APPENDICES - Laserfiche WebLink

Land Cover Classification Methodology The 2022 land cover classification was created using 2022 aerial imagery and lidar provided by the City. Dudek produced the land cover product by creating a composite image using red, green, and blue from the satellite imagery, normalized difference vegetation index (NDVI) (a measure of photosynthesis derived from infrared bands in the satellite imagery), and height data derived from the lidar. The composite image was then segmented, and representative training samples were selected from the segments for each land classification. The composite image and training samples were loaded into a deep learning model, using the Unet classifier architecture. The land classification was manually edited in problem areas. A stratified random sample of over 50 locations was used for each land classification to generate a confusion matrix to calculate overall accuracy measures. Canopy Change Analysis Methodology The canopy change analysis utilized urban tree canopy cover products from two different years. The 2012 canopy cover product was purchased from EarthDefine. The 2018 canopy cover product was freely downloaded from the U.S. Forest Service (USFS). According to the methodology, the 2018 product was created by EarthDefine using 2018 National Agriculture Imagery Program (NAIP) aerial imagery and lidar data collected by the U.S. Geological Survey. The NAIP imagery was acquired during the growing season and included four spectral bands (red, green, blue, and near infrared) at a 60-cm spatial resolution. The 2018 canopy cover product did not cover the entire city boundary. Approximately 12% of the city boundary, primarily non-urban and forested areas in the west and some areas in the east, were excluded from this analysis. The urban boundary defined in the 2018 product was used as the analysis boundary because it is expected to be freely available every four years. This provides a consistent boundary for future analyses, should the City choose to monitor canopy cover progress using this data source.