Context and objectives
An accurate knowledge of the state and dynamics of a population is a major prerequisite for a sound management of social and economic development; it is also an important specific tool, essential for the management of critical situations involving huge population displacements (wars, disasters …). Unfortunately, many countries, especially in Africa, only have at their disposal scarce scattered population data: population censuses are taken every ten year in the best case, and results come late and provide seldom if ever the expected regional detail; very few surveys are designed in a way suitable for the expected kind of data; the scope of most surveys is narrow (reproduction, public health), their regional coverage is often limited and its results are anyway unusable because of sample size; questionnaires and field procedures are often unclear and quite unreliable; finally, administrative data (population registers) are virtually non existing in most countries.
Since the late nineties, VHR images supplement the range of remote sensing data (aerial photo-graphs, satellite images in several resolutions …). They have quickly been felt to be a new oppor-tunity for a better knowledge of population. However, it is obvious that « everything » may not be seen from space (e.g. number of persons living in a house and their characteristics); but we also lack of methods enabling us to infer most of the invisible characteristics from remote sensing data.
The final objective is to identify, to test and to refine methods for estimating population parameters from HR/VHR images and ad hoc inputs from the field.
Project outcome
Remote sensing techniques and field demographic techniques have been jointly used with the final objective to develop an operational procedure for quick population estimates. Two pilot studies have been performed on Lubumbashi, RDC and Alibori, Benin; including field missions in order to collect demographic data and validate the extraction of features from satellite imagery.
From a demographer point of view remote sensing will remain an irreplaceable support for the mapping tasks prior to any demographic field survey, especially the light ones. This allows the demographers to save time and money. In Africa preparatory mapping tasks might represent over a quarter of the budget of a demographic survey when classical mapping methods are used.
The contribution of remote sensing to the demographer is not questionable. This was again proved on the two pilot studies in Lubumbashi and Alibori.
The final results on the urban area are very satisfactory. The estimate of the population in Lubumbashi (ca. 1.160.000 inhabitants) is close to what can be expected taking into account the last official population data and the characteristics of the population. The application of very high resolution remote sensing was extremely valuable and must be seen as a mandatory tool for future urban demographic studies.
The study on Benin has shown that high resolution imagery (SPOT5) was as useful as very high resolution imagery (Quickbird) to extract the villages and estimate their surface. The field survey has allowed to estimate the average population density per village. It also confirmed that the smallest villages that weren’t extracted from the SPOT5 images are not demographically significant.
From a financial point of view the use of SPOT5 instead of Quickbird is extremely relevant for rural areas.
Both exercises have shown the importance of knowing the history of the observed area. This is particularly true on cities like Lubumbashi. The pilot studies have emphasized the importance to rely on exact and agreed administrative limits to compare satellite-derived population estimates to field data.
Finally the tools and the methods developed have been transferred to the Keyobs team for future operational applications.
| Project leader(s): | KeyObs | |||
| Belgian partner(s) |
|
|||
| Location: |
Region:
|
|||
| Related presentations: | ||||
| Related publications: |
|
|||
| Website: | http://igeat.ulb.ac.be/fr/projets/details/project/population-estimation-by-remote-sensing/ | |||