Estimation of tree species proportions of forest compartments using ranging scatterometer

Tree species proportions of forest stands were estimated using a ranging scatterometer called HUTSCAT. The employed estimation method was a multilayer perceptron neural network with error backpropagation training algorithm. Different methods based on intensity and/or shape of measured profiles were tested. The best classification accuracy of the main tree species was about 88% and the mean error of estimation for tree species proportions was 0.26.