understanding data

[philippetournevase-spec]

I everybody and happy new year from France

i continue to work with ping mapper to improve the SI sonograms given by my humminbird and i have questions.

ping mapper not concerned but if somebody know ... for example, an humminbird record gives :
"B001_ds_highfreq_meta.csv" 3165 rows and an image "xxxxx.jpg" 3165 x 2013 px corresponding to Down Scan
"B002_ss_port_meta.csv" 2814 rows and a sonogram "yyyy.jpg" 2014 x 2103 px corresponding to side Scan ( range 2 x 40 m)
why is there the difference 3165 / 2014 ?
it looks like not all the ping emitted for down scan are used for the sidescan

an other point about the speed correcting of the sonograms in ping mapper
the process consist in expanding the x axis by increasing the x pixels
i tested by using "B001_ds_highfreq_meta.csv" where there is a collumn "trk_dist" with values from 0 to 171 m for previous example, how is it calculated i don't know
[note : there is a difference of around + 2 % by checking in using the geographical coordinates, ping mapper not concerned by this]
then in expanding x axis like this : x expanded = 2103 x 171/40 = 8990 px
the speed correct function of ping mapper gives 8296 px for the same case ie a difference of 8 %
so how is it calculated in ping mapper please ? [i'm not able to look at this in the soft]

[CameronBodine]

Happy New year to you as well! I am glad you are getting into the details.

TL;DR: Speed correction and georeferencing use different sources for calculating stretch and warping, resulting in different image sizes.

ping mapper not concerned but if somebody know ... for example, an humminbird record gives :
"B001_ds_highfreq_meta.csv" 3165 rows and an image "xxxxx.jpg" 3165 x 2013 px corresponding to Down Scan
"B002_ss_port_meta.csv" 2814 rows and a sonogram "yyyy.jpg" 2014 x 2103 px corresponding to side Scan ( range 2 x 40 m)
why is there the difference 3165 / 2014 ?
it looks like not all the ping emitted for down scan are used for the sidescan

It is unknown why there are differing number of pings between each of the sonar channels. PINGMapper exports everything that is present in the file. It is some type of issue with the hardware, I believe. There may be a difference in the ping frequency for each sonar channel. Maybe it is a tunable setting on the Humminbird. I do know there is at least a global setting for ping frequency.

an other point about the speed correcting of the sonograms in ping mapper
the process consist in expanding the x axis by increasing the x pixels
i tested by using "B001_ds_highfreq_meta.csv" where there is a collumn "trk_dist" with values from 0 to 171 m for previous example, how is it calculated i don't know

Speed correction is calculated for each chunk with doSpdCor():

PINGMapper/pingmapper/class_sonObj.py

Line 1331 in d7ec0ba

def _doSpdCor(self,

An important note is that speed correction assumes constant speed throughout a chunk. The along track distance covered for each transect is based on distance_end - distance_start. The most precise speed corrected image would use smaller values for chunk size.

trk_dist is calculated with _calcTrkDistTS() using time elapsed and speed of the vessel. The vessel speed and time elapsed are both reported by the Humminbird based on GPS:

https://github.com/CameronBodine/PINGVerter/blob/6c83cd1d91a8319f0a6d71543223a005fb8b7998/pingverter/humminbird_class.py#L1045

[note : there is a difference of around + 2 % by checking in using the geographical coordinates, ping mapper not concerned by this]
then in expanding x axis like this : x expanded = 2103 x 171/40 = 8990 px
the speed correct function of ping mapper gives 8296 px for the same case ie a difference of 8 %
so how is it calculated in ping mapper please ? [i'm not able to look at this in the soft]

Geographic positioning is reported to the nearest meter by the Humminbird, resulting in multiple pings sharing the same position despite a vessel moving (Bodine et al., 2022). For georeferencing, PINGMapper assumes the vessel is moving and reinterpolates position of every ping based on a spline fit to the original coordinates. More detail in:

Bodine, C. S., Buscombe, D., Best, R. J., Redner, J. A., & Kaeser, A. J. (2022). PING-Mapper: Open-source software for automated benthic imaging and mapping using recreation-grade sonar. Earth and Space Science, 9, e2022EA002469. https://doi.org/10.1029/2022EA002469

Please let me know if this answers your question.

[philippetournevase-spec]

i cameron

thank's this answers my questions

i found an interesting soft : https://imagej.net/ij/ and played with it, it allows particularly :

• inverting post sonograms then merging segmented sonograms (issued by ping mapper when the size of the input file is too much)
• extract all the pixels of an image with x, y and intensity value
• apply a nearly "CAG" by an "enhance local contrast" plugin giving interesting results

example, upper part enhanced

[CameronBodine]

Glad you found imagej. It's a great software. That corrected data looks very nice. You can also try out the EGN correction in pingmapper.

[philippetournevase-spec]

i camerone
i tested the best EGN correction i found : percent clip factor 0.01
but it is not convenient

[CameronBodine]

Did you mean to reopen this? If so, what are you needing?