Regression Analysis

The basic idea of the regression analysis is: firstly, establish a regression model between in-situ forest parameter measurements and LiDAR statistics at the plot level, and then use the obtained model to predict forest parameters from LiDAR statistics. Studies have shown that the regression methods can estimate forest parameters that cannot directly derive from LiDAR point cloud, such as aboveground biomass, and achieve a relative high accuracy without saturation (Popescu et al., 2004). However, building regression models requires a number of field measurement data. Please refer to Forestry FAQ for the selection of field measurements.

• Sample Data and Independent Variables

• Linear Regression

• Support Vector Machine

• Fast Artificial Neural Network

• Random Forest Regression

• Run Existing Regression Model

    @inproceedings{
        author={Popescu S C and Wynne R H},
        title={Seeing the trees in the forest: Using lidar and multispectral data fusion with local filtering and variable window size for estimating tree height},
        booktitle={Photogrammetric Engineering and Remote Sensing, 70(5): 589-604},
        year={2004}
    }