Data fusion using subtractive resolution merge (SRM) is limited because it currently requires fixed spectral band weights predetermined for particular sensors. This is problematic because there is an increasing availability of new and emerging sensors that have no predetermined band weights. There is also a need for fusion between sensors, which potentially requires a large number of sensor combinations and band weight calculations. This article demonstrates how the least sum of minimum absolute deviation (LAD, least absolute deviation) and ordinary least squares (OLS) regressions can calculate band weights for application in the SRM technique using QuickBird satellite and Vexcel aerial images. Both methods were effective in improving image details. The results of LAD and OLS are shown using qualitative and quantitative metrics and through unsupervised classification of freshwater habitat. OLS and LAD produce similar results; however, OLS is computationally simpler and easier to automate. The ability of the user to calculate their own scene specific band weights eliminates the dependence on predetermined sensor band weights. This research concludes that OLS band weight calculations should be integrated into the SRM technique to diversify its application.