Topics

1

Camera Calibration / How does Agisoft calculate the reflectance values for the calibration panel

« on: April 15, 2021, 09:29:09 PM »

Hi all,

Didn't find this mentioned earlier so made a new topic.
I am curious on how the reflectance values are calculated in agisoft based on the .csv file for the calibration detection module.
So for the following bands with these specific wavelengths, how does Agisoft reach the reflectance value per band?
Blue (B): 475 nm ± 32 nm;
Green (G): 560 nm ± 27 nm;
Red (R): 668 nm ± 14 nm;
Red edge (RE): 717 nm ± 12 nm;
Near-infrared (NIR): 842 nm ± 57 nm

when I am manually trying to calculate this with a script like below, I get these values
dtype: float64, reflectance 0.691253
dtype: float64, reflectance 0.702969
dtype: float64, reflectance 0.69715
dtype: float64, reflectance 0.688833
dtype: float64, reflectance 0.43085

import pandas as pd
reflectance_factors_excel_path = r"RP02-1603082-SC.xlsx"
df = pd.read_excel(reflectance_factors_excel_path)
df.info()
df2 = refl_df.set_index('wavelength')
middle_nm = 842
nm = 57
nm = int(nm/2)
first = int(middle_nm - (nm))
last = int(middle_nm + (nm))
row_list = list(range(first,last))
df2.loc[row_list].mean()

But agisoft gives different values, so I would like to know the details on how we get from csv to the panel calibration reflectance value if possible!

Kind regards,
Christina

2

Python and Java API / Transform DEM

« on: October 20, 2020, 08:08:34 PM »

See attachment: Is it possible to access this tool through the Python API?
Or can we change the cell size in another way?

In the GUI when building the DEM the resolution seems to be determined based on the data input. Is this also the case if we are working with python?

3

General / Best practises (flight planning, processing settings) for tree reconstruction

« on: February 27, 2020, 07:39:07 PM »

Hi all,

I would like to initiate a discussion for best practices for accurate tree reconstruction on orchards, specially when trees are small and have a sparse upper canopy during blossom season (apples, pears etc)!
Currently, we aim for very sharp rgb orthomosaics in order to detect the amount of blossom per tree.

Flight parameters: so far we use 80% overlap,  double grid flight plan, sensor looking on nadir and flying altitude of around 40m.

Based on your experiences what are the optimum flight plan settings:
- Overlap:
- Sensor angle:
- Speed:
- Field conditions (sunny, not windy etc): (of course we cannot control that but just for reference)

For agisoft processing settings we generate a mesh-based rgb orthomosaic since we noticed that a dsm-based rgb orthomosaic suffers more from distortions. But even a mesh-based orthomosaic sometimes has artifacts (see image attached).
Based on your experiences what would you recommend settings-wise for a sharp looking without distortions rgb orthomosaic tif?
Perhaps some blending mode suffers less or has a better seam-line estimation?

Thank you in advance!
Christina

4

General / Export the dense point cloud as .las coordinates issue

« on: February 14, 2020, 12:14:49 PM »

Hi all,

We are processing a DJI P4 (not rtk) normal rgb images, project with default WGS84 coordinates and later on we export the DEM & Orthomosaic in UTM 31N zone.

At the end of the project we also wish to export the dense point cloud as .las with the default WGS84 without specifying UTM zone but it goes wrong (see image attached).

Only if we export it also to WGS84/UTM 31N zone, that we specified in the orthomosaic and Dem then the export goes correct.

However, what is strange is that this is a bit random as for some other projects the default wgs84 export seems to work just fine.
Is there is something we miss here? Perhaps something with coordinates in the raw data?

Did anyone experience something similar?
Below screenshot of how the faulty one looks in CloudCompare, for confidenciality reasons I cannot share the data or a schreenshot of the normal point cloud.

Thank you in advance!

5

Bug Reports / Altum GSD issues

« on: July 18, 2019, 03:33:58 PM »

Not a 100% sure if this is a bug but according to Micasense, the Altum camera has a ground resolution of:

"5.2 cm per pixel (per EO band) at 120 m (~400 ft) AGL and 81cm per pixel (thermal) at 120m"

When processing Altum imagery the actual resolution that is automatically calculated does not seem to match and is way to low/coarse.

In the report of Agisoft we can see that the altitude is around 80 m and the ground resolution used for the OrthoMosaic is approx 10.5 cm which is way to coarse according to the MicaSense datasheet. The resolution used for export was automatically defined by Agisoft.

We are processing the data following this guide: https://agisoft.freshdesk.com/support/solutions/articles/31000148381-micasense-altum-processing-workflow-including-reflectance-calibration-in-agisoft-metashape-professi

Is this a bug or are we doing something wrong? Any ideas why the ground resolution could be lower?

Thanks in advance!

6

Python and Java API / Run "Set Brightness" tool to estimate brightness from python

« on: July 18, 2019, 02:50:25 PM »

Is it possible to run the "set brightness" in the tools menu from Python to automatically estimate the image brightness (and contrast)? I could not find it in the Python API reference.

Thanks in advance.

7

Python and Java API / Raster transform python multiple bands

« on: June 12, 2019, 12:55:08 PM »

Hi,

I need to export a multispectral image as reflectance with python. I am able to use a single band expression in the raster calculator in python but in this case I would like to export multiple bands.

It is unclear how to add more output bands to the expression in python. Is this possible? It is working correctly in the GUI see attached screenshot.

Thanks in advance