My question is about calculation surface temperature in that case the Thermal Infrared band remotely sensed by satellite or UAV does not exist. But accurate and detailed NIR (Near Infrared), Red Edge and red, blue, green and RGB exist.
Your question needs more clarification. Do you mean LST measurement by the satellite images using e.g. Landsat, MODIS, or other ones? Or you mean by the handheld or UAV measurement? What spatial and temporal resolution are you looking for?
If I understand well your question, the answer would be quite negative. If you try to use NIR bands they are obviously contaminated by reflectance throughout the atmosphere. It would be difficult to disentangle the thermal contribution and for doing that you would need thermal IR channels once more...
RGB imagery is composed by spectral bands and thus does not mean anything physical apart from the content of the various bands used to make up the composite. So, this is not a method that make too much sense to me.
However, you may want to explore passive microwave bands from several sensors. For example, you may look at the following literature
https://doi.org/10.1364/OE.25.00A940
Article A framework for the retrieval of all-weather land surface te...
You may find a more fundamental study on land surface emissivity retrieval from microwave bands at