALOS-4 mission is the successor of the currently operational ALOS-2, developed by Japan Aerospace Exploration Agency (JAXA). It will employ digital beamforming capabilities, which will allow for high-resolution wide-swath imaging. The scan-on-receive operation and phase spoiling of the antenna pattern are core features to achieve desired SAR imaging performance. Aiming for mission preparation and calibration studies, a framework to simulate digital beamforming systems for ALOS-4 is presented. A model of the instrument is developed and point target response is used as metrics to validate performance predicted. Simulation algorithms for calibration and evaluation strategies will be discussed in this work.
Aperture synthesis passive imaging system requires more input channels to achieve higher spatial resolution, leading to a square rate growth of connection between channels. To reduce the connection complexity in multi-channel aperture synthesis system, a novel double-layer distributed transmission and processing structure is proposed in this paper. The proposed structure is composed of fully-interconnected inner layers and ring-shaped outer layers, providing the ability for distributed computation and the scalability of the system. A data transmission method for the proposed structure is also presented to ensure that the correlation computation can be calculated in real time and without redundancy. The implementation of the proposed structure is analyzed, such as the uniform distribution of the calculation through the system, the tradeoff between the connection complexity and the transmission bandwidth, etc. The simulation results verify the effectiveness of the system and demonstrate that the proposed structure can tolerate high transmission delay and achieve good transmission utilization.
The main factors of land subsidence are excessive pumping of groundwater for industry, agriculture, snow melting, cooling, pumping of natural gas, and mines’ tunnel excavation. The city of Semarang, the capital of Central Java province, is located on the northern coast of Java island in Indonesia. Land subsidence in this city has enormous potentials for many hazards threatening people and urban infrastructures, such as buildings, roads, or whatever necessary for livelihoods. To accurately measure the land subsidence in Semarang, time series analysis of Small Baseline Subset (SBAS) or Permanent/Persistent Scatterers (PS) and Distributed Scatterers (DS) InSAR analysis is essential. This study shows the recent land subsidence volume in Semarang city with the SBAS technique and compares it to historical InSAR analysis. We also conducted the 2.5-dimensional analysis to convert the satellite line-of-site (LOS) displacement into the horizontal and vertical direction to see if there are differences between ascending and descending data. The SBAS analysis result showed the displacement volume in a range of - 18.2cm subsided and +6.8cm uplifted. The displacement rate was estimated to be substantially constant over time without any apparent seasonal effect. There is no significant difference in the results between ascending and descending datasets. The 2.5-dimensional analysis identified the horizontal and vertical displacements that we corrected any noise and inclined directional values using incidence angle.
Non-line of sight (NLOS) radar imaging is significantly promising in several environmental perceptions, such as urban environment perception and intelligent driving. However, the existing NLOS imaging radar mainly focuses on two-dimensional (2-D) imaging, which may suffer from information loss and geometric distortion of the real three-dimensional (3-D) underlying scene. In this paper, exploiting the scene sparsity and the specular scattering of the reflector, we proposed an NLOS 3-D radar imaging method based on Bayesian compressed sensing sparse reconstruction and mirror projection. In the scheme, the imaging model of NLOS radar with the geometric mirror scattering is first constructed. And then, the sparse Bayesian learning method is used to perform high-resolution imaging of the NLOS objects. In order to verify the technique, we set up a millimeter-wave multiple-input multiple-output (MIMO) array scanning system, and different types of metal balls and irregular targets are tested. The experimental results demonstrate that the presented 3-D NLOS imaging radar can obtain the near-field high-resolution 3-D imaging of the tested targets, and the effectiveness of the sparsity-driven algorithm is verified.
Synthetic aperture radar (SAR) is an advanced remote sensor, which can observe the earth's surface in all weather conditions, widely used in military reconnaissance and disaster rescue. However, due to the coherent summation of the return echoes and the random electromagnetic interference, a SAR image will be significantly affected by the noise, reducing the readability of the image. To deal with this issue, in this paper, we propose an iterative non-local denoising method based on multi-resolution. First, the cascade downsampling is performed to get the multi-resolution sub-images. Then, the 2D discrete cosine transform is applied to each fragment segmented from the sub-images. After that, grouping the similar fragments by using the pHash algorithm. And the basic denoising image can be obtained by performing collaborative filtering and aggregating. Finally, a denoising image can be acquired after iterating processing. Real airborne SAR data is used to validate the effectiveness of the proposed method.
Ionospheric dispersion will introduce quadratic and cubic phase error into spaceborne P-band ultrawideband synthetic aperture radar (SAR) signals, causing range defocusing. We propose an improved contrast optimization (CO) autofocus algorithm for compensating for this error. By analysis of the effect of quadratic and cubic phase error on image contrast we give the rationale for the improved algorithm. The procedures of the improved algorithm are presented. Its superiority to traditional algorithm is demonstrated using simulations based on point target in clutter background and PALSAR2 data.
High-speed digital signal processor (DSP) can solve the time-consuming issue of synthetic aperture radar (SAR) imaging for promoting its application in both military and civilian communities. In this paper, an entire processing chain using two multi-core DSPs for high-squint SAR data imaging, including azimuth filtering and decimation, improved polar format algorithm (PFA), phase gradient autofocus (PGA), image quantization, and return for display, is proposed. The multi-chip DSP architecture can improve the computational efficiency of high-squint and high-resolution SAR imagery generation. Finally, measured SAR imaging results demonstrate the algorithm chain and its parallel architecture that each multicore DSP TMS320C6678 can process the image of 4096*2048 pixels within 1.25 seconds. Moreover, while two chips cooperate in a ping-pong manner, two frames will be obtained after 1.53 seconds, enhancing the processing efficiency by 61 % approximately.
In this paper, we proposed an oil platform detection method in polarimetric SAR images based on level set segmentation and convolutional neural network (CNN). An improved level set segmentation method is used for the extraction of regions of interest (ROIs) at first. The classic CNN model is then used for the identification of oil platforms from the extracted ROIs. In the method, the offshore strong scattering targets are coarsely detected by a thresholding segmentation of the polarimetric entropy and alpha angle parameters. Then, a circle covering the initially detected targets is obtained using a proposed circle covering algorithm. The ROIs are extracted by using level set segmentation in the initialization of the circle. Oil platforms are finally detected by using the improved LeNet-5 model. The experimental results demonstrate the effectiveness of the proposed method using multiple sets of polarimetric SAR data from different sea regions acquired by RADARSAT-2.
Aviation turbulence is one of the dangerous events of a flight. This hazard impacts people and aircraft. One source of information on turbulence events comes from the NTSC report. Then one of the causes of flight turbulence is the presence of convective clouds. Indonesia, which is located in the tropics, is an area that is very active in its convective activities. Descriptive analysis was used in this study. There were three aviation turbulence events in 2016-2020, namely on 4 May 2016, 6 May 2016, and 24 October 2017. All incidents of aviation turbulence occurred wind shear around the incident location based on upper-air observation data. In two turbulence events, there are convective clouds around the scene.
The eruptions of Mount Raung, located in East Java, Indonesia was characterized as explosive and effusive. The majority of eruptions that occurred before 1956 produced ash high into the air, and since 1921 the majority of eruptions have been lava flows to the base of the caldera. The most recent eruption occurred on January 21, 2021 where the ash column was observed to reach 400 - 1000 meters above the peak; on February 8 - 12 it even reached an altitude of 2000 - 2500 meters. Monitoring of lava (hotspot) has been carried out since February 6, 2021, which proceeded to cover almost to the entire caldera base. This eruption is probably a series of lava eruptions for the period November 2014 - August 2015 and July-August 2020. Currently, the main volcanic activity of Raung is located at the base of the elliptical caldera, measuring 1750 x 2250 m at a depth of 400-550 m below the embankment. The deformation activity of Mount Raung was observed by combining interferometric method and GNSS observation. Around 268 unwrapped interferograms since 2015 until January 2021 were provided by LiCSAR and processed by LiCSBAS method. The deformations are clearly shown on crater with deflation trend, whereas inflation trend is detected on the caldera. In short period, i.e. June to January 2021, inflation and deflation are detected on the crater. For terrestrial method, GNSS data from 3 stations that observed since December 2018 using the Leica SR30 receiver and processed by GIPSYX software. The data showed indication of deflation in the body of Mount Raung from December 2018 to mid-January 2021. Moreover, between mid-January to late February 2021 there was an indication of inflation in the peak areas. The analysis using shorter period should be able to observe the dynamics changes in the deformation of Mount Raung, but the long-term ones can determine the pressure accumulation process at the depth and the possibility of future eruption.
Deep learning has been widely used in various areas, such as detecting materials, or estimating natural disasters. Especially, generative adversarial network (GAN), which is one of the deep learning models, is enhanced to CycleGAN for generation and discrimination of images even with unpaired datasets. In this paper, we design a model to generate real-like fake flood models, and we confirm that we distinguish between real and fake images by mixing them with real images. Based on this metric, a deep learning model is designed, and a dataset is generated using CycleGAN. We further perform data augmentation to assist in the dataset generation process. The program used to design the model is Python, which uses data from Sentinel-l. Input data is a collection of data from floods during 2019 in West Africa, Southeast Africa, Middle East Asia, and Australia. To determine the accuracy of the generated data, we compare the image using several indicators. The used indicators judge the accuracy and similarity of images such as SSIM and MSE, and PSNR. SSIM, MSE, and PSNR averaged 0.7192, 2014.0066, and 15.5745, respectively. Comparing images with these indicators, we confirm that the actual flood image and the generated flood image are similar. And using generated images, we use different deep learning model, to confirm how similar the real flood image is to the flood image produced in this paper.
Recently, radar-based near-rield three-dimensional (3D) imaging technology has attracted increasing attention owing to its unique characteristics. Various imaging methods have been developed for fast computation and accurate imaging. One of the most accurate and fast imaging algorithms is the fast factorized back-projection (FFBP) algorithm. This study presents a successive 3D-FFBP using a one-dimensional (ID) bistatic antenna array for 3D imaging for concealed weapon detection, which is derived as an extension of a well-known two-dimensional FFBP. The crucial parameters for these FFBPs to run efriciently are the prederined grid sizes of the imaging area. The optimal azimuth/elevation grid sizes for the proposed successive 3D-FFBP are formulated using certain approximations. The validity of the proposed successive 3D-FFBP using the optimal grid sizes is confirmed with a numerical simulation where a bistatic millimeter-wave radar array is used.
The combination of multi-input multi-output (MIMO) millimeter-wave radar sensor and synthetic aperture radar (SAR) makes the implementation of 3-D near-field millimeter-wave imaging cheaper and more practical. In this paper, a 3-D SAR imaging system based on the millimeter-wave sensor is constructed, through experiment contrast and analyzed based on the theory of matched filtering (MF) imaging algorithm and based on the theory of compressed sensing (CS) in the nearfield millimeter-wave imaging algorithm of 3-D imaging results of SAR system, the pros and cons of two kinds of algorithms are verified in theory, the imaging algorithm based on CS theory imaging effect is better than conventional MF imaging algorithm of imaging effect is concluded, along with various real imaging results.
This paper presents the technological and methodological aspects of the multifrequency MetaSensing systems for DEM generation. Three selected recent airborne campaigns, each at a different operating bands (X, C & P), specifically designed for InSAR applications are reported. The paper first provides a general overview of the installation of the different airborne SAR systems followed by a description of the processing and calibration steps. We introduce the use of the zero-meter-baseline single-pass interferogram for the phase calibration. The paper presents the results over hilly areas and forested terrains and discusses the resolution and the accuracy of the processed data.
We are conducting an experiment for Circularly Polarized Synthetic Aperture Radar (CP-SAR) using Unmanned Aerial Vehicle (UAV). Raw image data obtained by radar is processed by FPGA on UAV. The Range Doppler Algorithm (RDA) is used for our image processing. Currently, our image processing system uses the KC705 evaluation board with Kintex-7 FPGA, and it communicates with CP-SAR controller unit to perform image processing. This KC705 evaluation board is a multipurpose board and has parts that are not used in image processing. Therefore, we select the parts required for image processing and design a board for SAR image processing that is smaller and lighter than the KC705.
Remote sensing is useful to extract damages due to natural disasters. In this study, Synthetic Aperture Radar (SAR) images acquired from PALSAR-2 sensor onboard ALOS-2 satellite were used to observe damage situations due to the 13 February 2021 Off-Fukushima, Japan, earthquake. Landslides and rockfalls were reported at the Joban Expressway, a motor racing circuit, and seaside cliffs. Change detection using pre- and post-event PALSAR-2 intensity images was carried out and the results were compared with airborne optical images and field survey data. The landslide at the expressway and the motor circuit were recognized from the SAR intensity data, but rockfalls and other small-scale damages could not be identified due to the limitation of spatial resolution.
Unsupervised clustering is essential when in-hand information about crop phenological stages are not available. In addition, this technique assists in discriminating crop growth stages concerning the changes in the scattering mechanisms. Newly proposed parameter $\theta_{FP}$ for full polarimetric Synthetic Aperture Radar (SAR) data is shown to be enhancing the target characterization capability over the existing parameter such as $\overline{\alpha}$. Following this, the existing clustering technique by combinedly utilizing the Barakat degree of polarization and elements of the Kennaugh matrix along with scattering entropy H FP is effective over land cover classes. The clustering scheme is consisting of 12 zones where zones (Zl, Z2, Z3), (Z10, Zll, Z12), and (Z4, Z5, Z6, Z7, Z8, Z9) represent even, odd and multiple scattering types, respectively. In this study, we use the RADARSAT-2 FP SAR data from Jun to Nov 2019 acquired over the Indian test site of Vijayawada under the Joint Experiment for Crop Assessment and Monitoring (JECAM) initiative. The changes and shifts of the clusters are exciting according to the phenological development of rice fields. At the initial stage, dense clusters are evident in Z10, Zll, and Z12, while, during the advanced tillering stage, more data points are found in the Z3 zone. Following this, clusters in zone Z5, Z6, Z8 and Z9 are evident during the maturity stage.
The Persistent Scatterer Interferometry (PSI) analysis accurately estimates the change over time of only the PS points on the ground surfaces using the time variation of the phase change of pixels with a small amount of noise called Persistent Scatter (PS).The riprap on the external surface of a rockfill dam is well reflected by radar, and since many PS points can be obtained over the entire external surface of the dam, the amount of external deformation can be grasped in a planar manner. In this paper, the PSI analysis was conducted for six large-scale rockfill dams of TEPCO Renewable Power using the SAR (Synthetic Aperture Radar) data from Sentinel-1 C-band, with 22 to 32 scenes of southbound and northbound trajectories in FY2020, respectively. The external displacement of the rockfill dams was measured using the SAR data. As a result of comparing the results of PSI analysis with the external deformations measured by surveying, it is concluded that the results of PSI analysis could be used to evaluate the deformation of the rockfill dam for maintenance management.
This paper mainly studies despeckling method for Synthetic Aperture Radar (SAR) image. The denoising method based on deep learning usually needs a lot of samples and noiseless reference images to realize data regression. However, speckle noise is very different from random additive noise such as white Gaussian noise in that its distribution is affected by the background target and it is hard to obtain completely noiseless reference images. Therefore, we proposed a local learning method and constructed a deep learning model to overcome this challenge, which named Speckle-Denoise-Net (SD-Net). In proposed method, the speckle noise can be effectively suppressed without the need for noiseless reference images. It can not only solve the problem of lacking training samples and having excessive image size, but also take advantage of unsupervised learning for inherent noise to make a good balance between smoothing and structure-preserving compared With other methods. At present, the test results of this method (Including denoising effect, original information retention, and algorithm speed) have exceeded most of the existing traditional denoising algorithms and deep learning methods, which allow it to bring new inspirations to solve the problem of speckle noise.
