Techniques and solutions are described for improving automated user interface generation. Devices can automatically be recognized in one or more images of a system or environment. At least certain devices can be identified as standard devices, and associated with device profiles. The device profiles can include information useable to identify user interface controls that should be rendered on a user interface for the system or environment. The user interface controls can be rendered over an image or schematic diagram of the system or environment, including at locations that correspond to a geospatial location of the corresponding device, or a controllable element thereof. The user interface controls can be associated with commands of a standardized command schema, which in turn are mapped to specific commands that can be sent to control the devices or controllable elements thereof.