After testing with the previously given dataset, I collected the channel values again from the original image itself, I found the channel values to be quiet different from that of the previous dataset. Actually, the values I extracted matches with the color of the images. (verified with visual inspection). 
I re-plotted the new data \ref{240772}.