POST
|
I have been working through my own version of detecting swimming pools. I have use the same imagery (albeit, a subset) used by ESRI. I think I have taken reasonable training sites. I have tried to be meticulous with selecting training sites. The ones I choose seem similar to those found in the ESRI examples. I have used the same (and different parameters for training the model and detecting objects). I have taken the defaults, added values shown in ESRI examples, and tried some of my own options for parameters such as: image tile, stride size grid, scaling, ratio padding, threshold, etc epochs The "learning curve" has shown concave to "jagged" (which I believe means, overfitted). Like I mentioned, I have tried this scenario numerous times with the different combinations of parameters for both training and detecting. I am no where close to successfully detecting swimming pools. I mostly get features other than pools. I am attaching my PPKX with the training and detected objects (and the image), about 23 Mb. I do have a model zip file, but it is approx. ~200 Mb, so it may be too large. I am attaching the EMD file and the PNG that shows the loss curve. This is one is jagged. 25 epochs. The attachment has about 15 sample sites, but I have done this same process using a larger image and approx. 80 samples that I think are "good" samples of swimming pools...same kind of results. If anyone can provide some suggestions or guidance, I would like to successfully attempt to find swimming pools....and then work with detecting other kinds of features. I am using an XPS Dell business machine. Pro 2.5. I have the deep learning framework installed. No GPU or NVIDIA card. I can successfully run the full process, but the results are very bad. I have viewed and use: Deep Learning: Detect swimming pools as a guide in addition to the ESRI help and other ArcGIS.learn notebooks and other information provided by ESRI. Thank you for any insight or let me know if I should be contacting someone else at ESRI. I would also like to know if there is a process to go through to logically attempt different parameter values that can improve both training and detecting results. Nate
... View more
04-17-2020
05:02 PM
|
0
|
0
|
876
|
POST
|
Is there any documentation on the optional parameters for items like padding, threshold, block size, etc? I am interested in guidance on which values to enter and the reasoning behind it. The various canned samples have values, but no explanation and the ESRI help is not too useful, either. I am working through my own version of identifying pools using the ESRI imagery, Pro 2.5. I can't seem to access the "live" examples using my various ESRI logins... It would be most helpful to provide some additional guidance, logic, and reasoning for when and what to use for these parameters. The values I am using (which seem to be default or a best guess) are not providing good results to identifying objects. Also, what are all of the various parameters to use? Thank you, Nate
... View more
04-13-2020
11:40 AM
|
0
|
1
|
507
|
POST
|
I have a similar problem. I didn't see your post, but I removed the deeplearning clone and went to reinstall and it looked like all of the "installed" modules were still installed..When I redid the clone, my problem still persists. Did you actually "uninstall" the various Python packages? It seems like the ones I have are correct based on the instructions. Here is my post: https://community.esri.com/thread/248776-train-deep-learning-model-not-recognizing-training-folders Let me know if you or anyone has any other suggestions. I have not done anything differently than use the provided instructions both for the Python install and the steps to work on swimming pool ESRI example data. Thank you, Nate
... View more
02-26-2020
04:30 PM
|
0
|
1
|
1168
|
POST
|
I am on Pro 2.5.0 I followed the install instructions as per the Pro Help: Install deep learning frameworks for ArcGIS—ArcGIS Pro | Documentation I am using the sample image for swimming pools found here: https://www.arcgis.com/apps/MinimalGallery/index.html?appid=99c39f7512d54881bc365583c76c7da6 for the PyTorch swimming pool example. I created about 10 samples using the Label Objects for Deep Learning. I saved the samples and the labels and then ran the Export Training data which created the images and labels folder, the other files required for the next step to train the model. The input folder is never recognized (which is: \deep_learning\redlands_img that contains the above folders and files). I get the following error and cannot seem to fix it. Any help is appreciated. I am including my zip file of output. Also, the documentation for the specific tools and the overview don't really provide enough details to set tool parameters or suggestions for different types of scenarios. Some updated material would be most helpful for the deep learning community that support Pro 2.5. Thank you, Nate
... View more
02-24-2020
09:49 PM
|
0
|
7
|
2453
|
POST
|
Did you receive any resolution on this? I have almost the exact same problem. I had Anaconda installed to help someone with a script outside of ArcGIS. I uninstalled Anaconda. This did't help. I completely uninstalled 10.5.1 and reinstalled. This did not solve my problem. My sys.path shows this: >>> sys.path ['', u'c:\\program files (x86)\\arcgis\\desktop10.5\\arcpy', 'C:\\WINDOWS\\SYSTEM32\\python27.zip', 'C:\\Python27\\ArcGIS10.5\\DLLs', 'C:\\Python27\\ArcGIS10.5\\lib', 'C:\\Python27\\ArcGIS10.5\\lib\\plat-win', 'C:\\Python27\\ArcGIS10.5\\lib\\lib-tk', 'C:\\Python27\\ArcGIS10.5', 'C:\\Python27\\ArcGIS10.5\\lib\\site-packages', 'C:\\Program Files (x86)\\ArcGIS\\Desktop10.5\\bin', 'C:\\Program Files (x86)\\ArcGIS\\Desktop10.5\\ArcPy', 'C:\\Program Files (x86)\\ArcGIS\\Desktop10.5\\ArcToolBox\\Scripts'] >>> When I do import numpy, I see this problem Traceback (most recent call last): File "<stdin>", line 1, in <module> File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\__init__.py", line 187, in <module> from . import add_newdocs File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\add_newdocs.py", line 13, in <module> from numpy.lib import add_newdoc File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\lib\__init__.py", line 8, in <module> from .type_check import * File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\lib\type_check.py", line 11, in <module> import numpy.core.numeric as _nx File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\__init__.py", line 9, in <module> from . import numerictypes as nt File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\numerictypes.py", line 968, in <module> _register_types() File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\numerictypes.py", line 965, in _register_types numbers.Integral.register(integer) AttributeError: 'module' object has no attribute 'Integral' Any help or insight from anyone, is greatly appreciated. I do have ArcPro 2.0.1 Both Desktop and Pro were running fine. I don't think I "unchecked" the correct boxes for Anaconda, but with a full uninstall/re-install of ArcGIS 10.5.1, I thought this would solve my problem, but it did not. I do see that I still have the C:\Python27\ArcGIS10.4 folder still around, since I originally upgraded in place 10.5.1. Thank you for any assistance or insight. UPDATE: I deleted the ArcGIS10.4 folder, I completely uninstalled 10.5.1. I reinstalled the 10.5.1 software, took all of the defaults for the install. I made sure before I reinstalled, that Python was not on my system. I can type in Python at a CMD window and type in import numpy, import arcpy. A prompt returns, no problem. When I write a simple script in the IDLE window with print "test" import arcpy print "this is a test" The first line is printed in the shell, then I get an error that the import arcpy line doesn't work. See the errors below. If I use CMD to bring up a prompt, then type in "python" to get a python prompt and type in print "test" test (is printed) import arcpy (prompt is returned) print "this is a test" this is a test (is printed) I also tested within ArcMap using the Python window and did a Buffer. The buffer was successful using the Python window. I am really not sure what is/could be going on. I believe I have only one installation of Python (the 2.7.13 that comes with the ESRI 10.5.1 install). Traceback (most recent call last): File "C:\temp\junk4.py", line 8, in <module> import arcpy File "C:\Program Files (x86)\ArcGIS\Desktop10.5\ArcPy\arcpy\__init__.py", line 20, in <module> import numpy File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\__init__.py", line 187, in <module> from . import add_newdocs File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\add_newdocs.py", line 13, in <module> from numpy.lib import add_newdoc File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\lib\__init__.py", line 8, in <module> from .type_check import * File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\lib\type_check.py", line 11, in <module> import numpy.core.numeric as _nx File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\__init__.py", line 9, in <module> from . import numerictypes as nt File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\numerictypes.py", line 968, in <module> _register_types() File "C:\Python27\ArcGIS10.5\lib\site-packages\numpy\core\numerictypes.py", line 965, in _register_types numbers.Integral.register(integer) AttributeError: 'module' object has no attribute 'Integral' I also noticed this.... I had part of a simple script I wrote to demonstrate how to work with numbers. I called this script numbers.py and saved it in my "c:\temp" folder. When I posted my error messages above, the results of my numbers.py script were showing up and I could not figure out why. After running this script (or when re-installing ArcGIS and the 2.7.13 version of Python, I noticed that a numbers.pyc file was also created in c:\temp. I did not create the .pyc file. As a test, I renamed both the numbers.py and numbers.pyc files to numbersx.py and numbersx.pyc. When I reran my scripts, they all worked! I even did an independent test with a simple numpy script from here: https://docs.scipy.org/doc/numpy-dev/user/quickstart.html to test the use of the numpy module. A script using seems to work. I tested a script using the buffer routine to test the use of the arcpy module. This seems to work. I am not sure why having a script I wrote, called numbers.py in the c:\temp folder should matter, but it does. I did see that there is a numbers.py file in the C:\Python27\ArcGIS10.5\Lib which relates to the functioning of Python, so I am guessing that the file I created in my temp folder was causing the problem. If anyone has any insight on this, please share. I guess, I will be more careful with naming my python scripts. I was using what I thought was simple names so others newer to Python could use my scripts to learn. The teacher continues to learn, too. Thank you, Nate
... View more
01-15-2018
09:08 PM
|
0
|
1
|
768
|
POST
|
City of Sacramento. We do have an ADA coordinator. I took a look at the Section508 documentation and forwarded it to our Web Master who is aware of ADA compliance issues. I suspect we will follow up with our internal ADA coordinator for some recommendations. My hope is that we can provide interactive web maps (and apps) through the City's web page while being ADA compliant.
... View more
02-12-2016
11:08 AM
|
1
|
0
|
684
|
POST
|
I have the same question. Not just with Story Maps, but any "embedded" web map.
... View more
02-12-2016
09:22 AM
|
2
|
0
|
684
|
POST
|
I don't know if I am missing some concept, but I do have "time aware" data set up in my ArcGIS Online space, I have set the "Time" configuration to use a time range that contains "start" and "end" dates, My data "Date types" are set as "Date" When I configure the Time widget in Web App Builder, it looks like the time span is only looking at the "Start" times of my data and not the "end" times of my data. Let me know if I need to configure my data differently or if this is a possible bug. Beta site: ArcGIS Web Application and see the attached word document for some screen shots of the configuration. Thank you, Nate
... View more
02-05-2016
01:01 PM
|
0
|
0
|
2000
|
POST
|
I am trying to use the Directions template as "the template" for providing a web application that uses routing. I have been getting inconsistent results and functionality when trying to route from a user selected location or an address they type in the search box to a destination (which right now is a feature layer with polygons (i.e. parking garages). I have chosen to use the ESRI routing option. Because my results (such as turn directions and the route does not show up consistently), I am reaching out to the user community to see if they have any suggestions. I am aware that the Directions template is one of many that can be used as a Web App. The general work flow as I understand it is: 1. Add a feature layer to a map. Save the map 2. Choose to Create a Web App 3. Select a template if desired (I chose the Directions template) 4. Configure look/feel and functionality (such as routing). If you have any suggestions or insight on the routing inconsistency issue. Thank you for your reply. Nate
... View more
12-14-2015
02:32 PM
|
0
|
0
|
295
|
POST
|
I am trying to use the Directions template in Web App Builder (the one on ArcGIS Online). I have a polygon layer of parking garages that I would like to use as my destination. I have this layer in my web map that I created. I would like to use the ESRI provided Routing service (yes, I know it uses credits). When I attempt to use an address or click a point on the map and then click the "routing icon" on my list of parking garages, I get the following. "Route service description was not reached or recognized. Please check service URL and/or user credentials, if specified." I have found no real "help" on this and not sure what my options are. Also, I noticed that every time I refresh the map, the Index numbers change for the specific garages....this seems odd. Thank you for any assistance. Nate
... View more
10-26-2015
01:31 PM
|
0
|
2
|
3316
|
POST
|
This is not clear when working with AGOL. I was wanting to set up a "view" only type of map that a user could use to "view" the data but not edit. It seems like the Collector App could be used for this option too. For the collector app, it is easy to set up a data set and a map, then make it available for a Collector app user to use, but with a "view only" type option, this is not clear and seems like one "must have" an editable feature to actually use the Collector app. I was hoping to use the Collector app for both "data collecting" and for a "manager" to be able to "view" data while in the field, but not really needing to "change" any data. I would rather not produce two different solutions for "view only" and "collect/edit" data. Let me know of any feedback or suggestions. Thank you, Nate
... View more
04-01-2015
10:11 AM
|
0
|
1
|
109
|
POST
|
Signal Inventory | ArcGIS for Local Government I am just starting to consider this solution to start to create a signal/pole/cabinet (layer or layers). I would like some insight to this solution if anyone has used it/is using it, how the data is being managed, are there a bunch of relationship classes to manage (e.g. many signals on one pole, many poles related to a single cabinet, many parts of signals on one signal), issues/challenges with many different (potentially overlapping features) at a single location (e.g. many signals at a single pole location), using/configuring the data editing tools and rules. I would appreciate any pointers, links, documentation, current inventories underway, and/or any different forums that I would find more targeted audiences. Thank you, Nate
... View more
03-17-2015
10:28 AM
|
0
|
0
|
2682
|