Significant image enhancement technique for removal of noise in LiDaR images
A. Vijaya, M. Sundaresan · International Conference on Computing for Sustainable Global Development · 2016
This article tries to define and describe a method for removing noise and handling the edges in terms of image enhancement using LiDaR images, based on a derived model with variance and gradient coefficients. In this connection, some of the edge detection and filtering techniques are discussed as background study and the study facilitated to derive a new technique with simplest calculations of discrete functionalities. Edge detection operators are also considered since it has a vital role in terms of image enhancement. Neighborhood pixels are processed to identify the edges, signals and noises. The edges are predicted by variance of second order gradient factors and signal to noise ratio is evaluated by identifying the median value along with intensity coefficients of spectrum which produces the LiDaR image. The proposed filter concentrates on rate of image intensity and magnitude of the light which produces the image in the spatial domain. Gradient vector with wavelet transformation is instigated to remove the noise and enhance the quality of an image. Threshold value is used to evaluate the PSNR in the proposed methodology. The noise estimation parameters considered in this article are Minimum mean square error, standard deviation, correlation, variance, Mean, Median and PSNR. Least Square error has an effective role in estimating the noise over digital images. The proposed filter improves the quality of an image by detecting the edges and smoothening the signal values.