This course explores the theory, technology, and applications of remote sensing. It is designed for individuals with an interest in GIS and geospatial science who have no prior experience working with remotely sensed data. Lab exercises make use of the web and the ArcGIS Pro software. You will work with and explore a wide variety of data types including aerial imagery, satellite imagery, multispectral imagery, digital terrain data, light detection and ranging (LiDAR), thermal data, and synthetic aperture RaDAR (SAR). Remote sensing is a rapidly changing field influenced by big data, machine learning, deep learning, and cloud computing. In this course you will gain an overview of the subject of remote sensing, with a special emphasis on principles, limitations, and possibilities. In addition, this course emphasizes information literacy, and will develop your skills in finding, evaluating, and using scholarly information.

You will be asked to work through a series of modules that present information relating to a specific topic. You will also complete a series of lab exercises to reinforce the material. Lastly, you will complete paper reviews and a term project. We have also provided additional bonus material and links associated with surface hydrologic analysis with TauDEM, geographic object-based image analysis (GEOBIA), Google Earth Engine (GEE), and the geemap Python library for Google Earth Engine. Please see the sequencing document for our suggested order in which to work through the material. We have also provided PDF versions of the lectures with the notes included.

The classification of point cloud datasets to identify distribution wires is useful for identifying vegetation encroachment around power lines. Such workflows are important for preventing fires and power outages and are typically manual, recurring, and labor-intensive. This model is designed to extract distribution wires at the street level. Its predictions for high-tension transmission wires are less consistent with changes in geography as compared to street-level distribution wires. In the case of high-tension transmission wires, a lower ‘recall’ value is observed as compared to the value observed for low-lying street wires and poles.Using the modelFollow the guide to use the model. The model can be used with ArcGIS Pro's Classify Point Cloud Using Trained Model tool. Before using this model, ensure that the supported deep learning libraries are installed. For more details, check Deep Learning Libraries Installer for ArcGIS.InputThe model accepts unclassified point clouds with point geometry (X, Y and Z values). Note: The model is not dependent on any additional attributes such as Intensity, Number of Returns, etc. This model is trained to work on unclassified point clouds that are in a projected coordinate system, in which the units of X, Y and Z are based on the metric system of measurement. If the dataset is in degrees or feet, it needs to be re-projected accordingly. The model was trained using a training dataset with the full set of points. Therefore, it is important to make the full set of points available to the neural network while predicting - allowing it to better discriminate points of 'class of interest' versus background points. It is recommended to use 'selective/target classification' and 'class preservation' functionalities during prediction to have better control over the classification and scenarios with false positives.The model was trained on airborne lidar datasets and is expected to perform best with similar datasets. Classification of terrestrial point cloud datasets may work but has not been validated. For such cases, this pre-trained model may be fine-tuned to save on cost, time, and compute resources while improving accuracy. Another example where fine-tuning this model can be useful is when the object of interest is tram wires, railway wires, etc. which are geometrically similar to electricity wires. When fine-tuning this model, the target training data characteristics such as class structure, maximum number of points per block and extra attributes should match those of the data originally used for training this model (see Training data section below).OutputThe model will classify the point cloud into the following classes with their meaning as defined by the American Society for Photogrammetry and Remote Sensing (ASPRS) described below: Classcode Class Description 0 Background Class 14 Distribution Wires 15 Distribution Tower/PolesApplicable geographiesThe model is expected to work within any geography. It's seen to produce favorable results as shown here in many regions. However, results can vary for datasets that are statistically dissimilar to training data.Model architectureThis model uses the RandLANet model architecture implemented in ArcGIS API for Python.Accuracy metricsThe table below summarizes the accuracy of the predictions on the validation dataset. - Precision Recall F1-score Background (0) 0.999679 0.999876 0.999778 Distribution Wires (14) 0.955085 0.936825 0.945867 Distribution Poles (15) 0.707983 0.553888 0.621527Training dataThis model is trained on manually classified training dataset provided to Esri by AAM group. The training data used has the following characteristics: X, Y, and Z linear unitmeter Z range-240.34 m to 731.17 m Number of Returns1 to 5 Intensity1 to 4095 Point spacing0.2 ± 0.1 Scan angle-42 to +35 Maximum points per block20000 Extra attributesNone Class structure[0, 14, 15]Sample resultsHere are a few results from the model.

This layer displays a global map of land use/land cover (LULC) derived from ESA Sentinel-2 imagery at 10m resolution. Each year is generated with Impact Observatory’s deep learning AI land classification model, trained using billions of human-labeled image pixels from the National Geographic Society. The global maps are produced by applying this model to the Sentinel-2 Level-2A image collection on Microsoft’s Planetary Computer, processing over 400,000 Earth observations per year.The algorithm generates LULC predictions for nine classes, described in detail below. The year 2017 has a land cover class assigned for every pixel, but its class is based upon fewer images than the other years. The years 2018-2023 are based upon a more complete set of imagery. For this reason, the year 2017 may have less accurate land cover class assignments than the years 2018-2023.Variable mapped: Land use/land cover in 2017, 2018, 2019, 2020, 2021, 2022, 2023Source Data Coordinate System: Universal Transverse Mercator (UTM) WGS84Service Coordinate System: Web Mercator Auxiliary Sphere WGS84 (EPSG:3857)Extent: GlobalSource imagery: Sentinel-2 L2ACell Size: 10-metersType: ThematicAttribution: Esri, Impact ObservatoryWhat can you do with this layer?Global land use/land cover maps provide information on conservation planning, food security, and hydrologic modeling, among other things. This dataset can be used to visualize land use/land cover anywhere on Earth. This layer can also be used in analyses that require land use/land cover input. For example, the Zonal toolset allows a user to understand the composition of a specified area by reporting the total estimates for each of the classes. NOTE: Land use focus does not provide the spatial detail of a land cover map. As such, for the built area classification, yards, parks, and groves will appear as built area rather than trees or rangeland classes.Class definitionsValueNameDescription1WaterAreas where water was predominantly present throughout the year; may not cover areas with sporadic or ephemeral water; contains little to no sparse vegetation, no rock outcrop nor built up features like docks; examples: rivers, ponds, lakes, oceans, flooded salt plains.2TreesAny significant clustering of tall (~15 feet or higher) dense vegetation, typically with a closed or dense canopy; examples: wooded vegetation, clusters of dense tall vegetation within savannas, plantations, swamp or mangroves (dense/tall vegetation with ephemeral water or canopy too thick to detect water underneath).4Flooded vegetationAreas of any type of vegetation with obvious intermixing of water throughout a majority of the year; seasonally flooded area that is a mix of grass/shrub/trees/bare ground; examples: flooded mangroves, emergent vegetation, rice paddies and other heavily irrigated and inundated agriculture.5CropsHuman planted/plotted cereals, grasses, and crops not at tree height; examples: corn, wheat, soy, fallow plots of structured land.7Built AreaHuman made structures; major road and rail networks; large homogenous impervious surfaces including parking structures, office buildings and residential housing; examples: houses, dense villages / towns / cities, paved roads, asphalt.8Bare groundAreas of rock or soil with very sparse to no vegetation for the entire year; large areas of sand and deserts with no to little vegetation; examples: exposed rock or soil, desert and sand dunes, dry salt flats/pans, dried lake beds, mines.9Snow/IceLarge homogenous areas of permanent snow or ice, typically only in mountain areas or highest latitudes; examples: glaciers, permanent snowpack, snow fields.10CloudsNo land cover information due to persistent cloud cover.11RangelandOpen areas covered in homogenous grasses with little to no taller vegetation; wild cereals and grasses with no obvious human plotting (i.e., not a plotted field); examples: natural meadows and fields with sparse to no tree cover, open savanna with few to no trees, parks/golf courses/lawns, pastures. Mix of small clusters of plants or single plants dispersed on a landscape that shows exposed soil or rock; scrub-filled clearings within dense forests that are clearly not taller than trees; examples: moderate to sparse cover of bushes, shrubs and tufts of grass, savannas with very sparse grasses, trees or other plants.Classification ProcessThese maps include Version 003 of the global Sentinel-2 land use/land cover data product. It is produced by a deep learning model trained using over five billion hand-labeled Sentinel-2 pixels, sampled from over 20,000 sites distributed across all major biomes of the world.The underlying deep learning model uses 6-bands of Sentinel-2 L2A surface reflectance data: visible blue, green, red, near infrared, and two shortwave infrared bands. To create the final map, the model is run on multiple dates of imagery throughout the year, and the outputs are composited into a final representative map for each year.The input Sentinel-2 L2A data was accessed via Microsoft’s Planetary Computer and scaled using Microsoft Azure Batch.CitationKarra, Kontgis, et al. “Global land use/land cover with Sentinel-2 and deep learning.” IGARSS 2021-2021 IEEE International Geoscience and Remote Sensing Symposium. IEEE, 2021.AcknowledgementsTraining data for this project makes use of the National Geographic Society Dynamic World training dataset, produced for the Dynamic World Project by National Geographic Society in partnership with Google and the World Resources Institute.

Land use data is critically important to the work of the Department of Water Resources (DWR) and other California agencies. Understanding the impacts of land use, crop location, acreage, and management practices on environmental attributes and resource management is an integral step in the ability of Groundwater Sustainability Agencies (GSAs) to produce Groundwater Sustainability Plans (GSPs) and implement projects to attain sustainability. Land IQ was contracted by DWR to develop a comprehensive and accurate spatial land use database for the 2021 water year (WY 2021), covering over 10.7 million acres of agriculture on a field scale and additional areas of urban extent.The primary objective of this effort was to produce a spatial land use database with an accuracy exceeding 95% using remote sensing, statistical, and temporal analysis methods. This project is an extension of the land use mapping which began in the 2014 crop year, which classified over 15 million acres of land into agricultural and urban areas. Unlike the 2014 and 2016 datasets, the annual WY datasets from and including 2018, 2019, 2020, and 2021 include multi-cropping.Land IQ integrated crop production knowledge with detailed ground truth information and multiple satellite and aerial image resources to conduct remote sensing land use analysis at the field scale. Individual fields (boundaries of homogeneous crop types representing true cropped area, rather than legal parcel boundaries) were classified using a crop category legend and a more specific crop type legend. A supervised classification process using a random forest approach was used to classify delineated fields and was carried out county by county where training samples were available. Random forest approaches are currently some of the highest performing methods for data classification and regression. To determine frequency and seasonality of multicropped fields, peak growth dates were determined for each field of annual crops. Fields were attributed with DWR crop categories, which included citrus/subtropical, deciduous fruits and nuts, field crops, grain and hay, idle, pasture, rice, truck crops, urban, vineyards, and young perennials. These categories represent aggregated groups of specific crop types in the Land IQ dataset.Accuracy was calculated for the crop mapping using both DWR and Land IQ crop legends. The overall accuracy result for the crop mapping statewide was 97% using the Land IQ legend (Land IQ Subclass) and 98% using the DWR legend (DWR Class). Accuracy and error results varied among crop types. Some less extensive crops that have very few validation samples may have a skewed accuracy result depending on the number and nature of validation sample points. DWR revised crops and conditions from the Land IQ classification were encoded using standard DWR land use codes added to feature attributes, and each modified classification is indicated by the value 'r' in the ‘DWR_REVISE' data field. Polygons drawn by DWR, not included in Land IQ dataset receive the 'n' code for new. Boundary change (i.e. DWR changed the boundary that LIQ delivered, could be split boundary) indicated by 'b'. Each polygon classification is consistent with DWR attribute standards, however some of DWR's traditional attribute definitions are modified and extended to accommodate unavoidable constraints within remote-sensing classifications, or to make data more specific for DWR's water balance computation needs. The original Land IQ classifications reported for each polygon are preserved for comparison, and are also expressed as DWR standard attributes. Comments, problems, improvements, updates, or suggestions about local conditions or revisions in the final data set should be forwarded to the appropriate Regional Office Senior Land Use Supervisor.Revisions were made if:- DWR corrected the original crop classification based on local knowledge and analysis,-PARTIALLY IRRIGATED CROPS Crops, irrigated for only part of their normal irrigation season were given the special condition of ‘X’,-In certain areas, DWR changed the irrigation status to non-irrigated. Among those areas the special condition may have been changed to 'Partially Irrigated' based on image analysis and local knowledge,- young versus mature stages of perennial orchards and vineyards were identified (DWR added ‘Young’ to Special Condition attributes),- DWR determined that a field originally classified ‘Idle’ or 'Unclassified' were actually cropped one or more times during the year,- the percent of cropped area was changed from the original acres reported by Land IQ (values indicated in DWR ‘Percent’ column),- DWR determined that the field boundary should have been changed to better reflect the cropped area of the polygon and is identified by a 'b' in the DWR_REVISED column,- DWR determined that the field boundary should have been split to better reflect separate crops within the same polygon and identified by a 'b' in the DWR_REVISED column,- The ‘Mixed’ was added to the MULTIUSE column refers to no boundary change, but percent of field is changed where more than one crop is found,- DWR identified a distinct early or late crop on the field before the main season crop (‘Double’ was added to the MULTIUSE column); if the 1st and 2nd sequential crops occupied different portions of the total field acreage, the area percentages were indicated for each crop).This dataset includes multicropped fields. If the field was determined to have more than one crop during the course of the WY (Water Year begins October 1 and ends September 30 of the following year), the order of the crops is sequential, beginning with Class 1. All single cropped fields will be placed in Class 2, so every polygon will have a crop in the Class 2 and CropType2 columns. In the case that a permanent crop was removed during the WY, the Class 2 crop will be the permanent crop followed by ‘X’ – Unclassified fallow in the Class 3 column. In the case of Intercropping, the main crop will be placed in the Class 2 column with the partial crop in the Class 3 column.A new column for the 2019, 2020, and 2021 datasets is called ‘MAIN_CROP’. This column indicates which field Land IQ identified as the main season crop for the WY representing the crop grown during the dominant growing season for each county. The column ‘MAIN_CROP_DATE’, another addition to the 2019, 2020, and 2021 datasets, indicates the Normalized Difference Vegetation Index (NDVI) peak date for this main season crop. The column 'EMRG_CROP' for 2019, 2020, and 2021 indicates an emerging crop at the end of the WY. Crops listed indicate that at the end of the WY, September 2021, crop activity was detected from a crop that reached peak NDVI in the following WY (2022 WY). This attribute is included to account for water use of crops that span multiple WYs and are not exclusive to a single WY. It is indicative of early crop growth and initial water use in the current WY, but a majority of crop development and water use in the following WY. Crops listed in the ‘EMRG_CROP’ attribute will also be captured as the first crop (not necessarily Crop 1) in the following WY (2022 WY). These crops are not included in the 2021 UCF_ATT code as their peak date occurred in the following WY.For the 2021 dataset new columns added are: 'YR_PLANTED' which represent the year orchard / grove was planted. 'SEN_CROP' indicates a senescing crop at the beginning of the WY. Crops listed indicate that at the beginning of the WY, October 2020, crop activity was detected from a crop that reached peak NDVI in the previous WY (2020 WY), thus was a senescing crop. This is included to account for water use of crop growth periods that span multiple WYs and are not exclusive to a WY. Crops listed in the ‘SEN_CROP’ attribute are also captured in the CROPTYP 1 through 4 sequence of the previous WY (2020 WY). These crops are not included in the 2021 UCF_ATT code as their peak NDVI occurred in the previous WY. CTYP#_NOTE: indicates a more specific land use subclassification from the DWR Standard Land Use Legend that is not included in the primary, DWR Remote Sensing Land Use Legend.DWR reviewed and revised the data in some cases. The associated data are considered DWR enterprise GIS data, which meet all appropriate requirements of the DWR Spatial Data Standards, specifically the DWR Spatial Data Standard version 3.6, dated September 27, 2023. This data set was not produced by DWR. Data were originally developed and supplied by Land IQ, LLC, under contract to California Department of Water Resources. DWR makes no warranties or guarantees - either expressed or implied - as to the completeness, accuracy, or correctness of the data. DWR neither accepts nor assumes liability arising from or for any incorrect, incomplete, or misleading subject data. Detailed compilation and reviews of Statewide Crop Mapping and metadata development were performed by DWR Land Use Unit staff, therefore you may forward your questions to Landuse@water.ca.gov.This dataset is current as of 2021.

Classifying trees from point cloud data is useful in applications such as high-quality 3D basemap creation, urban planning, and forestry workflows. Trees have a complex geometrical structure that is hard to capture using traditional means. Deep learning models are highly capable of learning these complex structures and giving superior results.Using the modelFollow the guide to use the model. The model can be used with the 3D Basemaps solution and ArcGIS Pro's Classify Point Cloud Using Trained Model tool. Before using this model, ensure that the supported deep learning frameworks libraries are installed. For more details, check Deep Learning Libraries Installer for ArcGIS.InputThe model accepts unclassified point clouds with the attributes: X, Y, Z, and Number of Returns.Note: This model is trained to work on unclassified point clouds that are in a projected coordinate system, where the units of X, Y, and Z are based on the metric system of measurement. If the dataset is in degrees or feet, it needs to be re-projected accordingly. The provided deep learning model was trained using a training dataset with the full set of points. Therefore, it is important to make the full set of points available to the neural network while predicting - allowing it to better discriminate points of 'class of interest' versus background points. It is recommended to use 'selective/target classification' and 'class preservation' functionalities during prediction to have better control over the classification.This model was trained on airborne lidar datasets and is expected to perform best with similar datasets. Classification of terrestrial point cloud datasets may work but has not been validated. For such cases, this pre-trained model may be fine-tuned to save on cost, time and compute resources while improving accuracy. When fine-tuning this model, the target training data characteristics such as class structure, maximum number of points per block, and extra attributes should match those of the data originally used for training this model (see Training data section below).OutputThe model will classify the point cloud into the following 2 classes with their meaning as defined by the American Society for Photogrammetry and Remote Sensing (ASPRS) described below: 0 Background 5 Trees / High-vegetationApplicable geographiesThis model is expected to work well in all regions globally, with an exception of mountainous regions. However, results can vary for datasets that are statistically dissimilar to training data.Model architectureThis model uses the PointCNN model architecture implemented in ArcGIS API for Python.Accuracy metricsThe table below summarizes the accuracy of the predictions on the validation dataset. Class Precision Recall F1-score Trees / High-vegetation (5) 0.975374 0.965929 0.970628Training dataThis model is trained on a subset of UK Environment Agency's open dataset. The training data used has the following characteristics: X, Y and Z linear unit meter Z range -19.29 m to 314.23 m Number of Returns 1 to 5 Intensity 1 to 4092 Point spacing 0.6 ± 0.3 Scan angle -23 to +23 Maximum points per block 8192 Extra attributes Number of Returns Class structure [0, 5]Sample resultsHere are a few results from the model.