Differential absorption LIDAR signal denoising using empirical mode decomposition technique

Differential Absorption Lidar (DIAL) technique is a potential method for the remote detection of hazardous chemicals in the atmosphere. These hazardous chemicals can be due to industrial pollution or may be intentionally released by the terrorist groups or military forces of the enemy country to endanger both the military's personnel and the civilian population's lives. DIAL technique may be used for probing such chemicals from far-off distances of several km ranges and generating an early warning for the response teams. The output of the DIAL system normally consists of three parameters viz. name/class of hazardous chemical detected; its location in terms of distance and the concentration. The maximum standoff distance capability for any Lidar system depends on the signal to noise ratio which is governed by the parameters like atmospheric conditions, Lidar subsystem specifications, noises, etc. SNR is often limited by several noises embedded in the signal from various sources. Due to the presence of noises in the signal, the errors are introduced in the concentration estimation of chemicals from Lidar signal. The methods for improvement of SNR of lidar signal has been often limited by application of conventional denoising techniques like multi-pulse temporal averaging and spatial averaging and further requires nonlinear techniques for noise reduction due to nonlinear behavior of lidar signals. In the present work, Empirical Mode Decomposition (EMD) technique has been implemented on the Lidar signal from Differential Absorption Lidar system. The signal has been denoised and improved SNR is compared with that achieved from temporal averaging and spatial averaging. It was observed that the EMD technique is a better technique as compared to other conventional techniques like multi-pulse temporal averaging and spatial averaging for denoising the signal and increasing the Lidar SNR. It is seen that SNR can be improved 4–5 times the original SNR using EMD technique.