Environment synthesis for lighting an object

A method of rendering a reflective surface of a computer-generated reality (“CGR”) object based on surface synthesis in a CGR environment, comprising:

• at a device with one or more processors and a computer readable medium:
• determining a first geometry and textural information of a CGR environment, wherein the CGR environment is based on one or more camera images of a physical scene;
• determining that the CGR environment includes a surface having a captured surface portion and a missing portion, the surface in the CGR environment representing a surface in the physical scene and the missing portion of the surface in the CGR environment corresponding to an uncaptured portion of the surface in the physical scene;
• generating an extension portion for the missing portion based on an assessment of whether the captured surface portion has been cutoff by a boundary defined by a captured geometry, wherein the extension portion is generated by extending a 3D geometry of the captured surface portion based on a characteristic of the captured surface portion;
• determining a second geometry of the CGR environment representing the physical scene based on the extension portion;
• determining second textural information, wherein the second textural information is synthesized for the extension portion by:
• determining a characteristic of the captured surface portion that matches a stored reference characteristic of an object;
• or
• identifying a repeating pattern or texture of the captured surface portion;
• generating an environment map using the determined second geometry and the second textural information;
• and
• rendering a CGR object in the CGR environment, wherein an appearance of a reflective surface of the CGR object is determined based on the environment map.