I need to know if it is scientifically valid to fuse/pan-sharpen Landsat 8 surface reflectance products with pan band of that respective band? Landsat reflectance product details can be found at here. It needs to be mentioned that one needs to order surface reflectance product separately to get this product. This product contains only 7 band (30m) not IR and Pan band. So, again, my question is it valid to fuse 7 bands(30m) of surface reflectance product with normal(not surface reflectance) pan band(15m). I want to use this pansharpened image for segmentation and following land cover mapping.So I need to know that is there any established practice of this type of pan sharpening in academia with reference, if yes please cite.
Fundamentally the question here is "what does 'scientifically valid' mean". If you are looking to do spectral modelling on the data, then the answer is possibly different than if you are looking at doing classification / image segmentation. Pansharpening (depending on the method) is simply going to change the range of the values a fairly small amount and shouldn't put your reflectance values outside the realm of possibility.
All in all, it depends a lot on what application you are going to be using the data for. Furthermore, the impact of pansharpening may also be worth documenting as a partial side result in whatever study you are performing. The result may be that it doesn't add anything, except four times as many pixels, meaning four times as long a processing time, which in some cases is a showstopper.
Edit: My database of articles on this topic is not huge, but I have these two where pansharpend data is used (with reasonable results) for image segmentation:
Shackelford, A. K., & Davis, C. H. (2003). A combined fuzzy pixel-based and object-based approach for classification of high-resolution multispectral data over urban areas. IEEE Transactions on Geoscience and Remote Sensing, 41(10), 2354–2364. http://doi.org/10.1109/TGRS.2003.815972
Fernández, I., Aguilar, F. J., Aguilar, M. A., & Álvarez, M. F. (2014). Influence of Data Source and Training Size on Impervious Surface Areas Classification Using VHR Satellite and Aerial Imagery Through an Object-Based Approach. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 7(12), 4681–4691.
First of all - unless you REALLY know what you're doing and what you're experimenting with - you can not correctly convert PAN from DN to TOA reflectance. This data is made solely for the purpose of visual enhancement; and no spectral information is supposed to be derived from it.
TOA reflectance values are a re-scale from the 16bit data type as stated by USGS. Which means that you can use the PAN band directly as input with the multispectral TOA reflectance data. Especially since most - if not all - of the Pan-sharpening algorithms start with some sort of data normalization.
Another thing you can do - just to put your mind at ease - is to take two sample data (level 2 & level 1); apply pan-sharpening on the two, and do a spectral and spatial evaluations on both results.
P.S: Concerning the theme of your project
Last year, I worked on a project concerning the Evaluation of Pan-Sharpening effects on image classification, where the input data were Quickbird and Landsat 8 satellite imagery. Multiple algorithms and approaches were tested. And the results were very interesting. We have not yet gotten around to publishing the article so I can not disclose most of the things we did. But one thing I can say is: to try and use a combination of the original data (full bands) and Segmented pan-sharpened imagery. As most of the experiments done on Landsat data showed that the overall accuracy and Kappa coefficient dropped down comparing to the classification of the original data.