LM algorithm, the key is to treat with the estimated parameter vector function model in the field to do its linear approximation, ignore the above second-order derivative terms, which transformed into a linear least squares problem, can greatly reduce the computational load.

The key to the algorithm LM function is treated as a model to estimate parameters in its area of linear approximation, lose sight of attending more than 2 bands and translate into a linear most small 2 multiplied by issue, can greatly decrease the computation.

The LM algorithm key is estimates the parameter vector with the model function treatment to make the linear approximation in its domain, neglects two step above derivative item, thus transforms as the linearity is smallest two rides the question, may reduce the computation load greatly.

LM algorithm is the key to using the model functions within its area, treat the reassess parameter vector linear approximation, ignored second order derivative of the above, to convert to linear least squares problem, can greatly reduce the amount of calculation.

LM to treat estimate the parameter vector similar to with model function in their field linearly in key of algorithm, having finished neglecting derivative one of more than two steps, thus turn into linear minimum two questions, can greatly reduce the calculating amount.