Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Analyzing Poverty through Intra-Annual Time-Series: A Wavelet Transform Approach

Mohammad Kakooei, Klaudia Solska, Adel Daoud

TL;DR

The results indicate that integrating specific NDVI-derived features with multi-spectral data provides valuable insights for poverty analysis, emphasizing the importance of retaining intra-annual information.

Abstract

Reducing global poverty is a key objective of the Sustainable Development Goals (SDGs). Achieving this requires high-frequency, granular data to capture neighborhood-level changes, particularly in data scarce regions such as low- and middle-income countries. To fill in the data gaps, recent computer vision methods combining machine learning (ML) with earth observation (EO) data to improve poverty estimation. However, while much progress have been made, they often omit intra-annual variations, which are crucial for estimating poverty in agriculturally dependent countries. We explored the impact of integrating intra-annual NDVI information with annual multi-spectral data on model accuracy. To evaluate our method, we created a simulated dataset using Landsat imagery and nighttime light data to evaluate EO-ML methods that use intra-annual EO data. Additionally, we evaluated our method against the Demographic and Health Survey (DHS) dataset across Africa. Our results indicate that integrating specific NDVI-derived features with multi-spectral data provides valuable insights for poverty analysis, emphasizing the importance of retaining intra-annual information.

Analyzing Poverty through Intra-Annual Time-Series: A Wavelet Transform Approach

TL;DR

The results indicate that integrating specific NDVI-derived features with multi-spectral data provides valuable insights for poverty analysis, emphasizing the importance of retaining intra-annual information.

Abstract

Reducing global poverty is a key objective of the Sustainable Development Goals (SDGs). Achieving this requires high-frequency, granular data to capture neighborhood-level changes, particularly in data scarce regions such as low- and middle-income countries. To fill in the data gaps, recent computer vision methods combining machine learning (ML) with earth observation (EO) data to improve poverty estimation. However, while much progress have been made, they often omit intra-annual variations, which are crucial for estimating poverty in agriculturally dependent countries. We explored the impact of integrating intra-annual NDVI information with annual multi-spectral data on model accuracy. To evaluate our method, we created a simulated dataset using Landsat imagery and nighttime light data to evaluate EO-ML methods that use intra-annual EO data. Additionally, we evaluated our method against the Demographic and Health Survey (DHS) dataset across Africa. Our results indicate that integrating specific NDVI-derived features with multi-spectral data provides valuable insights for poverty analysis, emphasizing the importance of retaining intra-annual information.

Paper Structure

This paper contains 14 sections, 7 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Results
  3. Discussion
  4. Methods
  5. Data availability
  6. Code availability
  7. Acknowledgements (not compulsory)
  8. Author contributions statement

Figures (3)

  • Figure 1: Data collections and satellite imagery used in this study: (a) Landsat-8 monthly time-series data, featuring 7 bands: red, green, blue, NIR, SWIR1, SWIR2, and Thermal. (b) Aggregated Landsat-8 yearly data, generated using a pixel-based median operator; the left images display RGB visualization, while the right images show false color visualization using NIR-SWIR1-SWIR2. (c) VIIRS monthly composite of nighttime light data, consisting of a single band showing the calibrated monthly average value. (d) VIIRS yearly median value image. (e) VIIRS image spatial average value. (f) NDVI monthly time-series data, each image containing one band showing the NDVI average for that month. (g) Feature extraction from NDVI time-series data using the wavelet method.
  • Figure 2: Data pre-processing for DWT-based feature extraction
  • Figure 3: Example tile centered at Longitude 0.1515 and Latitude 35.996 in Algeria. (a) High-resolution Google imagery providing detailed visualization of the area. (b) Landsat-8 true color (RGB) visualization. (c) VIIRS nighttime light data. (d) Minimum LFC value. (e) Maximum LFC value. (f) Mean LFC value. (g) Minimum HFC value. (h) Maximum HFC value. (i) Mean HFC value.