Abstract A single pixel of a Landsat image has seven channels receiving 0. 1 to 10 microns of radiation from the ground within a 20x20 meter footprint. In principle, the pattern of seven values can be utilized to identify ground sources within the pixel footprint by using methodologies called spectral blind demixing of unknown sources when the reflectance matrixwij for the ith object and the jth band is either partially unknown or difficult to measure in theouter space.We can assume a linear mixing model for a long distance propagation of weak sources.The satellite measurements of seven bands in a single pixel U1 = W'V are obtained from Vobjects in the footprint of the pixel, of which both the set of objects V and the reflectance matrixWij are unknown. We postulate the constraint maximum entropy with Lagarangian multipliersto incorporate the measurement to be the Hopfield-like energy function . Use is made of thetraditional Lagrangian variation method to estimate the unknown sub-pixel object V in a