Application of principal component analysis to lidar data filtering and analysis

Principal Component Analysis (PCA) has proven to be a valuable tool for remote sensing data compression, pattern recognition, and for filtering out measurement noise. In this pa per, we present preliminary results on the application of PCA technique to reduce random noise present in lidar observations. Typically, the SNR at a given range can be improved either by increasing the integration time of the meas urements or by applying spatia l averaging. Th is procedure, however, improves the SNR at the expense of the instrument’s temporal and spatial resolution. The number of range bins needed to characterize backscatter features is far less than the number of components needed to characterize the distribution of these features in the atmosphere. The higher-order PCA components, which mainly serve to characterize noise, can be eliminated along with the noise that they characterize. The results of PCA noise filtering of lidar observations strongly depend on the variability of aerosol plum es. To avoid loss of information in the presence of highly variable aerosol plumes, it is necessary to use a conservative number of principal components higher then optimum for maximum noise reduction. Nevertheless, noise reduction factors of 2-8, depending on the lidar range and atmospheric variability, can still be achieved. Keywords: Remote sensing, Principal Component Analysis, lidar, aerosols, pollution, air quality