The completeness of road extraction is very important for application. However, existing deep learning (DP) methods often generate fragmented results. prime reason that DP-based use square kernel convolution, which challenging to learn long range contextual relationships roads. produce fractures in the local interference area. Besides, quality results will be subjected resolution remote sensing (RS) image. Generally, an algorithm worse fragmentation when used data differs from training set....

Roads are an important recognition target in synthetic aperture radar (SAR) image interpretation. Although a considerable number of high-quality SAR images now available, the method road extraction is lagging. To extract network with low missed and false rates, this paper proposed approach which includes line detection, segmentation, optimization. First, linear feature response direction map obtained from intensity using multiplicative Duda operation. Then, backscattering coefficient...

Low-grade roads have complex features such as geometry, reflection spectrum, and spatial topology in remotely sensing optical images due to the different materials of those also because they are easily obscured by vegetation or buildings, which leads low accuracy low-grade road extraction from remote images. To address this problem, paper proposes a novel deep learning network referred SDG-DenseNet well fusion method Synthetic Aperture Radar (SAR) data on decision level extract roads. On one...

In order to make the best use of available data in remote sensing database, this paper focuses on an important topic using complementary information multi-source data, that is, road network extraction based fusion technology with synthetic aperture radar (SAR) and optical images. Starting line segments achieved from segmentation maps, a decision-level method which mainly includes two stages is proposed paper. The first stage fusing geometric overlapping rules. second stage, approach takes...

In this paper, a method for road detection based on Duda and path operators has been presented. The roads are represented as slender dark regions with constant width reflectance in the high-resolution SAR images. (path openings closings) were performed morphological filters retaining linear structures. However, not sensitive to of feature. Focused limitation method, preprocessing procedure using was introduced before adopting profiles operators. When modified applied RADARSAT-2 datasets...

Interferometric synthetic aperture radar (InSAR) has become the primary means to obtain digital elevation models(DEM) on earth's surface, including several key steps such as removal of flatten effect, filter processing, and phase unwrapping. Due atmospheric interference, etc., distribution azimuth range quality SAR images is uneven. Thus, this paper proposes a method based weighted graph guide unwrapping, which considers difference contribution weight spatial noise graph. The weighting used...

In order to speed up the update of existing road maps, it is crucial develop a more efficient extraction method from remote sensing images. recent years, deep learning techniques have been widely used for applications. Among current CNN-based networks extraction, few works study shape convolution kernel, and contextual features dependency relationship not fully utilized. view these problems, an improved DLinkNet proposed in this paper. Firstly, convolutional layer which fuses information...

Road centerline extraction is the foundation for integrating segmented road map from a remote sensing image into geographic information system (GIS) database. Considering that existing approaches tend to have decline in performance and junction when structures are irregular, this paper proposes novel method which models network as sequence of connected spline curves. Based on motivation, ratio cross operators firstly proposed detect direction width features roads. Then, pixels divided...

In order to study the coherence characteristics of low-backscattering objects such as roads and rivers, we introduce a estimation approach based on clustering method. When is applied multi-temporal high-resolution TerraSAR-X images in urban areas, results show that higher than rivers shadows. This indicates features can be used distinguish water bodies. Further, this paper proposes road detection method synthetic aperture radar (SAR) path operators support vector machine (SVM), which...

Abstract. In this paper, a novel framework for spatial data processing is proposed, which apply to National Geographic Conditions Monitoring project of China. It includes 4 layers: storage, RDDs, operations, and query language. The storage layer uses HDFS store large size vector/raster in the distributed cluster. RDDs are abstract logical dataset types, can be transferred spark cluster conduct transformations actions. operations series on such as range query, k nearest neighbor join....

The inclination direction monitor of the power transmission tower can issue an early-warning for risk collapse, as well collapse direction. In this paper, we proposed a new method to retrieve using geocoding. First, backscattering coefficient was utilized separate from background, and figure up those coefficients along extract main body tower. Second, whole image traversed by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$3^{\ast}3$</tex>...

Using deep learning to extract roads from satellite images is one of the most popular methods. However, existing encoder-decoder-based networks usually produce fragmented roads, due complex spatial and color characteristics road. In this paper, motivated by road multi-scale information, we proposed a multi-direction feature fusion network (MSMDFF-Net) reduce fragmentation extraction results. The method mainly consists three processes: 1) initial stage, image details different directions were...

Abstract. In this paper, a novel Apache Spark-based framework for spatial data processing is proposed, which includes 4 layers: storage, RDDs, operations, and query language. The storage layer uses HDFS to store large size of vector/raster in the distributed cluster. RDDs are abstract logical dataset types, can be transferred spark cluster conduct transformations actions. operations series on such as range query, k nearest neighbour join. language user-friendly interface provide people not...

Due to different imaging mechanisms, the registration of optical and Synthetic Aperture Radar (SAR) image is a very challenging task. Many SAR methods have been proposed. But most them are for low-to-medium resolution images, less high-resolution images. Therefore, this paper proposes method using local self-similar descriptor based on edge feature. Firstly, Gauss-Gamma bi-windows algorithm used extract intensity maps images respectively. Its function eliminate non-linear gray-scale...

This paper introduces a method of road extraction using dual-temporal high-resolution synthetic aperture radar (SAR) images. Firstly, multiplicative Duda operators are applied to detect line features. Then, coherence and backscattering coefficient combined distinguish from river shadow the variation is used remove heterogeneous areas. Next, segmented via path opening thresholding. At last, novel thinning gap connection approach proposed gain thinned more complete map. The experiments were...

This paper investigates the problem of building extraction from very high resolution (VHR) satellite imagery. Deep learning methods are deemed as emerging trends for solving this due to their increasingly prominent effects. However, most state-of-the-art deep learning-based segmentation produce pixel-level masks rather than accurate polygon-level results required in real-world applications. introduces a Harris function based active contour network (HACNet) that extracts directly Our proposed...

This paper reports on an experiment conducted at the wind-wave tank in UESTC microwave chamber to characterize C- and X-band radar return from water surfaces covering oil films when observed different incidences. The measurements of Normalized Radar Cross Section (NRCS) with bi-objective calibration technique were carried out for full polarization various wind speeds. Comparisons are performed clean sea diverse spills. From this data set we validate Two-Scale Model (TSM) as a calculating...

This paper proposes a semi-empirical microwave model of methane emissions (CH4) based on the biogeochemical processes from rice paddy. By exploiting mechanism production, oxidation and emission, combined is developed to predict Simultaneously, main influencing factors, which concluded soil, underlying water body, vegetation, climate management, are analyzed in present model. During whole growth season, multi-polarization backscattering coefficients measured by ground-based radar...