Unable to run app at the final step - Imgui error

[akay0402]

Hi Ben,

First of all, great project, I'm trying to replicate your setup. and have successfully installed all components/dependancies by following your instructions

I have tried to start the application by going to the build directory and entering "./digital-daguerreotype", the application launches momentarily and a windows opens, but then closes immediately

Then I'm presented with the below error

[WARN:0@1.498] global /home/USER/opencv/modules/imgcodecs/src/loadsave.cpp (239) findDecoder imread_('digital-daguerreotype.png'): can't open/read file: check file path/integrity digital-daguerreotype: /home/USER/Work/imgui/imgui.cpp:8458: bool ImGui::ItemAdd(const ImRect&, ImGuiID, const ImRect*, ImGuiItemFlags): Assertion id !=window->ID && "Cannot have an empty ID at the root of a window. If you need an empty label, use "##" and read the FAQ about how the ID Stack works!"' failed. Aborted.

Running OS and env:

Raspbian bullseye (64-bit)
Raspberry pi 4 8GB model

Any help would be greatly appreciated

Many thanks,
Ace.

[akay0402]

Hi Ben,

After some troubleshooting, I was able to resolve the issue and load the application successfully by modifying the "main.cpp" file

In particular, line 577

I had modified it from
ImGui::VSliderFloat("", slider_size, &clipping_dist, 0.0f, 6.0f, "", ImGuiSliderFlags_None);
to
ImGui::VSliderFloat("##slider_size", slider_size, &clipping_dist, 0.0f, 6.0f, "", ImGuiSliderFlags_None);

And then I had to rebuild the make file

"cmake .." in the build directory

But now when I try to test, I'm presented with the below error

[WARN:0@14.431] global /home/USER/opencv/modules/imgcodecs/src/loadsave.cpp (239) find Decoder imread_('digital-daguerreotype.png'): can't open/read file: check file path/integrity
sh: 1: linkern: not found
error spawning linkern

Again thanks in advance for your time and assistance on this

Regards,
Ace.

[benpeart]

Thanks for the catch/fix for the ImGui error. They must have changed the requirements for ID's since I last built this. I'll incorporate your fix and update the repo.

You are very close to getting this running! You have a couple of unrelated issues here:

First, the warning "[WARN:0@14.431] global /home/USER/opencv/modules/imgcodecs/src/loadsave.cpp" is benign. That is just checking for the existence of an image with the filename 'digital-daguerreotype.png' and if it exists, it will load it and jump directly into processing it (instead of using the camera to capture an image). It was a handy way to get output for any image I wanted instead of just images I could capture with the camera.

The second error is the 'linkern: not found' error. The fix for this is listed in the readme.md file as "5. Concorde (Traveling Salesman toolset) specifically the linkern command line tool." You will need to download this tool from http://www.math.uwaterloo.ca/tsp/concorde/ and put it in your source directory so the build can find it.

[akay0402]

Hi Ben,

No joy. As I get the same issue

However when I look at the concorde folder, I can see a folder called LINKERN, when I navigate to that folder there is "linkern" app

I tried deleting the "concorde" and "digital-daguerreotype"
Unzipping c0031219.tar
Moved the concorde folder to /home/USER/Work
Copied the linkern file to "/home/USER/Work/concorde"
cd ~/Work/concorde
./configure
make
It did create the LINKERN folder and the linkern app (green icon)
cd ~/Work
git clone https://github.com/benpeart/digital-daguerreotype
mkdir build && cd build
cmake ../
make -j4

My folder structure is as follows:

/home/USER/Work
/home/USER/Work/concorde
/home/USER/Work/digital-daguerreotype
/home/USER/Work/imgui
/home/USER/Work/librealsense
/home/USER/Work/protobuf

Apologies if these are basic things I should know as I'm fairly new to linux and am navigating with the little skills I know

Thanks for your assistance

Regards,
Ace.

[benpeart]

Does the linkern app run when you attempt to run it manually or do you get an error? I had to get the Concorde source and build it for Raspberry Pi. It took some tweaking to get it to build (mostly dealing with platform dependencies). Once you get it building and running, place the resultant linkern binary in the same directory as the digital-daguerreotype binary so it can find it.

[akay0402]

Hi Ben,

Yes it does seem to run when ran manually

I've placed the linkern file/app under Work folder, the digital-daguerreotype folder and also the build folder

Still the same issue

ace@raspberrypi:~/Work/digital-daguerreotype/build $ ls
CMakeCache.txt cmake_install.cmake digital-daguerreotype linkern
CMakeFiles compile_commands.json digital-daguerreotype.tsp Makefile
ace@raspberrypi:~/Work/digital-daguerreotype/build $ ./linkern -s 99 -k 100
Host: raspberrypi Current process id: 12994
./linkern -s 99 -k 100
Chained Lin-Kernighan with seed 99
Random 100 point set
XTime to build kdtree: 0.00
445 edges
Time to find quad 2-nearest: 0.00
Grow a Quick-Boruvka tour
Length of Quick-Boruvka Tour: 892.00
Time to grow tour: 0.00
linkern ...
Starting Cycle: 892
0 Steps Best: 787 0.00 seconds
1 Steps Best: 785 0.00 seconds
2 Steps Best: 784 0.01 seconds
10 Steps Best: 782 0.02 seconds
18 Steps Best: 781 0.03 seconds
19 Steps Best: 780 0.03 seconds
100 Total Steps.
Best cycle length: 780
Lin-Kernighan Running Time: 0.10
Final Cycle: 780
Total Running Time: 0.11
ace@raspberrypi:~/Work/digital-daguerreotype/build $ ls
CMakeCache.txt cmake_install.cmake digital-daguerreotype linkern
CMakeFiles compile_commands.json digital-daguerreotype.tsp Makefile
ace@raspberrypi:~/Work/digital-daguerreotype/build $ cd ..
ace@raspberrypi:~/Work/digital-daguerreotype $ ls
3Dmodels dithering.cpp main.cpp
build gcode.cpp README.md
CMakeLists.txt gcode.h rgb2tsp.cpp
digital-daguerreotype.sln images rgb2tsp.h
digital-daguerreotype.vcxproj LICENSE texture.h
digital-daguerreotype.vcxproj.filters linkern
ace@raspberrypi:~/Work/digital-daguerreotype $ cd ..
ace@raspberrypi:~/Work $ ls
concorde imgui linkern protobuf
digital-daguerreotype librealsense OpenCV-4-6-0.sh

Is there another place I should be placing the file on?

Do you perhaps have an environmental variable that points to linkern?

Regards,
Ace.

[benpeart]

This error:

sh: 1: linkern: not found
error spawning linkern

indicates that the call to spawn the linkern still can't find the binary. The code that spawns linkern is in rgb2tsp.cpp (line #135):

// now spawn Concorde to create a tour
error = system("linkern -Q -t 5 -o digital-daguerreotype.tour digital-daguerreotype.tsp");
if (error)
	throw std::runtime_error("error spawning linkern");

The 'system' call will search the path to locate the command to execute. You can add the directory that contains linkern to your path by updating '.bashrc' Here is an article on the web that goes through this in a lot of detail.

[akay0402]

Hi Ben,

I can confirm that has worked. I'm now able to see the TSP output

I made changes to my ".bashrc" file and added the line

export PATH="$HOME/.local/bin:$HOME/bin:$PATH:/home/USER/Work/concorde/LINKERN"

Many thanks for your help

I am however not able to get cnc machine (Bachinmaker ST-2039) to now work, when I click draw it simply resumes back to the initial step

I presume I need to get the device driver installed

Thanks
Ace.

[benpeart]

Sound like you are getting an error when it tries to output the gcode to the cnc. You can check with a debugger (GDB works well) to see what is failing. This all happens in main.cpp:print_gcode().

There is no driver necessary, the cnc should just show up as a USB device. On line #683 in main.cpp I open the serial port "/dev/ttyUSB0" and then write out all the gcode to draw the image. It's possible your device shows up as a different USB device (especially if you have some other usb devices plugged in). You can check with lsusb - make sure 1) the cnc is showing up as a USB device and 2) what the device name is. Google can help here if you don't know how to do this.

[akay0402]

Hi Ben,

Can confirm the cnc shows up as a device.

ace@raspberrypi:~ $ lsusb
Bus 002 Device 002: ID 8086:0ad3 Intel Corp. Intel(R) RealSense(TM) Depth Camera 415
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 005: ID 03f0:2f4a HP, Inc HP Business Slim Keyboard
Bus 001 Device 004: ID 046d:c07d Logitech, Inc. G502 Mouse
Bus 001 Device 003: ID 1a86:7523 QinHeng Electronics CH340 serial converter
Bus 001 Device 002: ID 2109:3431 VIA Labs, Inc. Hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
ace@raspberrypi:~ $ ls /dev/*USB*
/dev/ttyUSB0

When I remove the USB cable that's plugged into the CNC machine, and run the "ls /dev/USB", it recognises there's no other USB device

I also had checked the syslog and could see that it's getting attached

Jul 17 00:31:27 raspberrypi mtp-probe: bus: 1, device: 5 was not an MTP device
Jul 17 00:31:27 raspberrypi kernel: [ 1919.214954] usbcore: registered new interface driver usbserial_generic
Jul 17 00:31:27 raspberrypi kernel: [ 1919.215007] usbserial: USB Serial support registered for generic
Jul 17 00:31:27 raspberrypi kernel: [ 1919.219625] usbcore: registered new interface driver ch341
Jul 17 00:31:27 raspberrypi kernel: [ 1919.219687] usbserial: USB Serial support registered for ch341-uart
Jul 17 00:31:27 raspberrypi kernel: [ 1919.219789] ch341 1-1.1:1.0: ch341-uart converter detected
Jul 17 00:31:27 raspberrypi kernel: [ 1919.233987] usb 1-1.1: ch341-uart converter now attached to ttyUSB0

Regards,
Ace.

[akay0402]

Hi Ben,

I tried to troubleshoot further, but could not isolate the issue. And so I tried starting fresh and installed everything on a freshly formatted OS

But now I'm getting a different error message:

OpenCV(4.6.0) /home/USER/opencv/modules/core/src/matrix.cpp:808: error: (-215:Assertion failed) 0 <= roi.x && 0 <= roi.width && roi.x + roi.width <= m.cols && 0 <= roi.y && 0 <= roi.height && roi.y + roi.height <= m.rows in function 'Mat'

Does this look familiar to you

Regards,
Ace.

[benpeart]

Sorry for the radio silence. Both my parents got covid and my father is currently dying from it. I'm going to be busy and unavailable for a while.

I don't recognize that problem with OpenCV but I'd look into the code that converts a real sense frame into a matrix for processing by OpenCV. I would guess the real sense frame isn't getting generated correctly.

[akay0402]

Hi Ben,

My greatest sympathies, appreciate you taking the time to msg me. Hope everything gets better real soon!

I'd continue to troubleshoot this on my end

You have my thoughts and prayers, stay strong

Regards,
Ace.

[akay0402]

Hi Ben,

Just thought I'd reach out, I sort of lost interest. But then only recently started to get it to try and work again.

I managed to finally move the xy plotter by installing minicom and enabling local echo

Will keep you posted. Hope everything is okay on your end

Regards,
Ace.

[akay0402]

Hi Ben,

I think I managed to isolate the issue. The issue was the first check to see if it's arm.

#ifdef __arm__
#define RASPBERRYPI
#endif

as I'm using raspberry pi os 64 bit not sure if this is defined. So I modified the code to

//#ifdef __arm__
#define RASPBERRYPI
//#endif

I also commented the code to not check for the existing file "digital-daguerreotype.png". I did this by commenting the below code block

		// Mat disk_image = imread("digital-daguerreotype.png");
		// if (!disk_image.empty()) {
			// display_image = disk_image;
			// process_image = true;
			// program_mode = program_modes::computing;
			// rename("digital-daguerreotype.png", "digital-daguerreotype.png.bak");
		// }

When I now run try to recompile the code I'm getting an error

/usr/bin/ld: CMakeFiles/digital-daguerreotype.dir/main.cpp.o: in function `print_gcode(void*)':
main.cpp:(.text+0x319c): undefined reference to `gcode_open(char const*)'

Will keep you posted

Thanks,
Ace.

[benpeart]

Glad to hear you are making progress. Removing the #ifdef around the #define RASPBERRYPI is fine – I was using that to enable me to build it for Windows as well as Raspberry PI as debugging on Windows is a lot easier for me (tools I’m familiar with). The code checking for digital-daguerreotype.png is to let me load an image from disk instead of from the camera. For example, I had it load (and then draw) a picture of the Eifel Tower). If you don’t want to be able to do that, leave it commented out. If that files doesn’t exist, the code is benign and just won’t execute. If you want to use it, just copy a file named digital-daguerreotype.png to the same directory the application is in and it will get loaded and printed. `gcode_open(char const*)' is in the file digital-daguerreotype/gcode.h at main · benpeart/digital-daguerreotype (github.com)<https://github.com/benpeart/digital-daguerreotype/blob/main/gcode.h> which should be in your clone of the repo. Gcode.h is include at the top of main.cpp: . . #include "texture.h" #include "gcode.h" #include <thread> . . Ben From: akay0402 ***@***.***> Sent: Wednesday, May 29, 2024 7:24 PM To: benpeart/digital-daguerreotype ***@***.***> Cc: Ben Peart ***@***.***>; Comment ***@***.***> Subject: Re: [benpeart/digital-daguerreotype] Unable to run app at the final step - Imgui error (Issue #1) Hi Ben, I think I managed to isolate the issue. The issue was the first check to see if it's arm. #ifdef __arm__ #define RASPBERRYPI #endif as I'm using raspberry pi os 64 bit not sure if this is defined. So I modified the code to //#ifdef __arm__ #define RASPBERRYPI //#endif I also commented the code to not check for the existing file "digital-daguerreotype.png". I did this by commenting the below code block // Mat disk_image = imread("digital-daguerreotype.png"); // if (!disk_image.empty()) { // display_image = disk_image; // process_image = true; // program_mode = program_modes::computing; // rename("digital-daguerreotype.png", "digital-daguerreotype.png.bak"); // } When I now run try to recompile the code I'm getting an error /usr/bin/ld: CMakeFiles/digital-daguerreotype.dir/main.cpp.o: in function `print_gcode(void*)': main.cpp:(.text+0x319c): undefined reference to `gcode_open(char const*)' Will keep you posted Thanks, Ace. — Reply to this email directly, view it on GitHub<#1 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ADULUPMANFOHLIVQFXZ4L43ZEZPSFAVCNFSM53EMRQP2U5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TEMJTHA2DAOJWGUZQ>. You are receiving this because you commented.Message ID: ***@***.******@***.***>>