[Qgis-user] (no subject)

[Loke Jian Munn]

Hi Kirk,
Thanks a lot for the input. It very helpful.

Thank you.

Loke Jian Munn
Information Technology (IT) Department
[cid:image001.png at 01D9A5B2.64B64990]
KUALA LUMPUR KEPONG BERHAD
Wisma Taiko,
No. 1, Jalan S.P. Seenivasagam,
30000 Ipoh,Perak Darul Ridzuan,Malaysia
T    :  +605 240 8000 (ext. 2443)
[A close up of a sign    Description automatically generated]<https://clicktime.symantec.com/15tpDKjABDCCmu4drBEZD?h=4-LzQeEUOcKmp7sCwswGafke694l_dZCDQJ4Rb8N8bc=&u=http://www.klk.com.my/>[A close up of a sign    Description automatically generated]<https://clicktime.symantec.com/15tpJ9vSdpsoBqtZPjdhq?h=bX-LFOj0HrSJV_t-3L_WhMP3i5PRx5AtNBaHlI02C7k=&u=http://www.linkedin.com/company/klkberhad>

From: kirk <kirk at nortekresources.com>
Sent: Wednesday, 21 June, 2023 5:09 PM
To: Loke Jian Munn <jm.loke at klk.com.my>; qgis-user at lists.osgeo.org; Richard McDonnell <richard.mcdonnell at opw.ie>
Subject: RE:

Hi Loke.

You came up with an interesting approach.  You are using the green band to complete an unsupervised classification using the Birch module in Scikit.  An unsupervised classification is just that, the implementation of an algorithm without training data.

I would suggest using your local point data from your local maxima routine as training data. You will have to come up with some logic to filter out the points that are actually palm centers, perhaps based on distance between trees.  I would then use one of the supervised classification modules to run a supervised classification on the training data ( local maxima). I would also experiment with all or 2 colour bands as more data may help the classification.

I must restate by previous comment that there is no bulletproof methodology to get 100 percent success.

Kirk Schmidt



Sent from my Galaxy


-------- Original message --------
From: Loke Jian Munn <jm.loke at klk.com.my<mailto:jm.loke at klk.com.my>>
Date: 2023-06-21 5:09 a.m. (GMT-04:00)
To: kirk <kirk at nortekresources.com<mailto:kirk at nortekresources.com>>, qgis-user at lists.osgeo.org<mailto:qgis-user at lists.osgeo.org>, Richard McDonnell <richard.mcdonnell at opw.ie<mailto:richard.mcdonnell at opw.ie>>
Subject: RE:

Hi,

As mention previously, I need NIR band to perform the NDVI, however my drone sensor does not record NIR band.

So, I trying another way which is crop counting from orthophoto with Python and Scikit Image. The whole process is geospatial as it works with raster and shapefiles and results are finally displayed on QGIS.

But the results I get are like below:

[cid:image004.png at 01D9A5B2.64B64990]

After Zoom in:
[cid:image005.png at 01D9A5B2.64B64990]

May I know what’s goes wrong ? I have attached the Python code above for references. For this project, I using an Orthomosaic layer with resolution of 5cm.

Appreciate if you can provide me some guidelines or insights in this as this is my first time touching python and scikit.

Thank you.

Regards,
Loke Jian Munn
Information Technology (IT) Department
[cid:image001.png at 01D9A5B2.64B64990]
KUALA LUMPUR KEPONG BERHAD
Wisma Taiko,
No. 1, Jalan S.P. Seenivasagam,
30000 Ipoh,Perak Darul Ridzuan,Malaysia
T    :  +605 240 8000 (ext. 2443)
[A close up of a sign  Description automatically generated]<https://clicktime.symantec.com/15tpDKjABDCCmu4drBEZD?h=4-LzQeEUOcKmp7sCwswGafke694l_dZCDQJ4Rb8N8bc=&u=http://www.klk.com.my/>[A close up of a sign  Description automatically generated]<https://clicktime.symantec.com/15tpJ9vSdpsoBqtZPjdhq?h=bX-LFOj0HrSJV_t-3L_WhMP3i5PRx5AtNBaHlI02C7k=&u=http://www.linkedin.com/company/klkberhad>

From: kirk <kirk at nortekresources.com<mailto:kirk at nortekresources.com>>
Sent: Tuesday, 6 June, 2023 6:09 PM
To: Loke Jian Munn <jm.loke at klk.com.my<mailto:jm.loke at klk.com.my>>; qgis-user at lists.osgeo.org<mailto:qgis-user at lists.osgeo.org>
Subject:

Hi Loke

Richard suggested using a NDVI approach which would require data in the Near Infrared (NIR) band.  An orthomisaic would contain 3 colour bands, but not the NIR.  NDVI provides an understanding of plant greenness and health and may not be suitable for detecting oil palm's.

If you are getting 90 percent plus with your current approach , you may find that you will spend 4 times longer capturing the additional 10 percent.

NIR data,are available from various satellite sensors however, their resolution would not work at the individual plant level.

Some drone sensors record the NIR band.

Kirk Schmidt



Sent from my Galaxy


DISCLAIMER : This email and any attachments are confidential and intended solely for the addressee. If you are not the intended recipient, any disclosure, copying, distribution either whole or partial, is prohibited and may be unlawful. In such case, please delete this message and kindly notify the sender.
Kuala Lumpur Kepong Berhad

DISCLAIMER : This email and any attachments are confidential and intended solely for the addressee. If you are not the intended recipient, any disclosure, copying, distribution either whole or partial, is prohibited and may be unlawful. In such case, please delete this message and kindly notify the sender.
Kuala Lumpur Kepong Berhad
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 8172 bytes
Desc: image001.png
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image002.png
Type: image/png
Size: 2313 bytes
Desc: image002.png
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image003.png
Type: image/png
Size: 1349 bytes
Desc: image003.png
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image004.png
Type: image/png
Size: 1479317 bytes
Desc: image004.png
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0008.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image005.png
Type: image/png
Size: 118345 bytes
Desc: image005.png
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20230623/0ce7d9b8/attachment-0009.png>