No, no, no, NO. You cannot state R,G,B values for such targets, "Image Analyst". The RGB colorspace is HIGHLY non-perceptually-uniform. Conversion to LAB would make it as perceptually uniform and standardized as you can get these days. And yes, you state that you use the sRGB values, but which sRGB whitepoint do you use? What observer degree? And so on. Convert your image to the appropriate LAB values and take the color space into consideration. Then compare LAB values.