This research demonstrates the potential of nighttime satellite imagery as provided by the Defense Meteorological Satellite Program's Operational Linescan System (DMSP OLS) for making spatially explicit estimates of human populations. The methods presented make estimates of aggregate and disaggregate population parameters. Disaggregate measures of population are intra-urban measures of population density based on measures of light intensity. An argument will be made that light intensity based estimates of population density represent a temporal average of population density that results from human mobility. Aggregate measures of population are estimates of total city populations. The total population of a city is estimated by measuring the areal extent of the city in the imagery. A strong linear relationship between the natural log of population and the natural log of city areal extent is demonstrated. Incorporating national statistics such as GDP per capita strengthens this linear relationship. Total national population estimates are obtained from these estimates of urban population by using published ratios of total to urban population for each country. The resulting estimate of the 1997 global human population is 6.3 billion, a number close to the generally accepted estimate of 5.8 billion. Chinese cities were excluded from the 'training' of the model resulting in an independent estimate of China's population at 1.512 billion, which is 22% higher than the accepted estimate of 1.242 billion. In addition, the method, which implicitly maps urban land cover, estimates 0.81% of the earth's land surface as urban. This is a factor of five greater than state of the art estimates of global land cover produced by the United States Geological Survey. Nighttime satellite imagery based estimates of urban populations and intra-urban population density show great promise as an inexpensive supplement to national censuses worldwide.