A couple months back, I posted a question about separating evergreens from deciduous trees in the output from the pretrained tree detection deep learning model. Now that I've completed a storymap on methods, I hope to get input on this from the community as to how I can improve my process. Please share any ideas you have. Though I don't have a formal accuracy assessment, I'm happy to say our urban foresters have found this useful in the field.
https://storymaps.arcgis.com/stories/57bc5e45a5d74c59a41601922d72a656
