One-dimensional lidar signal is achieved by the simulation combining an atmospheric model and the design of parameters of imaging lidar. A two-dimensional light cross image is recovered based on measured beam widths and features of the Gauss distribution. The noised images similar to the real signals detected by imaging lidar are obtained when we add Gauss white noise with different intensities and certain intensity of average background to the original image. The good de-noising effect is obtained when we denoise the noised lidar light cross image by the two-dimensional wavelet transform method. The relative error between echo signal after denoising and original echo signal is in the range of ±12%. The extinction coefficients of aerosol are retrieved with de-noising lidar signals. Comparing extinction coefficients under the input aerosol atmospheric model with retrieved extinction coefficients of aerosol, we find that the relative error is in the range of ±15% and their variation trends are coincident, which verifies the feasibility of the proposed method using wavelet transform in the de-noising for the noised images of lidar.