Semi-Automatic Classification Plugin Research Articles

Testing NDVI, tree cover density and land cover type as fuel indicators in the wildfire spread capacity index (WSCI): case of Montenegro

This paper presents an updated version of our previous GIS-based method developed for indexing the forest surfaces by their wildfire ignition probability (WIPI) and wildfire spreading capacity (WSCI). The previous study relied on a multi-criteria approach including a variety of factors of social, hydro-meteorological, and geo-physical character of the context. However, this study is challenging the drawbacks of the previous work, by introducing three new criteria regarding the vegetation properties in the area. Normalized Difference Vegetation Index (NDVI), Tree Cover Density (TCD), and land cover type are launched as indicators of fuel properties of the forest being indexed. The materials and software utilized here belongs to different open sources. CORINE Land Cover (CLC), Open Street Map (OSM), TCD via Copernicus high resolution data, and multispectral satellite images via Landsat 8 (Semi-Automatic Classification Plugin- SCP) are utilized as raw materials in a workflow in QGIS software. At this stage, the study area is the territory of Montenegro. Following the inventory stage, the indexing method relies on a normalizing procedure in QGIS and the assignment of weighted impact factor to each criterion via analytical hierarchy process (AHP). The WSCI value is derived as the sum of the products between the normalized class and the respective weighted impact factor of each criterion. Besides the methodological improvements the results of this work deliver tangible outputs in support of forest fire risk reduction in disaster risk management and fire safety agendas.

Automatic detection of stone pavement's pattern based on UAV photogrammetry

Pavement management system (PMS) is a set of tools that assist road agencies in finding optimal strategies for maintaining pavements in a serviceable condition over a period of time. Usually, municipalities base their PMS on the deterioration monitoring through a visual survey but the distresses identification is complex and the operations are based on visual and instrumental inspections. As regards natural stone pavements, which are very widespread in the road heritage of cities, in literature there are very few studies. The authors analyzed two supervised classification approaches (Semi-Automatic Classification Plugin for QGIS and a Convolutional Neural Network (CNN)), based on Unmanned Aerial Vehicle (UAV) photogrammetry, to detect stone pavement's pattern. This study showed that using a U-Net CNN on images obtained from UAV is an excellent alternative to the traditional manual inspection and can be implemented for other types of stone pavements, also with the aim of distress identification.

Using satellite images to study urban heat island and environmental monitoring

The paper describes a number of simplified methods of mapping of the urban environment surface temperature in order to study the structure of the urban heat island. It is very important to take into account the climatic and environmental variables of the territories when planning modern cities and settlements. To solve such problems, methods of remote sensing data are used increasingly. The study of multispectral satellite images provides more information about the characteristics of objects on the earth’s surface, as well as data on the distribution of various types of land use. Maps of the earth’s surface temperature (LST) can be used for studying the urban microclimate and localization of temperature anomalies of the city, as well as for urban planning tasks. The article considers the operation of processing a series of satellite images of urban areas for the summer period, followed by the study of temporary changes in the temperature portrait of the city. The paper considers the use of semi-automatic methods for obtaining satellite images in the thermal field, as well as a separate study of images in the near-infrared field to search for bright point sources of heat. Radiometric and atmospheric correction was performed on satellite data using the QGIS package and the Semiautomatic Classification Plugin extension. Based on the results of the study, thermal anomalies were localized in the city, and the boundaries of the heat footprint of new neighborhoods were determined.

Counting fuel properties as input in the wildfire spreading capacities of vegetated surfaces: case of Albania

Extreme weather conditions characterized by increased peak temperatures and stretched draught seasons are expected to boost up wildfire vulnerability in Mediterranean countries such as Albania. Thus, estimations about wildfire spread capacities of the territory are crucial. In this paper we introduce four new parameters into the indexing method for classifying the forested lands by their wildfire spreading capacity (WSCI). Land cover type via Corine Land Cover (CLC), Plant heat zones, Tree cover density (TCD), and Normalized difference vegetation index (NDVI) are integrated along with the previous set of criteria. The analytical steps of the process are performed in QGIS software including the Semi-Automatic Classification Plugin (SCP) which is useful in calculating NDVI values. The diversity among the inventory values of the selected criteria urges for a normalizing procedure within QGIS. Besides, each criterion is foreseen to have a specific impact on the WSCI value, which is weighted via Analytic Hierarchy Process (AHP). The sum of the products of the normalized class and the weighted impact factor of each criterion generates the WSCI value. The validation relies on the comparison between the index values of points being located within the burned areas and the values of the remaining locations. The results have shown that the former set of points have higher WSCI mean value then the latter group of points. Lastly, the parametric vulnerability assessment method presented here enables useful materials in support of wildfire risk reduction within the national priorities of disaster risk management and fire safety agendas in Albania.

Evaluating free and simple remote sensing methods for mapping Chinese privet (Ligustrum sinense) invasions in hardwood forests

Chinese privet (Ligustrum sinense) is a common invasive shrub in hardwood forests of the southeastern US and has been shown to negatively affect native herbaceous and woody plants. The ability to map the distribution of L. sinense on a property could help land managers plan and budget for control operations. We evaluated whether freely available moderate resolution multispectral imageries (Landsat 8 and Sentinel 2) and open-source GIS software (QGIS with the Semi-Automatic Classification Plugin) could be effective tools for this application. These tools are widely used by remote sensing and mapping professionals; however their adoption by field-level land managers appears limited, and their utility for mapping L. sinense invasions is untested. We evaluated how satellite type, image acquisition date, classification algorithm, and L. sinense cover affected detection accuracy. We found that Sentinel 2 imagery from March tended to produce good results, especially when analyzed using the maximum likelihood algorithm. Our best classifier obtained an overall accuracy of 92.3% for areas with ≥ 40% L. sinense cover. We recommend that land managers interested in applying this tool use an adaptive process for developing training polygons and test multiple images and classification algorithms in order to achieve optimal results.

Land-Use Change in New Moscow: First Outcomes after Five Years of Urbanization

Urbanization coincides with remarkable environmental changes, including conversion of natural landscapes into urban. Moscow megapolis is among the largest urbanized areas in Europe. An ambitious New Moscow project expanded the megapolis on extra 1500 km2 of former fallow lands, croplands and forests. The research aimed to monitor land use changes in New Moscow between 1989 and 2016 years. Landsat 5 and Landsat 8 images (30 m spectral resolution) and Sentinel – 2 images (10 m spectral resolution) were analyzed. All the images were collected for the similar summer period (from June to August). The images were preprocessed and classified by Semi-Automatic Classification Plugin in open source QGIS software to derive land cover maps. The following land cover classes were identified: water, built-up areas, bare soils, croplands and forested areas, and the total area covered by each class was estimated. The following land-use change pathways were reported: 1) reduction of the forested areas by 2.5% (almost 2000 ha) between 1989 and 1998; 2) partial reforestation (more than 1000 ha) and abandonment of croplands (more than 3000 ha) between 1998 and 2010 and 3) intensive urbanization (more than 11000 ha) between 2010 and 2016. New build-up areas and infrastructures were constructed on former forested areas and croplands. Although, some uncertainties in the absolute estimates are expected due to the classification errors, the general urbanization trend can be clearly distinguished as a principal outcome after the five years of New Moscow project.

A New Approach to High-Resolution Urban Land Use Classification Using Open Access Software and True Color Satellite Images

Urbanization nowadays results in the most dynamic and drastic changes in land use/land cover, with a significant impact on the environment. A detailed analysis and assessment of this process is necessary to take informed actions to reduce its impact on the environment and human well-being. In most parts of the world, detailed information on the composition, structure, extent, and temporal changes of urban areas is lacking. The purpose of this study is to present a methodology to produce high-resolution land use/land cover maps by the use of free software and satellite imagery. These maps can help to understand dynamic urbanizations processes to plan, design, and coordinate sustainable urban development plans, especially in areas with limited resources and advancing environmental degradation. A series of high-resolution true color images provided by Google Earth Pro were used to do initial classifications with the Semi-Automatic Classification Plug-in in QGIS. Afterwards, a new methodology to improve the classification by the elimination of shadows and clouds, and a reduction of misclassifications through superimposition was applied. The classification was carried out for three urban areas in León, Nicaragua, with different degrees of urbanization for the years 2009, 2015, and 2018. Finally, the accuracy of the classification was analyzed using randomly defined validation polygons. The results are three sets of high-resolution land use/land cover maps of the initial and the improved classification, showing the detailed structures and temporal dynamics of urbanization. The average accuracy of classification reaches 74%, but up to 85% for the best classification. The results clearly identify advancing urbanization, the loss of vegetation and riparian zones, and threats to urban ecosystems. In general, the level of detail and simplicity of our methodology is a valuable tool to support sustainable urban management, although its application is not limited to these areas and can also be employed to track changes over time, providing therefore, relevant information to a wide range of decision-makers.

Land use\u2212land cover changes and Mopani worm harvest in Mangwe District in Plumtree, Zimbabwe

BackgroundMopani worms have been considered a critical food security resource as people living in Mopani woodlands depend on the worms both as a food source and as an income generating resource. These become a readily available substitute if agriculture fails and droughts loom. However the yields from these worms have been dwindling over the years and this has been associated with land use−land cover changes as the worms depend on vegetation. This research sought to investigate the relationship between land use−land cover changes and Mopani worm harvests in Mangwe District in Plumtree from the period 2007 to 2016 in Zimbabwe.MethodsSatellite imagery was collected using LandSat 5 and LandSat 8 satellite and then classified using the Semi-Automatic Classification plugin in Quantum Geographic Information System to identify trees, dams, bare soil and settlements. Thematic maps were then produced and used to quantify extent of Land Use–Land Cover changes in the period from 2007 to 2016. Ground control data was collected using hand held Global Positioning System. Harvests trends (and reasons thereof) were estimated through usage of interviewer administered questionnaires on selected Mopani worm harvesters and harvest data kept by the community leaders.ResultsResults showed that settlements and bare soil cover had greatly increased from 2007 to 2016. There was a significant decline on vegetation cover from 2007 to 2016. There was also decline in the Mopani worm harvest over the years. Reasons for reduction of harvest were cited as deforestation, drought, overharvesting, cooking with lids on and occurrence of veld fires.ConclusionsFindings of this study highlight that land use−land cover changes impacted negatively on Mopani worm harvests over the period 2007 to 2016. It is also of paramount importance to note that major reasons for this decline has been associated with Land Use changes that are associated with human activities. There is need for Natural Resource Conservation Agencies to embark on awareness campaigns to encourage reforestation and also control cutting down of Mopani woodlands. Research needs to be conducted to determine the association between cultural beliefs and Mopani worm conservation and utilisation.

Estimation of the Thermodynamic Temperature of the Earth's Surface Using Satellite Data Based on the Land Cover Classification in the Optical Radiation Range

The method for estimation of the thermodynamic temperature field of the Earth's surface using satellite data of the long-wave infrared range is studied and implemented in the environment of Quantum GIS using the Semi-Automatic Classification Plugin. This method is based on the land cover classification in the optical radiation range using the machine learning. The supervised land cover classification into four main macroclasses was carried out using the maximum likelihood method according to the reflectivity for the formed data set of spectral channels in the optical radiation range. When classifying, for each macroclass several training areas are created, each of which defines the certain child class. Training regions are formed by region growing method by attaching adjacent pixels to some selected pixel-seed based on the proximity of their spectral signature vectors. The reclassification of the resulting classification raster was performed, and for each macroclass the characteristic known value of the thermal emissivity was assigned. The research results are illustrated by the example of the estimation of the surface thermodynamic temperature of the wetland mineral and peat soils in the lowlands of the Kyiv region using satellite images of Landsat-8 (OLI, TIRS). It has been established that the contours of heated terrain areas, obtained from the conditions of exceeding of the experimentally selected threshold values of the thermodynamic temperature, based on the constructed temperature raster of the land surface, are consistent with the information of the State Emergency Service of Ukraine about dates and places of the peatland fires.

Land use/land cover dynamics using landsat data in a gold mining basin-the Ankobra, Ghana

The Ankobra River basin, which forms part of the Southern-Western River System of Ghana, has become a hub for both large and small-scale mining activities. Some of the activities of these mines threatens the sustainability of its resources in the basin. A study in the basin which focuses on changes in LULC patterns in the basin over the past years is therefore paramount to understand the changes that has taken place and its potential impact on resources in the basin. This study assessed the pattern of Land use/cover changes, and the possible drivers of change, in the basin. The study used multi-spectral Landsat images of 30 m resolution for the years 1991, 2002, 2008 and 2016. The Semi-Automatic Classification Plugin (SCP) in QGIS was used for atmospheric correction, image classification using the spectral angle mapping algorithm, and post processing. The overall difference between the reference and the classified map of 2016 was 5.5% with the overall quantity, exchange and shift components as 1.5%, 3% and 1% respectively. Findings from the study show that, the Ankobra River basin has witnessed noticeable changes over the 25-year study period. Closed forest which occupied 40.4% of the total basin area in 1991 reduced drastically to 22.8% in 2016. The dominant LULC change patterns in the basin are from Closed Forest to Open Forest; Open Forest to Farmland, Settlement/Bare land and Mining Area; and the increase in surface area of Water consequently resulting from the increase in Mining Area. Mining activities, particularly illegal mining was identified to be the dominant driver of deforestation in the Ankobra Basin between the years 2008 and 2016 where mining activities in the basin sharply increased.

DEVELOPING A LAND COVER CLASSIFICATION OF SALT MARSHES USING UAS TIME-SERIES IMAGERY AND AN OPEN SOURCE WORKFLOW

Abstract. Salt marsh ecology classification is difficult using traditional coarse resolution remote sensing techniques. Salt marshes exhibit a spatial pattern of vegetation zonation that are visually identifiable using imagery that has an improved 0.04 meter per pixel resolution. This project applies high resolution unmanned aerial system (UAS) imagery to aid in multi-temporal classification of our study area (Horseneck Beach) in Westport, Massachusetts, USA. We flew a DJI Phantom Pro 3 at low- and high-tide to capture effects the changing tide has on vegetation in an effort to predict effects of the rising sea level on saline plant species. We implement an open source software workflow using OpenDroneMap and the Semi-Automatic Classification Plugin for QGIS to create the necessary orthomosaics and to conduct vegetation classification required of this project. We compare land cover classifications using one-time-point RGB imagery to a multi-time-point (low tide, high tide) RGB image stack to investigate whether the multi-time point stack improves land cover classification accuracy. We find it does. More generally, this paper provides a model for others wishing to use low-cost UAS equipment carrying a simple low-cost RGB camera, and free and open source for geospatial (FOSS4G) tools, to develop multi-band image stacks to improve land cover classification accuracy. Further, we provide some reflections and technical notes on our experience. The approach we present here could be extended to include other image layers that UAS can provide when equipped with other sensors, such as multi-spectral (e.g., NIR, thermal), or by adding another band with photogrammetry-produced digital elevation data.