This SALTIEL_readme20220420.txt file was generated on 20220420 by Troy Saltiel. ------------------- GENERAL INFORMATION ------------------- 1. Dataset for: Howard Slough Waterfowl Management Area Multispectral Imagery at Various Resolutions and Convolutional Neural Network Training Data 2. Author Information Principal Investigator Contact Information Name: Troy M. Saltiel Institution: University of Utah Address: University of Utah Department of Geography, 260 Central Campus Drive, #4625, Salt Lake City, UT 84112 Email: troy.saltiel@utah.edu Associate or Co-investigator Contact Information Name: Philip E. Dennison Institution: University of Utah Address: University of Utah Department of Geography, 260 Central Campus Drive, #4625, Salt Lake City, UT 84112 Email: dennison@geog.utah.edu Alternate Contact Information Name: Michael J. Campbell Institution: University of Utah Address: University of Utah Department of Geography, 260 Central Campus Drive, #4625, Salt Lake City, UT 84112 Email: mickey.campbell@geog.utah.edu Name: Thomas R. Thompson Institution: University of Utah Address: Division of Forestry, Fire and State Lands, 1594 West North Temple, Suite 3520, Salt Lake City, Utah 84114 Email: tomthompson@utah.gov Name: Keith R. Hambrecht Institution: University of Utah Address: Division of Forestry, Fire and State Lands, 1594 West North Temple, Suite 3520, Salt Lake City, Utah 84114 Email: khambrecht@utah.gov 3. Date of data collection (single date, range, approximate date) 20200811 4. Geographic location of data collection (where was data collected?): 41.1127203, -112.1482218 5. Information about funding sources that supported the collection of the data: N/A -------------------------- SHARING/ACCESS INFORMATION -------------------------- 1. Licenses/restrictions placed on the data: CC BY NC – Allows others to use and share your data non-commercially and with attribution 2. Links to publications that cite or use the data: Saltiel, T.M., Dennison, P.E., Campbell, M.J., Thompson, T., and Hambrecht K. 2022. Tradeoffs Between UAS Spatial Resolution and Accuracy for Deep Learning Semantic Segmentation Applied to Wetland Vegetation Species Mapping. Submitted to Remote Sensing. 3. Links to other publicly accessible locations of the data: N/A 4. Links/relationships to ancillary data sets: N/A 5. Was data derived from another source? No 6. Recommended citation for the data: Saltiel, T.M.; Dennison, P.E.; Campbell, M.J.; Thompson, T.; Hambrecht K. 2022. Dataset for: "Howard Slough Waterfowl Management Area Multispectral Imagery at Various Resolutions and Convolutional Neural Network Training Data." The Hive: University of Utah Research Data Repository https://doi.org/10.7278/S50d-h9z0-5ft8 --------------------- DATA & FILE OVERVIEW --------------------- 1. File List A. Filename: imagery.zip Short description: Original image captures, resampled (coarsened) images, code to coarsen imagery, and tool to transfer exif data from original imagery to resampled imagery. B. Filename: train_test_images.zip Short description: Train and test images at various resolutions for training a convolutional neural network model. C. Filename: Short description: 2. Relationship between files: The training and testing images (train_test_images.zip) may be used to train a convolutional neural network model, evaluate its accuracy, then predict on the original imagery (imagery.zip). For visualization purposes, the original imagery should be stitched to a single orthoimage. This is possible with commerical software like Pix4D (such as in the cited study), but also possible with open source software like OpenDroneMap. 3. Additional related data collected that was not included in the current data package: Ground reference points are not downloadable but are accessible on the Utah DNR ArcGIS Server: https://services.arcgis.com/ZzrwjTRez6FJiOq4/ArcGIS/rest/services/GSL_Phrag_Mapping_Ground_Control_Points/FeatureServer Note that points classed as 'other' were reclassified using the associated ground photo and feature notes (see cited publication, Figure 3). 4. Are there multiple versions of the dataset? yes/no? No -------------------------- METHODOLOGICAL INFORMATION -------------------------- 1. Description of methods used for collection/generation of data: Imagery captured with Parrot DISCO unoccupied aerial vehicle and Red-Edge MX multispectral image sensor. Ground reference points captured with GPS enabled smartphone and ArcGIS Collector application. Training and testing imagery created in ArcGIS Pro. See citing article for additional methodology. 2. Methods for processing the data: imagery.zip is raw/collected data. train_test_images.zip is derived from imagery.zip. imagery.zip was orthorectified in Pix4D software, clipped to a study area, an image segmentation was generated in ArcGIS Pro, then the segments were manually labeled to produce the classified images (see citing article for more detailed methodology). 3. Instrument- or software-specific information needed to interpret the data: Original imagery captured by MicaSense Red-Edge MX multispectral image sensor. In the citing article, imagery was orthorectified with Pix4D Mapper 4.6.4 and imagery was clipped and labeled with ArcGIS Pro 2.7. 4. Standards and calibration information, if appropriate: N/A 5. Environmental/experimental conditions: Imagery captured under mostly clear skies. 6. Describe any quality-assurance procedures performed on the data: N/A 7. People involved with sample collection, processing, analysis and/or submission: Sample collection by Christian Hardwick (Utah Geological Survey), data processed and submitted by Troy Saltiel. ----------------------------------------- DATA-SPECIFIC INFORMATION FOR: imagery.zip ----------------------------------------- 1. Number of variables: 5 (image bands), suffix of image file names 2. Number of cases/rows: N/A 3. Variable List A. Name: 1 Description: Blue B. Name: 2 Description: Green A. Name: 3 Description: Red B. Name: 4 Description: Red-Edge A. Name: 5 Description: Near-Infrared 4. Missing data codes: 65535 (digital number or pixel value) 5. Specialized formats of other abbreviations used N/A ----------------------------------------- DATA-SPECIFIC INFORMATION FOR: train_test_images.zip ----------------------------------------- 1. Number of variables: 6 (digital number, pixel values of classified/labeled images) 2. Number of cases/rows: N/A 3. Variable List A. Name: 0 Description: Phragmites australis B. Name: 1 Description: Cattail B. Name: 2 Description: Water A. Name: 3 Description: Non-Photosynthetic Material B. Name: 4 Description: Algae A. Name: 5 Description: Bulrush 4. Missing data codes: N/A 5. Specialized formats of other abbreviations used N/A