I have several square km of colour aerial imagery at 12.5cm resolution, showing vegetated wetland areas with visible deer trackways. These can be manually traced but it takes considerable time. Is there scope to use Deep Learning to extract linear features? I see that the Label Objects for Deep Learning tool offers only polygons.
I see there is online training at https://www.esri.com/training/catalog/5eb18cf2a7a78b65b7e26134/deep-learning-using-arcgis/ but if linear feature extraction is not possible then I'd rather not spend the time on this.
I'm a university educator with several years of general desktop GIS experience, although I have not really used Python and only a little bit of supervised image classification. Any comments on the likely learning curve with deep learning? Other recommended resources?
Thanks for any pointers!
