Mapping dolomitization through a co-regionalization of simulated field and image-derived reflectance spectra: a proof-of-concept study
Field spectral measurements are often conducted to guide the processing and analysis of remote sensing data. However in most cases they serve a descriptive purpose by characterizing spectrally the materials of interest rather than an analytical purpose by integrating quantitatively these data with the image data (e.g., through regression or co-regionalization). The few previous studies that have been conducted in the integration of field and image data were in agriculture applications and vegetation studies where the two variables of interest (i.e., the field and image measurement) were the same namely leaf area index. The work presented in this Letter is a proof-of-concept study to integrate field observations on the degree of dolomitization with radiance values from one single band out of a GER imaging spectrometer data set. 'Field measurements' are simulated from GER data to generate a test set and a validation set. Direct interpolation of the 'field measurements' using ordinary kriging, regression using the best linear fit between the 'field measurements' and the corresponding radiance values, and co-kriging exploiting the spatial cross correlation between the two variables are used to produce estimates of dolomitization. The results are evaluated using the validation set demonstrating the value of incorporating spatial dependencies. From the study, criteria are derived to be followed in implementation in field practice.