تکنیک کاهش سر و صدا برای سیگنال لیودار اکو براساس تجزیه و تحلیل مولفه مستقل پالس تطبیقی ​​مستقل

Speckle noise has always been a particularly tricky problem in improving the ranging capability and accuracy of Lidar system especially in harsh environment. Currently, effective speckle de-noising techniques are extremely scarce and should be further developed. In this study, a speckle noise reduction technique has been proposed based on independent component analysis (ICA). Since normally few changes happen in the shape of laser pulse itself, the authors employed the laser source as a reference pulse and executed the ICA decomposition to find the optimal matching position. In order to achieve the self-adaptability of algorithm, local Mean Square Error (MSE) has been defined as an appropriate criterion for investigating the iteration results. The obtained experimental results demonstrated that the self-adaptive pulse-matching ICA (PM-ICA) method could effectively decrease the speckle noise and recover the useful Lidar echo signal component with high quality. Especially, the proposed method achieves 4 dB more improvement of signal-to-noise ratio (SNR) than a traditional homomorphic wavelet method.