Speaking as a non-professional, I have seen no such detailed datasets.

Your research should not depend on clueless strangers, though. My first hit on Chinese statistics was: http://www.stats.gov.cn/english/statisticaldata/ , and I'd think they are worth exploring. Same for India.

Gathering detailed income data is fraught with difficulty on many levels: the ones with access to the high-fidelity data (local tax authorities) won't talk for a host of reasons, apart from confidentiality (quite often they won't be bothered to help other government agencies), while aggregation over villages/counties/whatever lowest-level administrative units is not something in high demand from national and supranational decision-makers. I leave aside the obvious problems of tax evasion, non-response, and outright fabrication.

Thus, for countries with less than well-funded statistical agencies, one is forced to look for proxies.

The usual proxy is electricity consumption; while the utilities are loath to give outsiders access, as a rough proxy, one can simply look at the night lighting: http://geology.com/articles/satellite-photo-earth-at-night.shtml.

I'd recommend against turning to land use data - they can't tell you much about city-dwellers or income levels. You can ask folks at GIS SE for details, though.

Access to sanitation is another proxy (albeit a non-linear one). World Bank, local authorities, utility companies, food inspection agencies can possibly serve as sources, I'd guess.

Whatever you do, please remember to compile list of references, prior and related studies, validate and cross-check your data. Ideally, you would do an on-site survey for that (assuming you have got some money to spend). Beware of systematic bias creeping in (it will!) without you noticing or telling users of your data.