Messages

16

General / Re: Two questions - Number of Markers and Alignment

« on: December 22, 2017, 04:46:17 PM »

Hey ljurjevic,

You can set your coordinate system to whichever one you want, I believe as default it is doing WGS84 .

17

General / Re: cancelling batch process

« on: December 21, 2017, 10:19:50 PM »

Hey Weissspice,

As Alexey mentioned you would only lose the current step. This happened to me multiple times and I haven't lost the entire project. Once you cancel it however make sure  you save it instantly because if you don't and close your project you will lose it all.

18

General / Re: Camera calibration in Agisoft Photoscan Pro v1.4.0

« on: December 21, 2017, 10:14:27 PM »

Hey Nikos.

For NDVI you don't have to do anything else in in order to normalize the data as long you use the following formula:
(B4-B2)/((B4+B2).
 This formula is going apply normalization for NDVI within the formula. However for the other bands such as Green, Red, Red Edge and NIR, you have to normalize the data.

When you apply "B1" formula in the raster calculator under Palette tab, it needs to be refreshed. Once it is refreshed the minimum and the maximum value will appear. However those values aren't the normalized and they need to be normalized.

Once you get the minimum and the maximum value then you need to divide the maximum value by the band. (Example: B1/45000)

Once you normalize the band  it needs to be refreshed using blue arrows. Once it is refreshed the reflectance values would appear between 0 and 1.

**This is under the assumption that you are going to disable the calibration targets.

I am not sure what you mean by the "barrel effect", but if your refectance map looks too bright or too dark, you may definitely end up with wrong reflectance values. 

19

General / Re: Two questions - Number of Markers and Alignment

« on: December 21, 2017, 07:49:36 PM »

Hey Andreg,

If you are going to have 11 chunks, they need to be treated as separate projects so I would do 10 markers/ GCP's for each chunk. (You could probably get away with 6-7 per chunk). I am not sure about the orthophoto since I have never exported an orthophoto before however for the orthomosaics I would use High Alignment (preferably), if not Medium  Alignment.  Moreover you can either generate them using the DEM, Mesh, Sparse Cloud or a Dense Cloud. I would use the DEM as my first option. However the advantage of using Mesh over DEM is that sometimes DEM is not generated accurately and if you are creating your orthophoto or  your orthomosaic based on Mesh, your orthomosaic wouldn't be affected even if your DEM is not accurate.

20

General / Re: Identifying bands from multispectral images in folder structure

« on: December 21, 2017, 07:31:53 PM »

Hey Robert,

I am guessing you are using the MicaSense Red Edge camera since it has 5 bands. Are you trying to export the reflectance maps manually or are you using a script to export them? If you are not specifying individual bands while exporting, it would export the RGB output as default. You need to specify each bands separately if you want 5 different reflectance outputs, it would correspond to Band 1, Band 2, Band 3, Band 4 and Band 5.   

It would be something like this for the Band 2 and similar for the other bands as well:
chunk.raster_transform.formula = ["(B2)"]
chunk.raster_transform.calibrateRange()
chunk.exportOrthomosaic("Path"mosaic_.tif", image_format = PhotoScan.ImageFormatTIFF, raster_transform = PhotoScan.RasterTransformValue, write_world=True, write_alpha=False, tiff_big=False, white_background=True)

Or alternatively you can do it manually by using the Raster Calculator and specify each band then export and repeat it for each band. Just type B1 under Output Bands then hit apply and export and do it for the other bands as well.
I have spend a lot of time trying to figure that out but it seems to be working now. Please let me know if you need more info. FYI what version of Agisoft Photoscan are you using?
 

21

General / Re: Camera calibration in Agisoft Photoscan Pro v1.4.0

« on: December 21, 2017, 06:57:10 PM »

Hey Nikos,

I had the same issue with the version 1.4. I believe that version is still un-stable therefore it is not incorporating the calibration targets as it should be. You may have noticed that your reflectance values may really low (something like 0.001 instead of 0.1 for the low range) and that is probably the reason why you are ending up with a white reflectance map. Since there is no variance between the values it is only incorporating the white color. The other thing you could do would be disabling the calibration targets. Then you would have to normalize the data so that you would end up with the reflectance values between 0 and 1 for the Green, Red, Red Edge and the NIR bands and between -1 and 1 for the NDVI. I have been doing that for a while, and it seems to be working well for me.