How to Extract BPL and DPL Syntax_Extract HowTo Extract BL HowTo Extract Soil Line |
- The -deglint argument is disabled by default: no deglinting upon extraction of calibration data
- except in the AutoCAL process if a glint.shp shapefile is available
- The Smooth and Smart options are used as/if enabled. Smoothing is advised, but be wise.
- Some shallow areas may need to be avoided as follows:
- mSE and mBPL : this may be achieved through the mBPL parameters in the -Extract argument.
- Any pixel with a value of mSE<mBPL is skipped upon the extraction of BPL data
- Any pixel with a value of mSE>=mBPL and mBPL<mSOIL is considered in the extraction of BPL data.
- for example, if you set BPL at 2, and mSOIL at 21 all pixels with mSE>=2 and mSE<=20 are considered, and all pixel with mSE<2 or mSE>=21 are bypassed
- Mask : alternatively, by enabling the Mask_N option in in the -Extract... argument, all pixels with mSE in the range N-21 are considered for BPL, while all others are bypassed
- for example, with Mask_3: only pixels with mSE==3 through to mSE=20 are considered, all others are bypasseded.
|
BPL Brightest Pixels Line For each pair of bands, the BPL is made of all pixels that are brightest in band i for each DN in band j - The BPL may be seen as the outer envelope of the shallow pixels exponential bulge
- It intends to sample the brightest type of bottom substrate that is present across the whole range of bottom detection. tmnov_how_nice_x12
- LsNirMaxBPL : in order to prevent selecting heavily glinted pixels in pairs which do not involve the NIR band, a NIR threshold is applied through the LsNirMaxBPL parameter of the -Extract argument
- A value of LsNirMaxBPL=5 above Lsw niris usually recommended
- this is to ensure that pixels which are badly damaged by glint effect are discarded
|
DPL Darkest Pixels Line For each pair of bands, the DPL is made of all pixels that are darkest in band i for each DN in band j - For an image with dryland exposure, most Darkest Pixels are at near null depth, from bright to dark
- The DPL is extremely usefull for the determination of spectral La
- For an image without any dryland exposure like an atoll submerged at high tide, the DPL, from bright to dark, is likely to represent pixels at a certain depth:
- The DPL is also averaged into SL_DPL: see tarawa-subset-m0cal
|
Extraction window - A rectangular extraction window may be specified if desired at the end of the -Extract argument , so that the extraction of calibration data shall be limited to pixels inside that window extraction is strictly limited to shallow pixels with mSE=Mask
- -extract/v/NIRband4/NIRmax5/mBPL_2/mSOIL_21/Mask_4 (this overrides the value specified for mBPL)
- Mask_2: ==> 4SM writes and shall read text file negril85.m2.cal
- Mask_3: ==> 4SM writes and shall read text file negril85.m3.cal
- Mask_N: ==> 4SM writes and shall read text file negril85.mN.cal
Upon calibration : don't get mixed up if Mask_N is specified (1>=N<mSOIL) | 4SM shall read negril85.mN.cal | else | 4SM shall read negril85.m0.cal | |
See How to read calibration data textfile |
Look at the location of BPL pixels in the tarawa-subset image They are mapped in channels 15, 16 and 17 by the Extraction process see tarawa-subset  Here is a color composite of these three pairs of BPL |