Temporal analysis of multispectral video/satellite imagery for the detection and monitoring of salinity on agricultural lands.

摘要

A study was conducted on an area of known salinity problems to evaluate the use of multispectral video and satellite imagery for the delineation of salinity problems on agricultural lands. Multispectral video and satellite imagery along with airborne and ground radiometric data were used in a carefully monitored experiment to detect salinity stress in agricultural fields.;A model based on minimal ground truth data using multispectral and thermal imagery was developed to detect salinity stress on agricultural plant canopies and soils. The model used visible and near-infrared reflectance and vegetation indices along with thermal imagery from the different surfaces to detect stress on crops. The output of the model when compared to an extensively ground-truthed image closely matched the ground truth data.;The model applied to a satellite image failed to differentiate between salinity and non-salinity-related stress.;The use of multitemporal, multispectral satellite imagery for the detection and mapping of salt-affected agricultural lands with minimal ground information was also tested in this study. The degree of stress affecting an agricultural field was related to the canopy uniformity in individual fields. The methodology successfully determined areas of stress on three multidate satellite images of the same area. Areas that showed non-uniformities over several years were considered to be salt affected. Other non-uniform areas that showed up in only one of the multitemporal images were judged to be related to other stress factors. The temporal method alone could not accurately differentiate between salinity and non-salinity-related canopy stress; however, a salinity range was determined.;To improve the multitemporal detection of salinity, an integrated geographic information system study was conducted that incorporated layers of topography, soil drainage properties, and Soil Conservation Services soil salinity survey data. The geographic information systems layers used in an integrated model along with the 1995 temporal salinity image differentiated between salinity and non-salinity-related stress.