kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
A novel method for locating the local terrestrial laser scans in a global aerial point cloud
1
26
FA
Amin
Baghani
K.N. Toosi University of Technology
Mohammad Javad
Valadan Zoej
K.N. Toosi University of Technology
Mehdi
Mokhtarzade
K.N. Toosi University of Technology
In addition to the heterogeneity of aerial and terrestrial views, the small scale terrestrial point clouds are hardly comparable with large scale and overhead aerial point clouds. A hierarchical method is proposed for automatic locating of terrestrial scans in aerial point cloud. The proposed method begins with detecting the candidate positions for the deployment of the terrestrial laser scanner in the aerial point cloud. After that, by simulating the performance of the laser scanner, the visible portion of the aerial point cloud is detected and it is extracted as the candidate deployment aerial point cloud. As a result, the problem of scan locating is converted to a corresponding one between several local terrestrial point clouds and several local aerial point clouds. In order to increase the comparability of these two datasets in the corresponding process, the main geometric structures of each point cloud are extracted using four predesigned geometric feature indexes, and they are organized in the form of four feature-maps of each point cloud. The feature-maps generated for each point cloud are described by the rotation invariant Fourier-HOG descriptor. Afterward, the corresponding problem is structured in the form of a k-nn classification among the classes established for these descriptors. Finally, the location of each terrestrial scan is obtained based on the classification results. The evaluation results of the proposed method on an urban dataset, showed an average accuracy of about 5 meters for locating the terrestrial scans in aerial point cloud. The obtained accuracies seem to be sufficient to enter the process of registering the terrestrial scans to the aerial point cloud.
kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
Investigating the effect of the removing well-known periods from TEC time series data in identifying seismo-ionospheric anomaly
27
43
FA
Zahra
Sadeghi
K. N. Toosi University of Technology
masoud
Mashhadi-Hossainali
K. N. Toosi University of Technology
Iran is situated on one of the two largest seismic belt of the world which is called Alpa. The occurrence of damaging earthquakes all around the country causes a lot of pecuniary losses and casualties. The prediction of this phenomenon can have a profound impact on reducing the risks which are posed by it. Among the earthquake predictors, which are of interest to many scholars today, we can mention the occurrence of abnormal changes in the ionospheric parameters. The studied ionospheric parameter in this paper is the Total Electron density Content (TEC) obtained from the Global Inosphere Map (GIM). The time series of the TEC data comprises well-known periods whose causes are not seismic. It seems that, by eliminating the effect of these factors, as much as possible, the identification of the ionospheric-seismic anomaly could be done much more accurately. In this study, changes in TEC before the occurrence of two earthquakes in Iran have been investigated by using the mean method two times, once before the elimination of the periods and once afterwards. In fact, after the removal of known frequencies for Khuzestan and Semnan earthquakes in the area of identification of seismo-ionospheric anamolysis, the improvement of more than 50% and 10% is seen in most of the places respectively.
kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
A multi-scale convolutional neural network for automatic cloud and cloud shadow detection from Gaofen-1 images
45
68
FA
Mehdi
Khoshboresh-Masouleh
School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran
Reza
Shah-Hosseini
School of Surveying and Geospatial Engineering, College of Engineering, University of Tehran
The reconstruction of the information contaminated by cloud and cloud shadow is an important step in pre-processing of high-resolution satellite images. The cloud and cloud shadow automatic segmentation could be the first step in the process of reconstructing the information contaminated by cloud and cloud shadow. This stage is a remarkable challenge due to the relatively inefficient performance of the current methods in complex scenes in multispectral high-resolution satellite images. In recent years, using deep convolutional neural networks has largely improved the performance of cloud and cloud shadow segmentation. Increasing the generalization capability of cloud and cloud shadow segmentation is one of the problems of deep convolutional neural networks. In this paper, we focus on tackling the poor generalization performance of automatic cloud and cloud shadow segmentation in Gaofen-1 (GF-1) images. In this regard, we propose a deep learning multi-scale method, founded on multi-dimension filters, for accurate segmentation of cloud/cloud shadow in single date GF-1 images which is based on a new multi-scale deep residual-convolutional neural network called MultiCloud-Net. The cloud/cloud shadow masks are extracted based on a new loss function to generate the final cloud/cloud shadow masks. The MultiCloud-Net was implemented in the Google Colab and was validated using 12 globally distributed GF-1 images. The quantitative assessments of test images show that the average F1 score, the average Jaccard Similarity Index (JSI), and the Kappa coefficient for cloud (cloud shadow) segmentation are about 97 (95.5), 96 (94.5), and 0.98, respectively. The experimental results using the GF-1 images demonstrate a more reasonable accuracy and efficient computational cost achievement of the proposed method compared to the automatic cloud/cloud shadow segmentation performance of two advanced deep learning and statistical methods.
kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
Generating an Indoor space routing graph using semantic-geometric method
69
80
FA
Javad
Sadidi
Kharazmi University
Zahra
Joudaki
Kharazmi University
Hani
Rezayan
Kharazmi University
The development of indoor Location-Based Services faces various challenges that one of which is the method of generating indoor routing graph. Due to the weaknesses of purely geometric methods for generating indoor routing graphs, a semantic-geometric method is proposed to cover the existing gaps in combining the semantic and geometric methods in this study. The proposed method uses the CityGML data model, which is actually a semantic modeling of building space. The output of the method is also presented with several test scenarios, and their results. Using semantic information and semantic graphs is in fact a good strategy for purely geometric methods, and according to the results, the proposed semantic-geometric method for producing indoor routing graphs seems to be efficient.
kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
Study of soil moisture change effects on L-band DInSAR phase
81
101
FA
Sadegh
Ranjbar
University of Tehran
Mehdi
Akhoondzadeh
University of Tehran
The Differential Synthetic Aperture Radar Interferometry (DInSAR) technique is recognized as a potential remote sensing tool for detecting ground surface displacements with less than a centimetre accuracy. The surface soil moisture changes ( ) during the time between the two images as an effective parameter on interferometry phase ), leads to incorrect calculation of ground movement . In this research, the amount and the way that affects on wheat, rapeseed, weed, pea and idle land fields have been investigated empirically using a regression model. To do this investigation, airborne data UAVSAR (L-band) along with ground-based data in the CanEx-SM10 campaign in 2010 were used. According to the scattergraphs between and , and observing a direct and approximately linear relationship between them, some hypotheses were taken into consideration in order to use a regression modeling . Comparing the estimated π using the calibrated regression model and calculated π from the interferometry technique shows that the model provided the best results for the bare field in VV and HH polarizations (RMSE) of 0.3 to 0.6 rad and R2 of 69% to 72%. In general, the results of the regression model showed that without other factors’ effects on , this parameter can be modelled based on a regression function in bare fields. The model also provided acceptable results in vegetated fields (RMS of 0.6 to 0.99 rad and R2 of 40% to 55% depending on the different vegetation types and different polarizations). Comparing polarizations, fluctuations in co-polarizations (HH and VV) showed a higher correlation with . Consequently, φ is directly affected by , and significant changes in brings about a considerable error in displacement estimation.
kntu
Engineering Journal of Geospatial Information Technology
2008-9635
8
4
2021
3
1
Bridge Modeling using Segmentation of Point Cloud Captured from Photogrammetric UAV
103
127
FA
Mansour
Mehranfar
University of Tehran
Hossein
Arefi
University of Tehran
Fatemeh
Alidoost
University of Applied Science (HFT), Stuttgart
In recent years, great efforts have been made to generate 3D models of urban structures in photogrammetry and remote sensing. 3D reconstruction of the bridge, as one of the most important urban structures in transportation systems, has been neglected because of its geometric and structural complexity. Due to the UAV technology development in spatial data acquisition, in this study, the point clouds generated from UAV-based images are used for 3D modeling of the four main elements of a bridge structure, including the railing, body, base and abutment elements. For this, a knowledge-based algorithm is proposed to provide 3D models of different types of bridge structures in GIS-based data format using the knowledge in the shape, structure and geometric relationships between the bridge’s elements. First, the fuzzy c-means clustering method including height and spectral values as well as point-based features such as the 3D density, normal vectors ββand planarity is used to segment the point cloud. Next, a projection-based reconstruction technique, which is developed based on the geometric and structural features of each bridge element, is proposed to generate a 3D model for that element. The proposed reconstruction workflow includes the projection of point clouds to a 2D space, fitting the primitive geometric models to 2D points, locating the primitive coordinates of the models in the 2D space, and then developing 2D models into 3D space. To evaluate the proposed method, the dimensions of the structural elements in the bridge design plans are compared with the dimensions of the elements in the generated 3D model. Despite the many challenges in modeling steps, the results of this study indicate a high accuracy and ability for the proposed algorithm in 3D modeling of bridges with different geometry and designs, with a mean error and a standard deviation of about 3 cm and 1cm, respectively.