A platform for crowdsourcing the creation of representative, accurate landcover maps
| Autor: | Stephanie R. Debats, Lyndon Estes, Kelly K. Caylor, Jonathan Choi, R. Zempleni, W. Guthe, G. Ragazzo, Tom Evans, D. McRitchie, D. Luo |
|---|---|
| Rok vydání: | 2015 |
| Předmět: |
0106 biological sciences
Environmental Engineering 010504 meteorology & atmospheric sciences Cover (telecommunications) business.industry Computer science Ecological Modeling media_common.quotation_subject Crowdsourcing Machine learning computer.software_genre 010603 evolutionary biology 01 natural sciences Representativeness heuristic Job market Pattern recognition (psychology) Quality (business) Artificial intelligence business computer Software Representative sampling 0105 earth and related environmental sciences Geometric data analysis media_common |
| DOI: | 10.7287/peerj.preprints.1030v2 |
| Popis: | Accurate landcover maps are fundamental to understanding socio-economic and environmental patterns and processes, but existing datasets contain substantial errors. Crowdsourcing map creation may substantially improve accuracy, particularly for discrete cover types, but the quality and representativeness of crowdsourced data is hard to verify. We present an open-sourced platform, DIYlandcover, that serves representative samples of high resolution imagery to an online job market, where workers delineate individual landcover features of interest. Worker mapping skill is frequently assessed, providing estimates of overall map accuracy and a basis for performance-based payments. A trial of DIYlandcover showed that novice workers delineated South African cropland with 91% accuracy, exceeding the accuracy of current generation global landcover products, while capturing important geometric data. A scaling-up assessment suggests the possibility of developing an Africa-wide vector-based dataset of croplands for $2-3 million within 1.2-3.8 years. DIYlandcover can be readily adapted to map other discrete cover types. A crowdsourcing platform that uses human pattern recognition skill to create accurate, geometrically rich landcover maps.Primary features: representative sampling, worker-specific accuracy assessment, and connection to online job markets.A cropland mapping trial showed 91% accuracy, and potential to make an Africa-wide map for $2-3 million within 1.2-3.8 years. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|