Underwater videos are made with two red laser dots of fixed separation distance and visible in captured frame. Red laser dots in captured frame are used for distance resolution estimation in order to associate statistics with area. This paper proposes a method for automatic detection of the dots using wavelet sub-band template matching and laser dot tracking. Segmentation based on color information and 8-connectivity are used as preprocessing steps to improve detection accuracy and optimize computational cycles of template matching stage. Template matching in the Stationary Wavelet Transform (SWT) domain is used for further isolating red laser dots. Horizontal sub-band at 3rd level decomposition using Haar basis function is used to eliminate unwanted details in template matching. Finally, a distance error minimization with constrained feature is used in detecting the red laser dots. This tracking step minimizes the training requirement of template matching stage whilst improving detection rate in the undersea video.