The snapshot area is given in device pixels (i.e. pixels on the resulting image if you took a full screenshot). That is different to layout co-ordinates (i.e. object positions) which depends on scrolling, scaling, the viewport size, etc.
The system LayerToCanvasX/Y expressions can convert a layout position to canvas CSS co-ordinates, and then if you multiply that by PlatformInfo.DevicePixelRatio, that should give you the device co-ordinates.