Deep Learning Analysis

Deep Learning Analysis in Global Mapper is a set of deep learning-based imagery analysis tools.

Deep Learning Analysis is comprised of the following tools:

Deep Learning Analysis tools can be accessed from the Deep Learning Dropdown Menu, or from the Deep Learning Toolbar.

• Land Cover Classification - The Land Cover Classification tool can be used to classify land cover data from aerial imagery.
• Object Detection - The Object Detection tool allows for users to extract building footprints, and identify vehicles from high resolution aerial imagery.
• Fine-Tuning - The Fine-Tuning tool can be used to adjust the built in models included in the Deep Learning Analysis tools to better fit user data, as well as create new custom models from scratch.

Note: These tools will utilize CUDA-enabled GPUs if available by default. Using a machine with a dedicated Nvidia CUDA-enabled GPU can greatly improve processing time. If a machine does not have a compatible GPU, the tools will run on CPU. There is also an option in the Deep Learning Models section of Configuration to turn off GPU usage even on machines with a compatible GPU.