Thirty years of use and improvement of remote sensing, applied to epidemiology: From early promises to lasting frustration