A noise is an inherent entity of the imaging technologies that tend to deteriorate the quality of processed images at all levels. At the hardware level they appear as the dark current, shot noise etc. however at the imaging side they may involve artefacts arising from interference patterns, undesired shadows, flickering etc. Eliminating such signals are challenging. Though suggesting a hardware changes and improving the imaging technologies is one way, the problem still remains. An ideal de-noising technique should know apriori various estimates of the noisy data both spatially and temporally. In the context of devising an ideal de-noisy method, I chose to estimate the limits of the intensities levels for the raw data and the edges determined by abs-Laplacian, Robert, Sobel, Prewitt and Canny’s method during day time, with incandescent lightning compared with darkness. Results show that the conventional wavelet based de-nosing can only deteriorate the edge profiles and are not useful in real time applications. abs-Laplacian still stands out as the better edge detection technique in comparison but with large bandgap. There is a negligible change in temporal distributions. A change of 12 units was observed during prolonged imaging of a static background that comprises about 4.7% of the maximum intensity. It appears that prolonged imaging has an effect of sharpening the edges or in other words a set of subsequent images would be useful in enhancing edge profiles. Keeping in view the timing constraints for real time applications, the only choice left in formulating rapid de-noising technique would be in learning the ways the noise manifest itself.