Search

Relative radiometric normalization is a prerequisite for multi‐temporal remotely sensed data analysis. Dark object subtraction (DOS) and pseudo‐invariant feature (PIF) methods are two commonly employed scene‐based methods for relative radiometric normalization. Both these methods have known limitations. The authors describe a soil line transformation method for multi‐date relative radiometric normalization, which circumvents the limitations of the DOS and PIF methods.

Data from seven validation campaigns in the coastal waters of the Bay of Bengal (BOB) were used to develop a regional SPM (suspended particulate matter) algorithm. The in situ data sets for this algorithm are obtained from 86 stations with optical and 68 stations with total SPM measurements encompassing SPM concentrations between 10 and 189 mg l −1 ,w here most of the observations are in shallow (case 2, average depth ∼45 m) waters and a limited number of samples in open case 1 waters. From simple statistical analyses we found a close relationship between SPM concentration and diffuse attenuation coefficient at 555 nm (K 555) in the BOB coastal waters. The linear regression to the fit has ac oeff icient of determination (r 2 ) 0.96 with a standard error of estimates (σ )12. 5m g l −1 for the above SPM range. The algorithm relating K 555 to [L wn443/L wn670] has been evaluated through a regression analysis of radiometric profiles in the BOB. However, the new SPM algorithm overestimates in case 1 waters (SPM range 0.05–25. 0m g l −1 ), wher ea spectra lr eflectance ratio algorithm (Tassan 1994 Appl. Opt. 33 2369–78) appears to produce better results. An integration of both the approaches performs better in generating the routine IRS-P4 OCM (ocean colour monitor) SPM mapped product.

With a spring‐loaded platform mounted on a precision minidrilling machine, it is possible to rapidly and inexpensively drill sharp sample holes down to about 100 μm in hardened metal gaskets for diamond anvil cell experiments.