slcOLI_20180118

Investigating a LUT approach using Landsat 8 data

home

home
scene LC80340432018018LGN00, January_18th_2018

Data and Deglinting

TOA TCC: raw image logarithmic enhancement	TOA TCC deglinted image logarithmic enhancement
Glint regressions using SWIR1 R2 values are fairly high 15 m resampling does not increase noise
Location of Profile_RED backdrop is a TCC of deglinted data Profile_Yellow is also shown dLsw shapefile is also shown

Optical calibration for the whole image

Calibration diagram for the whole image for bands Blue, Green, Red and NIR	Calibration diagram for the whole image for bands Blue, Green, PAN, NIR CoefKPan=1.00
K_BLUE/K_GREEN~=0.94 Jerlov water type Coastal_1 nice linear BPL line: denotes homogeneous brightest bottom in homogeneous waters	This calibration suits the whole image

Calibration for bands Blue, Green, PAN, NIR for San Lorenzo Channel , under mask_4	Homogeneous The SL Channel , under maskSE=4 does not exhibit anything particular 0-25m water type Coastal_1 Kblue/Kgreen~=0.94

Ready for modeling

True Color Composite of BOAWCC radiances scaled 0-250 B=Coastal band in channel 34 G=Green band in channel 35 R=Blue band in channel 36	Retrieved depth see legend
Maximum BOAWCC reflectance (0-1) for this scene -RM/0.133/0.175/0.227/0.265/0.315 /0.441/0.425 To convert BOAWCC radiances into calibrated reflectances CoastalRefectance=CoastalBOAWCC0.133/250 BlueRefectance = BlueBOAWCC0.175/250 GreenRefectance = GreenBOAWCC*0.227/250	Maximum shallow bottom reflectance (0-1) for coral sands -RM/0.310/0.380/0.440/0.550/0.560/0.670/0.700 Coral sand is approx twice brighter than brightest sands in this scene.
Average bottom brightness see legend for image B	15 m GSD, smart-smoothed not_PANsharpened: sharpening becomes redundant/pointless, thanks to: improved co-registration of the PANchromatic band with the MULTI bands (might still be improved further) combined with 15 m resampling of MULTI bands to match the PANchro GSD smart-smoothing then becomes desirable/necessary

Mid_winter time on January 18th

Computed depth along Profile Yellow WCC is also shown for the Coastal, Blue and Green bands	Profile Yellow location Backdrop is True Color Composite of BOAWCC radiances scaled 0-250 B=Coastal band in channel 34 G=Green band in channel 35 R=Blue band in channel 36
Computed depth along Profile Yellow compares the Green and the PAN solutions	Z3 vs Z4 GREEN solution versus PAN solution
Dormant dormant: waters and bottom substrates appear to be dormant Reflectance of brightest bottoms unchanged, while most bottoms would be darker than usual in winter dark and dull: such tight correlation of Green vs PAN solutions is only possible provided there is a lack of contrast in bottom substrate spectral signatures Green and PAN solutions yield approx the same depth over the whole depth range, except at very shallow depth (<3-4 m). This is mid-winter time in spite of Coastal water type, calibration is straightforward	? PANsharpening of MULTI bands ? no obvious benefit; adds some noise PANsharpened.............R2=0.97 RMSE=0.56 m NOTsharpened.............R2=0.99 RMSE=0.45 m Over very shallow bottoms GREEN solution yields under-estimated depth over very shallow bottoms PAN solution does not, apparently