We propose an estimation methodology of per-capita incomes using satellite information on night luminosity (DMSP-OLS Nighttime Lights Time Series), using a structure of continuous spatial random effects and correction for measurement errors. This methodology allows the construction of income measures for disaggregated units and also building income estimates for years without census information. We apply this methodology using information for the Brazilian territory, and the results indicate that inclusion of spatial random effects is critical to the accuracy of the estimated income, and that these spatial effects are persistent in time.

中文翻译：

---

使用夜光度的收入估算：连续空间模型

我们提出了一种使用连续夜空亮度效应结构和测量误差校正的夜间亮度卫星信息（DMSP-OLS夜间灯光时间序列）的人均收入估算方法。这种方法不仅可以构建分类单位的收入度量，还可以在没有普查信息的情况下建立多年来的收入估算。我们使用巴西领土信息使用这种方法，结果表明，空间随机效应的纳入对于估算收入的准确性至关重要，而且这些空间效应在时间上是持久的。