Extraction, processing, production and display of geographic data
Fatemeh Ahmadi; Yasser Ebrahimian Ghajari; Abbas Kiani
Abstract
Extended AbstractIntroductionUrbanization can be defined as social and economic development and thus, urban planners need timely information for service provision and management in urban areas. Due to the difficulties of traditional methods, automatic and semi-automatic methods have gained special importance. ...
Read More
Extended AbstractIntroductionUrbanization can be defined as social and economic development and thus, urban planners need timely information for service provision and management in urban areas. Due to the difficulties of traditional methods, automatic and semi-automatic methods have gained special importance. Therefore, remote sensing data and image classification techniques have been used to help identify different types of land use. Nighttime light emission data can help researchers effectively identify human activities and urban areas. These satellite images are collected from the surface of the earth at night and can clearly separate nighttime light emission of urban areas from the surrounding dark areas. Thus, it can be concluded that various types of data with different nature and capabilities (spectral, nighttime light, etc.) are available for any specific area each of which has its own advantages and limitations. As a result, using a combination of these data types will increase accuracy and reduce uncertainty. Algorithms and scientific methods enabling this combination are thus of great importance. The present study applies a combination of nighttime light emission and daytime multispectral images to produce automatic and high-quality optimal training samples and locate built-up areas.Material & MethodsTwo study areas (in Babol and Kerman) with two different climates have been investigated in the present study. Also, DMSP and VIIRS nighttime light emission images and Landsat 5, 7 and 8 images collected during the statistical period have been used.Research MethodsThe present study has proposed an approach consisting of four main phases of pre-processing, feature extraction and production of initial training samples, selection of optimal training samples and finally classification and evaluation. Nighttime light emission images were corrected and primary samples including two classes of built and unbuilt areas were produced using the limit of automatic thresholds. Nighttime light emission is generally related with human activities, and thus, built-up areas usually have a higher nighttime light emission value compared with unbuilt areas which have a lower or zero value. Due to the saturation and blooming problems occurring in DMSP images and the relatively low spatial resolution of nighttime light emission data, training samples extracted from built areas using these data still include unbuilt areas such as water bodies and vegetation cover. Therefore, an index has been developed using features extracted from nighttime light emission and Landsat images. Considering the inverse relationship between various features of urban and rural areas (vegetation cover and soil) in LST images obtained from the thermal band of Landsat images and the NDVI vegetation index obtained from Landsat and features of urban areas in nighttime light emission image, an index was provided which maintains the main characteristics of urban areas in nighttime light emission images while minimizing saturation and blooming. Finally, time series of classified images was investigated and urban expansion was analyzed.Result & DiscussionFollowing nighttime light emission data correction, an upward trend was observed for the values of pixels collected from each city which verifies the pre-processing stage. Then, an appropriate automatic threshold limit was selected in accordance with the features of each nighttime image and applied to produce the initial training samples. Nighttime light emission images were corrected using the introduced index to minimize saturation and blooming in urban and suburban areas. Training samples thus optimized were used for final classification. Due to the low quality of initial training samples, classified pixels obtained from urban areas did not confirm to reality. Thus, classification faded in Kerman city in some years and practically no classification was performed which shows the low quality of initial training samples. Due to the low spatial resolution of nighttime light emission images, the size of samples collected from built-up areas was falsely detected to be large, and thus, there were definitely samples related to vegetation, soil, and etc. in the specified range. In the next step, classification was performed using optimal training samples in which built-up regions were modified. In this way, results got closer to the reference data and reality. In fact, using a combination of nighttime light emission and Landsat data can overcome the limitations of both methods.Conclusion Selection of training samples is considered to be the main and fundamental challenge of classification. With a valid training sample, classification is precisely performed. Since, traditional and manual methods of obtaining training samples are costly and time-consuming, automatic and semi-automatic methods have become specifically important. Therefore, the present study has classified and extracted built-up areas using satellite images. The initial training samples can be obtained automatically from nighttime light emission images, however high saturation and blooming of these images have reduced their quality. To solve this problem, a nighttime light index has been developed based on the relationship between the characteristics of urban areas in optical images and nighttime light emission images which has minimized related problems in both study areas with two different climates to a great extent. This shows the flexibility and effectiveness of the proposed method. High-quality training samples thus obtained were highly effective in the final classification phase. Investigating urban expansion time series has shown that urban growth and expansion have generally occurred around the city.
Extraction, processing, production and display of geographic data
Maryam Kouhani; Abbas Kiani; Yasser Ebrahimian Ghajari
Abstract
Extended AbstractIntroductionVegetation has always been affected by various environmental and human factors that have directly or indirectly affected the conditions and performance of the environment over time. Consequently, monitoring and investigating the vegetation cover in the northern regions of ...
Read More
Extended AbstractIntroductionVegetation has always been affected by various environmental and human factors that have directly or indirectly affected the conditions and performance of the environment over time. Consequently, monitoring and investigating the vegetation cover in the northern regions of Iran is also highly considered important. Research suggests that the destruction and change of vegetation cover and forests are among the most important factors influencing natural hazards such as floods, erosion, and earthquakes. In addition to processing and presenting well-known spatial data, remote sensing can also be used to improve human understanding of annual changes in vegetation cover, from a local to a global scale. In this regard, the anomaly evaluation criterion with high differentiation can separate and display anomalous areas in order to recognize the change process and reveal the areas with anomalies over time. Thus, medium-resolution images, vegetation indices, and anomaly criteria can be used to evaluate long-term vegetation changes. Therefore, a positive step in reducing the environmental effects of a region can be made by locating the urban areas that have experienced changes over time and making decisions related to future planning.Material and methodsThis study utilized a time series of Landsat 5, 7, and 8 images downloaded from the Google Earth engine. To get the best representation of the vegetation in this study, spring and summer were chosen because vegetation at this time is at its greenest. The main focus of this study was on the evaluation of vegetation changes over time quantitatively and qualitatively, using remote sensing data from Google Earth Engine to prepare a map of vegetation changes over time. The general process of implementing this research can be summarized in 7 phases. The first phase involves taking Landsat images and preparing statistical meteorological data. In the second phase, the time series images were collected according to the specific period and in the third phase, the obtained images were corrected and pre-processed. As a next step, the EVI index is extracted from all Landsat images, and then to determine the anomaly of changes, a series of statistical analyses, including the mean and standard deviation, are applied. The next step involves generating the map of anomalous time series changes and extracting the map of vegetation changes to improve understanding. The end of the process also includes evaluating the results obtained from this research. Results and DiscussionSince vegetation and drought changes are non-uniform depending on location and distance from the sea and humid areas, and vegetation is destroyed to build villas, residential areas, commercial areas, and towns, several study areas were divided into smaller pieces. Then each area was analyzed and evaluated separately for its changes. It has been observed in the first and third study areas that vegetation has generally been on the rise in the past 36 years, although sometimes there have been anomalies and fluctuations in EVI value. It was significant to see the reduced vegetation in 2008 in both regions. For example, 262.5 mm of precipitation in the first region fell this year, indicating a rain shortage. The results obtained from the second region, considered one of the coastal regions, indicate that the anomaly graph in the region during the period had a downward slope in the direction of decreasing vegetation, and EVI values reached 0.14 in 2005 and 0.09 in 2013. The 4th and 5th regions have shown a lot of fluctuations in anomalous changes and EVI values, although the trend has generally been downward. Results obtained in the 4th region show that vegetation cover peaked in 2004 and 2011. Rainfall in the 5th region, a highland region, in 2008 was deficient, with 259.8 mm reported by the meteorological station. The anomaly value in this year was -1.96. According to the Department of Meteorology in Mazandaran province, most droughts that have affected the underground water in the province have taken place in coastal and plain areas in the province's east and center, and in western cities, they have mostly affected mountainous areas.ConclusionThirty-six years of EVI time series images obtained from Landsat images were utilized in this study to investigate the changes and identify anomalies. In order to conduct a more detailed investigation, the study area was divided into several different regions, and each region was evaluated separately. The results obtained with existing meteorological statistical data were analyzed because vegetation can be affected by climatic and environmental conditions such as weather conditions. According to the results from study area )4(, vegetation cover has consistently decreased over the last three decades due to various factors like tree cutting, landslides, or land use changes. As shown in the map showing the obtained changes, there appears to be an increase in the value of the vegetation index in some northern areas of Chalus city until around 2002, indicating an improvement in greenness. While In some areas close to the Caspian Sea and the coastline, because of the construction of villas and commercial areas, there has been a loss of vegetation, such as in area (2) based on the changed map, a major part of the vegetation in that area has been destroyed because of the establishment of a settlement and construction of a road. As a result of comparing the evaluation of two anomaly approaches, it has also been concluded that both modes show almost the same trend of changes, but the graphs in "Anomaly compared to the overall average" mode compared to "Anomaly compared to the average of each set" display the change process better.
Mina Mohammadi; Abbas Kiani
Abstract
Extended Abstract
Introduction
DEMs (digital elevation models) are of critical importance in different areas such as land use planning, infrastructural project management, soil science, hydrology and flow direction studies. Across greater spatial scales, their usage is the key for contouring topographic ...
Read More
Extended Abstract
Introduction
DEMs (digital elevation models) are of critical importance in different areas such as land use planning, infrastructural project management, soil science, hydrology and flow direction studies. Across greater spatial scales, their usage is the key for contouring topographic and relief maps. A DEM represents the bare surface, eliminating all natural and artificial features, while the digital surface model (DSM) captures both natural and artificial features of the environment. DSM is of significant interest for applications such as environmental planning, map updating, or building detection. Ground filtering is the removal of the points belonging to the above-ground objects in order to retrieve ground points to be used in generating DEM. DEM can be effectively obtained from LIDAR or digital photogrammetry. Lidar point clouds have great success in representing the objects they belong to; but since the Lidar data acquisition is still a costly process, using point clouds generated by the photogrammetric process to produce DSM is a reasonable alternative. Since DSM represents the information of surface of the land objects and is also affected by ground slope, it cannot be useful lonely for interpreting the data; therefore, to make optimal use of it, a distinction is required between the land and non-land pixels. On this basis, due to the large volume of the high-resolution images and with regard to complex urban structure, a fast yet simple and accurate method is desirable.
Material & Methods
Based on the filtering algorithms, the provided digital surface model is classified into ground and off-ground pixels. For all the off-ground pixels, the closest ground point is assumed to be the relevant low point, thus, through the height difference of the off-ground point with the assigned ground point, the so-called normalized height is computed. However, most of the filtering algorithms are mainly developed to filter Lidar data and will require the adjustment of a number of complex parameters to achieve high accuracy. At the same time, the processing time, degree of effectiveness in different scenes, and degree of automation of these methods are also important. Scene details and topographical complexity, for example in urban areas, make the filtering process even more challenging. For optimal results, users should try to adjust various parameters until they find the desired filtering result, which is a time-consuming and costly process. Due to the lack of a comprehensive study on the efficiency, automation, and computational complexity of different filtering methods on the points cloud obtained from photogrammetry, in this study, different and most widely used algorithms in this field of study were compared with each other. The studied methods were analyzed in terms of class filtering quality, processing time (execution time), scene complexity, and number of algorithm parameters (indicating the degree of user involvement in data processing to determine the amount of automation). Results of this analysis can be useful in order to better understanding the performance of filtering methods on the DSM obtained from high resolution images (dense point clouds from aerial and UAV images). In addition, it can be suitable for different users according to the parameters of time, hardware, scene type, and output accuracy.
Result & Discussion
Ground filtering is essential for DEM generation. In this paper, for ground filtering, at first, a suitable algorithm was selected and, after setting the initial parameters, they were applied to the point clouds. Comparing the obtained results, it can be seen that in the building class with sloping roofs, Morph and ATIN methods performed better, but in buildings with flat roofs, only Morph method had good accuracy. In the mono-tree class, the Morph and ATIN methods in Metashape software were able to perform the separation well, and in the tree row class, both methods performed well. The ATIN method in Metashape software was able to differentiate the road class more accurately than other methods. It also performed well in the river class. Therefore, according to the results of this study, if the goal is to identify high tolls in urban areas, due to the lower computational cost of the Morph method than the ATIN method, the Morph method is recommended. But if the goal is to produce good quality DTM, the ATIN method will be the priority.
Conclusion
In this research, ATIN, ETEW, MLS, MORPH1D, and MORPH2D algorithms for land extraction were evaluated. Thus, first the algorithms were examined on the test data and, then, the results were analyzed with the ground true images. In this study, five filtering methods were examined and compared on three images of urban areas, which included various natural and human-made features, including streets, trees, and buildings. The data were related to the digital aerial imagery taken by Intergraph/ZI DMC sensor in Vaihingen city, Germany. DSM data sets were defined on the grid with the ground resolution of 9°cm. Comparing the results of all the three data sets, it can be seen that the difference in accuracy between the one- and two-dimensional morphology algorithms was very small and they had similar performance. In terms of processing time, the ATIN method had longer execution time than other methods and the ETEW method had shorter execution time than other algorithms. Also, the number of algorithm parameters indicated the degree of user participation in data processing. Therefore, due to the point that the ETEW algorithm had fewer parameters, its degree of automation was higher than other algorithms. Comparing and reviewing the results obtained from the test data demonstrated that MLS and ETEW algorithms had the lowest efficiency in the urban area. On the other hand, in features such as buildings with sloping roofs, single trees, and tree rows, two ATIN and Morph algorithms provided favorable results. According to the obtained results, the suitable algorithm was Morph algorithm for flat-roofed buildings and ATIN algorithm for road and parking. In general, it is recommended to use the Morph algorithm for urban and small areas due to time savings and less effective parameters.