If you'd like to support our journey and help us provide even better GIS and Remote Sensing insights, consider contributing through 'Buy Me a Coffee'. Your support ensures we can continue creating tutorials, sharing tricks, and making our content as informative and enjoyable as possible. 🙏 Contribute💰💰 Here: www.buymeacoffee.com/jessebuyungo 🙏 Contribute💰💰Here: www.buymeacoffee.com/jessebuyungo
Dear Sir, when I run OBT_Vector Classifier, it does not produce any files. I think it is a problem reading the *.model file. How can I check this? Please can you help me? thank you very much
Hello, I have some problem with TrainVectorClassifier. When I choose "joined layer" as Input Vector Data I see a messege: "Unable to load selected layer or table. Dependent field cannot be filled in". How can I solve that? Thanks for your help.
How to fix this error using trainvectorclassifier ??? (FATAL) TrainVectorClassifier: itk::ERROR: FieldParameter(0000025D49A515F0): Value ['_mean', '_median', '_stdev'] not found in the list of choices: DN, _count, _sum, _mean, _median, _stdev, _min, _max, _range, _minority, _majority, _variety, _variance, id, class_id.
hi sr. It gives me this error when I apply ''segmetaion': '“.raster.out.tif” files are not supported as outputs for this algorithm' why and how should I correct it? you could tell me. I am looking forward to your response. thank you
Hello Jesse, I face a challenge, during segmentation the message "Self-intersection at or near point" appears, and I can not continue the segmentation process. how to fix this problem?
Thanks for your video. I have a problem with the last step ''Vector Classifier''. When I run the process, it doesn't make me the classified_output.shp. Do you know what happen? I need to do a special step?
I had to the same issue. Solved it by making sure that the file path didn't have any spaces in the name. Had 1 and replaced it with an '_'; things work!
Hi@@JorgeAlbertoGutierrezVillagran i have the same problem. When i run 'vector classifier' it doesn't create the classified_output.shp file. I downloaded QGIS version 3.32.1 Lima and OTB version 8.1.2, but the problem is not resolved. What can i do?
I still get the "Could not load selected layer/table. Dependent field could not be populated" error even when I saved joined_layer as shp file, can you please help?
2023-08-25 11:41:41 (WARNING) Segmentation: The input image has no data values but this application does not handle no-data. No-data pixels will be treated as regular pixels. 2023-08-25 11:41:41 (FATAL) Segmentation: itk::ERROR: No OGR driver known to OTB to create and handle a DataSource named . i don't know but my otb doesn't work at all
Hello, Jesse. Thanks for the great video. I encounter the problem, that my segmentation results generate dozens (more than 1000) self-intersecting polygons. With those I was not able to proceed the process, so I tried to just use the "fix geometry"-tool an so far it worked out fine.
@@jorritxl Hey Jorrit, you can find the "fix geometry"-tool in the qgis toolbox. For further explanations, check out the documentation or here: www.qgistutorials.com/en/docs/3/handling_invalid_geometries.html. Greetings
Hi, Thank you for your great video. I have to face a problem with the "Train Vector Classifier". I received an error message " TrainVectorClassifier: itk::ERROR: FieldParameter(0000022C97C15A30): Single selection mode is on, but there are 2 selected items ". Please help me.
Hi Jesse, when trying to run the TrainVectorClassifier I get the Error message that directory does not exist... What would be the work around for this?
I have a question there are three mean and statistics calculations Does that signify 3 bands of zonal statistics, Which means otb zonal statistics can process multiband zonal statistics calculation is that correct ??
Hello Jesse, thanks for creating such a valuable tutorial on OBIC. It's been incredibly helpful. I was wondering if you could kindly share the raster image you used.
OTB currently support only gdal and ogr provider. Parameter 'inzone.vector.in' uses 'memory' provider Execution failed after 0.04 seconds so, the second half of a video makes no sence to watch
Thanks for Video Jesse. Everything works up until the VectorClassifier part of the video. When I run VectorClassifier I get the following: 2023-03-31 21:22:23 (INFO) VectorClassifier: Default RAM limit for OTB is 256 MB 2023-03-31 21:22:23 (INFO) VectorClassifier: GDAL maximum cache size is 398 MB 2023-03-31 21:22:23 (INFO) VectorClassifier: OTB will use at most 8 threads Could not read file C:\GIS\back Any idea what the issue is?
Idem problem, no .model no .txt 2023-06-25 20:00:53 (INFO) VectorClassifier: Default RAM limit for OTB is 256 MB 2023-06-25 20:00:53 (INFO) VectorClassifier: GDAL maximum cache size is 788 MB 2023-06-25 20:00:53 (INFO) VectorClassifier: OTB will use at most 16 threads Could not read file C:\Users\Usuario\Desktop\Nueva
hey Jesse thanks for the video! I have a question in Train Vector Classifier part, if i have mean_3 and stdev_3 do i also put that? also why do we use mean and stdev in Train Vector Classifier?
Steve, Mean are parameters that are computed during the compute statistics stage, they are related to the mean pixel and the std of the pixel in each segment.......that's y the number (n) depends on the number of bands the image has.
@@jessebuyungo Hi, I have installed OTB (7.3.0) and QGIS (3.14), Win-10,64Bit. However, the Segmentation menu looks differant than you presented (LargeScaleMeanShift) , is it only a change in OTB UI ?
Professor, how can I create a DEM elevation model file from the surface of the external appearance using the Qgis Orfeo Toolbox plugin? Please answer me, and thank you very much, Doctor.
if I have sentinel 2 data, which data should I enter? are all bands or bands stacked? sorry if my english is not polite sir, i use google translate. Thank you sir
Thank you very much, clear, new and good tutorials: I have used a 0.15*0.15 high resolution digital orthophotos following your step, I get a good result. Thank you very much again
Dear Sir, thank you so much for this nice video! One question: how do you do accuracy assessment after the classification? Is there some kind of confusion matrix that I can get in QGIS after the classification? Thanks for your help
Yes please, prepare test and train data shapefile differently..... during model training...input the test data tooo.....and a confusion matrix will be output
Dear Karina, I also Face a challenge to perform an accuracy assesment in a train vector calssifier. Have you got a solution to accuracy assessment? please share me!!!
Hi there, am sorry, but I haven't used AutoCAD in this tutorial, Incase you are facing challenges, I can take you through under a private arrangement please
@@jessebuyungo I once again tried all the steps at the final classification stage to open the output shp file, I cannot find it despite the message specifying that the document has been classified, can you review this on your side maybe there was an error
hello, I face some struggle during the "VectorClassifier". The shp is missing at the end. I can find it nowhere and even I try to run it on temporary file, it doesn't show up either. Thanks in advance
@@jessebuyungo hello, so basically the algorithm work perfectly but I can't find the output. Plus there's no error message after running the "vectorclassifier". I just don't understand why it's missing. I choose the file for the results and everything but it's still missing
Hey great vid, I tried to copy but it resulted in this error: Warning 1: Self-intersection at or near point 145032.5 408201.5, do you know what went wrong?
Thank you so much Michael. Well when it comes to the data, you need to download satellite imagery. You could consider downloading freely available sentinel 2 or landsat 8.
@@jessebuyungo if I have sentinel 2 data, which data should I enter? are all bands or bands stacked? sorry if my english is not polite sir, i use google translate. Thank you sir
Hi sir I would like to ask if you have a tutorial on how to segment the tree crown using OTB or anything that will segment tree crown using the UAS data
@@jessebuyungo I used this approach but it was oversegmented and there was an error "Self-intersection at or near point". And I also used UAV image so it will have a high resolution. Can you recommend how to segment only the tree crown and the parameters to used for this? please
I am having a hard time with this, my results layer ends up flipped like a mirror above my satmap. I think this is making the rest of this project not work.
...I can't say that there is an accurate model, different model will be accurate depending on the given situation, maybe the landscape, the samples, the satellite imagery etc.....
@@jessebuyungo I undestand, another thing, im trying to do the segmentation with a landsat but it doesnt work, the output is just a big polygon that cover the study area. Do I need to change some settings for landsat? I tried to resample at10 meters but It doesnt work
sir is there any way you could demonstrate sun glint and water column correction and OBIA using planet data in qgis or arcgis for benthic habitat mapping?
Hi, I have the same problem. I am trying to apply classification to images acquired with UAV on the water surface (full motio video) to detect posidonia. Have you had any feedback to your question? can you give me any suggestions please? thanks
Thanks for this great video! I am trying your method and so far I have no problem from the start until the part of "vector classifier". It keeps showing this: ectorClassifier: itk::ERROR: ListViewParameter(000002495504F300): Value mean_0 not found in the list of choices: Value, Count, NbPixels, Flat, Round, Elong, Perim, meanB1, meanB2, meanB3, stdB1, stdB2, stdB3, MedB1, MedB2, MedB3, VarB1, VarB2, VarB3, KurtB1, KurtB2, KurtB3, SkewB1, SkewB2, SkewB3, NbPixels_1, Flat_1, Round_1, Elong_1, Perim_1, meanB1_1, meanB2_1, meanB3_1, stdB1_1, stdB2_1, stdB3_1, MedB1_1, MedB2_1, MedB3_1, VarB1_1, VarB2_1, VarB3_1, KurtB1_1, KurtB2_1, KurtB3_1, SkewB1_1, SkewB2_1, SkewB3_1.
Have you computed statistics for the input layer, if Yes make sure the stat layer is joined and the input to the classifier is that joined layer, if you have done this, you can share a snapshot of the error and I see how to help
When I do categorized after the classified output, i get only 1 class; though i have created 5 classes during the training data/sample. What seems to be the problem?
Please open the attribute table of that layer and check first the predicted classes, if you see all the five classes then go on to the symbology and categorize, Alternatively if you see only one class predicted then please you need to recustomize your training parameters and re-run the algorithm.
My solution to this problem was that my file path to the input/output layers had a space in it. I created a new filepath without spaces and it was successful.
i have 6 class_id´s but predicted only classifies everything as the same (i.e. 1 of 6), even those polygons which have been given class ids. potential issues might be that im using sentinel-2 data, which results in 9x mean and stdev fields. additionally, after the join by location, are you using a filtered dataset for the training (ie only those polygons that have been assigned class_ids) or does your training dataset contain all polygons with both class id´s assigned and polygons with class id NA?
Hey, great work. However, I can´t use TrainVectorClassifier, it keeps showing the same error, "Could not load selected layer/table. Dependent field could not be populated", after loading the joined layer. How can I fix it? Thank y
@@jessebuyungo I have saved it as a .shp file but its still showing the same error when I am trying to carry out the TrainVectorClassifier. How can I resolve this issue?
Well I you like to do accuracy assessment, make sure you pick testing samples and the same way you did for the training data. And supply it to the algorithm during the training. It will give you a confusion matrix which you can analyze ......and generate producer, user, overall accuracy parameters, etc
Hi great tutorial right there. If you don’t mind answering, i would love to ask. Can the analysis be done at urban/regional scale spatial unit? Because from your tutorial, the analysis is done at neighborhood scale
Thanks for this great video! I am trying your method and so far I have no problem from the start until the part of "vector classifier". It keeps showing this: "2021-08-10 22:20:06 (FATAL) VectorClassifier: itk::ERROR: Failed to open GDALDataset file C:\Users\" and it does not produce the classified output layer. I am using QGIS 3.18 Zurich and OTB (7.3) Windows 64bits with OTB master plugin. I wonder what could be the problem? I would appreciate your help
Hi, very informative video but I have this error saying: "Classified_output" has invalid geometry. Please fix the geometry or change the Processing setting to the "Ignore invalid output features" option. This is occured when I try to compute or Dissolve the Classified Output.
If you'd like to support our journey and help us provide even better GIS and Remote Sensing insights, consider contributing through 'Buy Me a Coffee'. Your support ensures we can continue creating tutorials, sharing tricks, and making our content as informative and enjoyable as possible.
🙏 Contribute💰💰 Here: www.buymeacoffee.com/jessebuyungo
🙏 Contribute💰💰Here: www.buymeacoffee.com/jessebuyungo
Dear Sir, when I run OBT_Vector Classifier, it does not produce any files.
I think it is a problem reading the *.model file. How can I check this?
Please can you help me? thank you very much
Hi, I am having the same problem
I am having the same problem
Hi, I am having the same problem
I am having the same problem
I have the same problem. How can i solve it?
Hello, I have some problem with TrainVectorClassifier. When I choose "joined layer" as Input Vector Data I see a messege: "Unable to load selected layer or table. Dependent field cannot be filled in". How can I solve that?
Thanks for your help.
Great. Can you make a separate video for accuracy assessment as it is necessary after classification to prove our results?
How to fix this error using trainvectorclassifier ???
(FATAL) TrainVectorClassifier: itk::ERROR: FieldParameter(0000025D49A515F0): Value ['_mean', '_median', '_stdev'] not found in the list of choices: DN, _count, _sum, _mean, _median, _stdev, _min, _max, _range, _minority, _majority, _variety, _variance, id, class_id.
Super helpful and clear instructions, can't thank you enough for making this.
hi sr. It gives me this error when I apply ''segmetaion': '“.raster.out.tif” files are not supported as outputs for this algorithm' why and how should I correct it? you could tell me. I am looking forward to your response. thank you
Thanks for the help. Signed by the Water Sciences Master's degree from Lyon (France).
Is it possible to have another tutorial on how to use the Segment Anything Model in QGIS?
Hello! Great video!
Which satellite image is this? Its prett high resolution!
Hello Jesse, I face a challenge, during segmentation the message "Self-intersection at or near point" appears, and I can not continue the segmentation process. how to fix this problem?
Hi, can you help me, I couldn’t find the vector classifier output in the folder even though the process is finished
Hi, i have the same problem. Did you solve it?
Thanks for your video.
I have a problem with the last step ''Vector Classifier''. When I run the process, it doesn't make me the classified_output.shp.
Do you know what happen? I need to do a special step?
I had to the same issue. Solved it by making sure that the file path didn't have any spaces in the name. Had 1 and replaced it with an '_'; things work!
@@Technologic911 Thanks, it isn't the problem. It was the qgis version because the otb version needs a upgrade qgis and just I download the new qgis
This is happening to me, what version of qgis and otb did you need to get it to work? Thanks
@@aislingoflynn6167 3.20 or upper
Hi@@JorgeAlbertoGutierrezVillagran i have the same problem. When i run 'vector classifier' it doesn't create the classified_output.shp file. I downloaded QGIS version 3.32.1 Lima and OTB version 8.1.2, but the problem is not resolved. What can i do?
Very good instructions and it really helped me. Thank you!
This is a very good tutorial, thanks for posting.
I still get the "Could not load selected layer/table. Dependent field could not be populated" error even when I saved joined_layer as shp file, can you please help?
I have the same problem, plus my joined layer has only 1 set of mean and stdev fileds
2023-08-25 11:41:41 (WARNING) Segmentation: The input image has no data values but this application does not handle no-data. No-data pixels will be treated as regular pixels.
2023-08-25 11:41:41 (FATAL) Segmentation: itk::ERROR: No OGR driver known to OTB to create and handle a DataSource named .
i don't know but my otb doesn't work at all
Hello, Jesse. Thanks for the great video. I encounter the problem, that my segmentation results generate dozens (more than 1000) self-intersecting polygons. With those I was not able to proceed the process, so I tried to just use the "fix geometry"-tool an so far it worked out fine.
Hey hannes, i had the same error, were can i find the 'fix geometry' thing?
@@jorritxl Hey Jorrit,
you can find the "fix geometry"-tool in the qgis toolbox.
For further explanations, check out the documentation or here: www.qgistutorials.com/en/docs/3/handling_invalid_geometries.html.
Greetings
👍👍👍 Am closely following once I try it out will have my reactions
No worries Isaac.
Make sure you have a working installation of the OTB tool box
train vector classifier
'Could not load selected layer/table. Dependent field could not be populated'. So no sence to watch AFTER THAT AS WELL
The OBIA_model.model doesn't want to save on our computer as an output-file during the TrainVectorClassifier-step. Do you know what the problem is?
This have helped me a lot in my work. Thanks a lot!
Hi, Thank you for your great video. I have to face a problem with the "Train Vector Classifier". I received an error message " TrainVectorClassifier: itk::ERROR: FieldParameter(0000022C97C15A30): Single selection mode is on, but there are 2 selected items ". Please help me.
my classified output is not appearing. What could be the problem?
sorry , what could be the problem
Hi Jesse, when trying to run the TrainVectorClassifier I get the Error message that directory does not exist...
What would be the work around for this?
That is a lot of work! Thanks for sharing
Great Explanation! Thanks!
What means "More than 1000 errors or warnings have been reported. No more will be reported from now." ? Please help me.
Super 🥰 ! Thanks... 🤗
mil graciassssss estaba muy perdida me ayudo mucho el video saludos desde Colombia
Thank you so much
I have a question there are three mean and statistics calculations Does that signify 3 bands of zonal statistics,
Which means otb zonal statistics can process multiband zonal statistics calculation is that correct ??
Is it possible to reuse the segmentation on other similar satellite images? So you dont have to set new datapoints on a new image?
yes please, you have to segment the imagery each time
Super helpful 🎉
Please help! Vector classification does not have an output for me, maybe its because of the model? Im using 3.32.1 - Lima version
Please share a screenshot error
Please share a screenshot error
i encountered an error whilist activating the otb
it said there is a typeError in the python so how can i solve that
Thank you so much, sir, it was really helpful!
Glad it helped!
Hello Jesse, thanks for creating such a valuable tutorial on OBIC. It's been incredibly helpful. I was wondering if you could kindly share the raster image you used.
Wow, this is a Pléiades satellite image taken in Kampala CBD, Uganda
@@jessebuyungo Thanks. I could not find it on EarthExplorer. I wonder if you still have that image.
It's a commercial satellite image. So you have to buy it at a cost
OTB currently support only gdal and ogr provider. Parameter 'inzone.vector.in' uses 'memory' provider
Execution failed after 0.04 seconds
so, the second half of a video makes no sence to watch
Thanks for Video Jesse. Everything works up until the VectorClassifier part of the video. When I run VectorClassifier I get the following:
2023-03-31 21:22:23 (INFO) VectorClassifier: Default RAM limit for OTB is 256 MB
2023-03-31 21:22:23 (INFO) VectorClassifier: GDAL maximum cache size is 398 MB
2023-03-31 21:22:23 (INFO) VectorClassifier: OTB will use at most 8 threads
Could not read file C:\GIS\back
Any idea what the issue is?
Idem problem, no .model no .txt
2023-06-25 20:00:53 (INFO) VectorClassifier: Default RAM limit for OTB is 256 MB
2023-06-25 20:00:53 (INFO) VectorClassifier: GDAL maximum cache size is 788 MB
2023-06-25 20:00:53 (INFO) VectorClassifier: OTB will use at most 16 threads
Could not read file C:\Users\Usuario\Desktop\Nueva
I cannot find OTB Zonal statistics tool on my QGIS 3.22
Te amo mi rey. Sos un Genio!!!! Mis aplausos para Usted!!!!
hey Jesse thanks for the video! I have a question in Train Vector Classifier part, if i have mean_3 and stdev_3 do i also put that? also why do we use mean and stdev in Train Vector Classifier?
Steve, Mean are parameters that are computed during the compute statistics stage, they are related to the mean pixel and the std of the pixel in each segment.......that's y the number (n) depends on the number of bands the image has.
Nice video and understandable instructions. please sir, can you share my your presentation slide on this video or pdf.
What's the sourse of your satellite image. And how did you get your resolution to be that good?
.model file not creating
plz tell me what to do
Thank You, Which Ver. of QGIS is required ?, How do I check that OTB Tool box is installed ?
You're welcome,
For this specific tutorial, I used Qgis Hannover, after installing the OTB, you will begin seeing it's algorithms in the tool box menu
@@jessebuyungo Hi, I have installed OTB (7.3.0) and QGIS (3.14), Win-10,64Bit. However, the Segmentation menu looks differant than you presented (LargeScaleMeanShift) , is it only a change in OTB UI ?
@@davidbehr7865 Ohh could you please share the snapshot in an email please @jessebuyungo48@gmail.com
Professor, how can I create a DEM elevation model file from the surface of the external appearance using the Qgis Orfeo Toolbox plugin? Please answer me, and thank you very much, Doctor.
Does anyone know how to increase the number of iterations for the SVM classification method using the train vector classifier
heey brother l need your help. Isent you an email concerning this topic
please resend the email I have failed to identify it , I am sorry I haven't replied, yet.
@@jessebuyungo l managed to figure it out.thank you Jesse
I have this series QGİS 3.20.3. Is it possible with it?The segmentation tool is not found when searching in the toolbox.Please help me
Please have you installef the OTB plugin already??
what version are you using?
Hannover
if I have sentinel 2 data, which data should I enter? are all bands or bands stacked? sorry if my english is not polite sir, i use google translate. Thank you sir
No worries,. Its fine.
Sentinel 2 stack B,2,3,4 to make RGB color image
which version of qgis did u use
Will it work with landsat 7 data also
yes but at a lwoer resolution
Thank you very much, clear, new and good tutorials: I have used a 0.15*0.15 high resolution digital orthophotos following your step, I get a good result. Thank you very much again
Dear Sir, thank you so much for this nice video! One question: how do you do accuracy assessment after the classification? Is there some kind of confusion matrix that I can get in QGIS after the classification?
Thanks for your help
Yes please, prepare test and train data shapefile differently..... during model training...input the test data tooo.....and a confusion matrix will be output
Dear Karina, I also Face a challenge to perform an accuracy assesment in a train vector calssifier. Have you got a solution to accuracy assessment? please share me!!!
@@jessebuyungo Hello Sir. Can you make a separate video for accuracy assessment as it is necessary after classification to prove our results?
can you explain why u use autocad and not shp in the last, i didnt see my file output any help pls ?
Hi there, am sorry, but I haven't used AutoCAD in this tutorial, Incase you are facing challenges, I can take you through under a private arrangement please
@@jessebuyungo I once again tried all the steps at the final classification stage to open the output shp file, I cannot find it despite the message specifying that the document has been classified, can you review this on your side maybe there was an error
Ohhh sorry chief. Try checking in the directory you saved it
hello, I face some struggle during the "VectorClassifier". The shp is missing at the end. I can find it nowhere and even I try to run it on temporary file, it doesn't show up either. Thanks in advance
@jessybuyungo I need your help
Hello chief, what does it say after classification?, does it give an error?
@@jessebuyungo hello, so basically the algorithm work perfectly but I can't find the output. Plus there's no error message after running the "vectorclassifier". I just don't understand why it's missing. I choose the file for the results and everything but it's still missing
I can't understand the problem, maybe we can arrange an online meet and you take me through and I can guide you where you are going wrong
@@jessebuyungo ohh really, thanks in advance.
Hey great vid, I tried to copy but it resulted in this error: Warning 1: Self-intersection at or near point 145032.5 408201.5, do you know what went wrong?
Did it stop there or it proceeded???
Dear jorrit, I also face this challenge? have you get a solution?
Hello Jesse, thank you for your great tutorial video. I have a question regarding the OTB plugin. I couldn't find the plugin in QGIS 3.16.
Please you need to install otb, watch my tutorial on how to
Great video. Can you share the tutorial data, please? Can I use Google Earth Imagery for this?
Thank you so much Michael. Well when it comes to the data, you need to download satellite imagery.
You could consider downloading freely available sentinel 2 or landsat 8.
@@jessebuyungo if I have sentinel 2 data, which data should I enter? are all bands or bands stacked? sorry if my english is not polite sir, i use google translate. Thank you sir
@@einsor7260 stack b2,3,4
Hii there! the segmentation process shows completed within seconds and there are no results...please help
Could you please share a screenshot of the error/ result screen
Hi sir I would like to ask if you have a tutorial on how to segment the tree crown using OTB or anything that will segment tree crown using the UAS data
Try using this approach and check the results please, however this requires high resolution imagery
@@jessebuyungo I used this approach but it was oversegmented and there was an error "Self-intersection at or near point". And I also used UAV image so it will have a high resolution. Can you recommend how to segment only the tree crown and the parameters to used for this? please
I am having a hard time with this, my results layer ends up flipped like a mirror above my satmap. I think this is making the rest of this project not work.
Could you please share a snapshot of the output
I have the same problem, how can I send you a screenshot?
is there a solution for this?
Thank you sir for the tutorial. I want to ask you how to extract lineament using obia?
Well, the procedure is the same
I don't have the segmentation option in my toolbox
Really?
if I use polygons instead of points for the training sample it the classification goin to be more accurate?
We. Are not using points actually, we are selecting points which we later use to select our polygons
@@jessebuyungo got it!! in your opinion what model is the most accurate? Im trying em all but it seems that linear and polinomial are the better ones
...I can't say that there is an accurate model, different model will be accurate depending on the given situation, maybe the landscape, the samples, the satellite imagery etc.....
@@jessebuyungo I undestand, another thing, im trying to do the segmentation with a landsat but it doesnt work, the output is just a big polygon that cover the study area. Do I need to change some settings for landsat? I tried to resample at10 meters but It doesnt work
Can I give you a tutorial for that at a discounted price?
sir is there any way you could demonstrate sun glint and water column correction and OBIA using planet data in qgis or arcgis for benthic habitat mapping?
Hi, I have the same problem. I am trying to apply classification to images acquired with UAV on the water surface (full motio video) to detect posidonia.
Have you had any feedback to your question? can you give me any suggestions please? thanks
you did not mention the source of the imagery
Please feel free to use any high resolution imagery
Good day Sir, Can I ask the Name of the Plugins used for segmentation? Thank you
orfeo tool box
OTB
@@jessebuyungo Where do I get this plugin from? I have Orfeo Toolbox provider plugin installed but it's not showing in the panel.
Thanks for this great video! I am trying your method and so far I have no problem from the start until the part of "vector classifier". It keeps showing this: ectorClassifier: itk::ERROR: ListViewParameter(000002495504F300): Value mean_0 not found in the list of choices: Value, Count, NbPixels, Flat, Round, Elong, Perim, meanB1, meanB2, meanB3, stdB1, stdB2, stdB3, MedB1, MedB2, MedB3, VarB1, VarB2, VarB3, KurtB1, KurtB2, KurtB3, SkewB1, SkewB2, SkewB3, NbPixels_1, Flat_1, Round_1, Elong_1, Perim_1, meanB1_1, meanB2_1, meanB3_1, stdB1_1, stdB2_1, stdB3_1, MedB1_1, MedB2_1, MedB3_1, VarB1_1, VarB2_1, VarB3_1, KurtB1_1, KurtB2_1, KurtB3_1, SkewB1_1, SkewB2_1, SkewB3_1.
Have you computed statistics for the input layer, if Yes make sure the stat layer is joined and the input to the classifier is that joined layer, if you have done this, you can share a snapshot of the error and I see how to help
You should try input "meanB1 meanB2" etc. for training vector classifier.
Great video.
Thanks
When I do categorized after the classified output, i get only 1 class; though i have created 5 classes during the training data/sample. What seems to be the problem?
Please open the attribute table of that layer and check first the predicted classes, if you see all the five classes then go on to the symbology and categorize,
Alternatively if you see only one class predicted then please you need to recustomize your training parameters and re-run the algorithm.
did you resolve the problem? i haver the same issue (only one predicted classs, even for polygons that have been classified as a different class)
Thank you for this video, can we use google street view or image, and use Qgis for image classification without using deep learning?
No please. We have to use Aerial or satellite imagery
why isn't it possible to use an image you took by yourself?
@@jessebuyungo
when the vector classifier process there is an error : failed to open GDALDataset file D,
how to handle it
Could you please share a snapshot of the error screen and we see how to help you please
My solution to this problem was that my file path to the input/output layers had a space in it. I created a new filepath without spaces and it was successful.
i have 6 class_id´s but predicted only classifies everything as the same (i.e. 1 of 6), even those polygons which have been given class ids. potential issues might be that im using sentinel-2 data, which results in 9x mean and stdev fields. additionally, after the join by location, are you using a filtered dataset for the training (ie only those polygons that have been assigned class_ids) or does your training dataset contain all polygons with both class id´s assigned and polygons with class id NA?
2021-12-01 15:29:08 (INFO) VectorClassifier: Default RAM limit for OTB is 256 MB
2021-12-01 15:29:08 (INFO) VectorClassifier: GDAL maximum cache size is 813 MB
2021-12-01 15:29:08 (INFO) VectorClassifier: OTB will use at most 8 threads
2021-12-01 15:29:08 (INFO) VectorClassifier: Model loaded
2021-12-01 15:29:08 (INFO) VectorClassifier: mean used: [0]
2021-12-01 15:29:08 (INFO) VectorClassifier: standard deviation used: [1]
2021-12-01 15:29:08 (INFO) VectorClassifier: Loading model
log files from vectorClassifier output.
May be you Didn't configure the training very well.
Could you please share a screenshot of the error?
@@jessebuyungo greeting of peace, I have same problem.
can i use this methodology for a binary classification?
Not so sure, it may not work, still you can do a multi class classification and visualize it as a binary one...... Quite simple to do
Hey, great work. However, I can´t use TrainVectorClassifier, it keeps showing the same error, "Could not load selected layer/table. Dependent field could not be populated", after loading the joined layer. How can I fix it? Thank y
Save as .shp
@@jessebuyungo I have saved it as a .shp file but its still showing the same error when I am trying to carry out the TrainVectorClassifier. How can I resolve this issue?
Which Qgis version are you using?
@@jessebuyungo QGIS DESKTOP 3.16.1
@@bidiptaghosal7411 I don't know now because I use the same version
Great job @Jesse!! I would like to look at the Accuracy assessment results. How accurately did the algorithm predict the class?
Well I you like to do accuracy assessment, make sure you pick testing samples and the same way you did for the training data.
And supply it to the algorithm during the training.
It will give you a confusion matrix which you can analyze ......and generate producer, user, overall accuracy parameters, etc
Hi great tutorial right there. If you don’t mind answering, i would love to ask. Can the analysis be done at urban/regional scale spatial unit? Because from your tutorial, the analysis is done at neighborhood scale
Yes please you can do the analysis at a regional scale
Jesse Buyungo does it require survey or sampling ?
@@hernandbagaskara sampling is fine
Can I do it in ArcGIS?
I think you should be able to
Thanks for this great video! I am trying your method and so far I have no problem from the start until the part of "vector classifier". It keeps showing this: "2021-08-10 22:20:06 (FATAL) VectorClassifier: itk::ERROR: Failed to open GDALDataset file C:\Users\" and it does not produce the classified output layer. I am using QGIS 3.18 Zurich and OTB (7.3) Windows 64bits with OTB master plugin. I wonder what could be the problem? I would appreciate your help
well could you please share a snapshot of your Input parameters plus the error screen snapshot please
@@jessebuyungo would you mind if I share it in your email instead here? thanks
@@sarahjane4484 kindly contact me using: jessebuyungo48@gmail.com
Whatever file you save, and the folders in which the files are located make sure their names don't have a space, rather use _ instead of spaces.
@@jessebuyungo I have sent you an email regarding the same problem.
Hi, very informative video but I have this error saying: "Classified_output" has invalid geometry. Please fix the geometry or change the Processing setting to the "Ignore invalid output features" option.
This is occured when I try to compute or Dissolve the Classified Output.
Could you share the error snapshot please