Hi,
I am very new to remote sensing. I got raw images from planet scope for my area of interest through QGIS plugi. however before I calculate NDVI for my area. I need to normalised or scale the bands. I am using this formula
lower_percentile = 30
upper_percentile = 70
y1=float(0.2)
y2=float(0.5)
def scale(band): # scale values for display purposes
band = band.astype('float64')
x1 = np.log10(float(np.percentile(band(band!=0], lower_percentile)))
x2 = np.log10(float(np.percentile(bandlband!=0], upper_percentile)))
bandband==0]=np.nan
band = np.log10(band)
def equations(p):
a, b = p
return (np.log(a)-x1*b-np.log((1-y1)/y1), np.log(a)-x2*b-np.log((1-y2)/y2))
a,b = fsolve(equations, (1, 1))
band=1/(1+a*np.exp(-b*band))
return band.
Can somebody please tell why could be different method for scaling and is there any paper which can recommend for scaling the planet images?