An evaluation of satellite remote sensing for crop area estimation in the west bank, Palestine

Abstract

This thesis investigates the use of field and satellite data for crop area estimation in the northern part of the West Bank, Palestine. The satellite data were obtained by the SPOT HRV on 19 May 1994. The satellite data were geometrically corrected to the Palestine Grid using 1: 50,000 Israeli topographic maps. The study investigated the ability of SPOT HRV data to produce accurate crop area estimation of the northern part of the West Bank that is characterised with small field sizes and complex physical environment. A land cover classification scheme appropriate to the study area was designed. Twenty-three land cover classes were produced from the SPOT HRV classification. Land cover classes were developed to produce thematic land use classes. The classification accuracy obtained from SPOT HRV image classification was 81%. Classification results were assessed by using the known land use information obtained from the field during the training stage and the field sampling survey. The study area was divided into five strata and the field survey was conducted by applying a stratified random sampling methodology. Seventy three 1 km(^2) sample units were randomly chosen and surveyed by the author using maps, aerial photographs, satellite photographs, a questionnaire, camera photographs, and sketches. The field area measurements were taken and the final hectarage estimates were obtained for each crop type. The SPOT HRV and the field data were combined in regression analysis using a double sampling method and a hectarage estimate was produced for each crop in the study area. The results obtained showed that the regression estimator was more efficient than the field estimator and a gain in precision was achieved. The results were analysed on stratum and crop type basis. Remote sensing and thematic agricultural perspectives were used in the analysis. Results of the study suggest that it is possible to improve image classification accuracy by using better spatial and spectral resolution imagery and the integration of remote sensing data with agricultural data using the Geographical Information Systems (GIS).