Color balancing is an important domain in the field of photography and imaging. Its use is necessitated because of the color inconsistencies that arise due to a number of factors before and after capturing an image. Any deviation from the original color of a scene is an irregularity which is dealt with color balancing techniques. Images may deviate from their accurate representation because of different illuminant ambient conditions, non-linear behavior of the camera sensors, the conversion of file format from a wider color gamut of raw camera format to a file format with a narrower color gamut and so on. Many approaches exist to correct the color inconsistencies. One of the basic techniques is to do a histogram equalization to increase the contrast in an image by utilizing the whole dynamic range of the brightness values. To remove color casts introduced due to false illuminant selection at the time of image capture many white balancing techniques exist. The white balancing can be employed before image capture right in the camera using hardware filters with dials to set illuminant conditions in the scene. A lot of research has been done regarding the effectiveness of white balancing after image capture. The choice of color space and the file format is quite important to consider before white balancing. Another side to color balancing is color transfer whereby the image statistics of one image are transferred to another image. Histogram matching is quite widely used to match the histogram of a source image to that of a target. Other statistics for color transfer are to match the mean and standard deviation of a source image to a target image. These two approaches for color transfer are analyzed and tested in this thesis on images displaying the same scene but with different color casts. Color transfer matching the means and standard deviations is selected because of its superior color balancing and ease of implementation. While a lot of color balancing work has been done in 2D images, no significant work is done in the 3D domain. There exist 3D scanners which scan a scene to build its 3D model. The 3D equivalent of the 2D pixel is a scan point which is obtained by reflecting a laser beam from a point in a scene. Hundreds of thousands of such points make up a single scan which displays the scene that was in the view of the 3D scanner. Because a single scan cannot capture scene behind obstructions or the scenes out of the scanners’ range, more than one scans are undertaken from different positions and time. More than one scans grouped together make up a data structure called a point cloud. Due to these changes in position and time, luminance conditions alter. As a result the scan points from different scans representing the same scene show a considerably different color cast. Color balancing by matching the means and standard deviation is applied on the point cloud. The color inconsistencies such as sharp color gradients between points of different scans, the presence of stray color streaks from one scan into another are greatly reduced. The results are quite appealing as the resulting point clouds show a smooth gradient between different scans.