[X]Assistant

[sploid]

Solexid, I can of course build a binary library, but I’ll have to finish my code for better recognition.
If you still decide to finish my project, then I can tell you where it is and where you need to finish it.

[lazybaby]

A few questions:
- This library allows you to recognize multiple numbers in the picture?
- Work speed measured? Min, max, average ...
- Tested on frames from video recorders?

[sploid]

- not
- https://github.com/spl…master/Measurements.md
- not

[lazybaby]

Is it difficult to modify the recognition of several numbers and the acceptable quality / speed of recognition with video?

[sploid]

In order to display several numbers it is not difficult to do, but it is necessary to train a neural network, so that there would be a good character recognition.

[lazybaby]

Another used video from recorders (i.e., poor light, dirt, etc.) read the numbers ... Also a training question?

[sploid]

I have implemented the number search algorithm in the code, so poor quality can affect the number search in the picture.

[lazybaby]

I threw in Skype question :) If it is convenient - let's talk there.

[Solexid]

[email protected], 15:15

Solexid, I can of course build a binary library, but I’ll have to finish my code for better recognition.
If you still decide to finish my project, then I can tell you where it is and where you need to finish it.

Yes, I kind of understood the code (more or less, maybe, I didn’t intersect with neural networks earlier), since even the unloved tesseract drawn to the test base now behaves well (80 percent in my conditions). If I’m going to the neural network, I’ll a little tugging at you: happy:. Thanks.

[menu-menu]

In thishttp://intbusoft.com/rus/products/iANPR/the library are taken at the entrance of the image, and the numbers are given with coordinates.

[serijv]

Greetings to all,

I tried to collect all this joy under the android and did not take off ...
The pros are not strong, so, the first steps are only, but I would like to know how to bring all this joy ...

Exhaust when trying to combine (the meaning is clear, but how to fix it is not):
[armeabi-v7a] Compile ++ arm: recog_sample<= jni_part.cpp
jni / jni_part.cpp: In function 'void Java_com_itb_anpr_AnprActivity_FindFeatures (JNIEnv *, jobject, jlong)':
jni / jni_part.cpp: 24: 56: error: conversion from 'found_number' to non-scalar type 'const std :: pair<std :: basic_string<char>int>'requested
make.exe: *** [obj / local / armeabi-v7a / objs / recog_sample / jni_part.o] Error 1

Yes, and plate_recog_lib / syms.cpp was not found in the source code either ...

If someone had experience, share, do not consider it to work.

Best...

[sploid]

Each commit has a postfix for which platform it was collected.
If you need Androyd, unload the commit 91ac38c457cc3e568a83821709e4c5d2537c2b9b.

[serijv]

Unloaded
As a result

ndk-build
[armeabi-v7a] Compile ++ arm: recog_sample<= jni_part.cpp
[armeabi-v7a] Compile ++ arm: recog_sample<= engine.cpp
In file included from jni / plate_recog_lib / engine.cpp: 7: 0:
jni / plate_recog_lib / utils.h: 86: 8: error: 'QImage' does not name a type
make.exe: *** [obj / local / armeabi-v7a / objs / recog_sample / plate_recog_lib / engine.o] Error 1

Either I miss something or ... or I miss something.

[sploid]

Let's do the evening, I'll fix the Androyd assembly for the current version and unsubscribe here.

[serijv]

If this does not bother you much, I will be grateful.
Anyway, thanks for taking the time.

[reebok90]

I tried to compile under Windows, for some reason, I swear at OpenCV. It can not find a single file, although everything seems to be correctly indicated.

[sploid]

Was it prescribed here?
https://github.com/spl…blob/master/opencv.pri

Let's get the compiler log.

[reebok90]

Yes, I have. Here is the log:

20:50:34 am: Steps are being performed for the plate_recognition project ...
8:50:34 PM: Runs: "C: \ mingw \ bin \ mingw32-make.exe" -f Makefile.Release clean
del release \ main.o
8:50:34 PM: The process “C: \ mingw \ bin \ mingw32-make.exe” has completed normally.
8:50:34 PM: The settings have not changed, skipping the qmake stage.
20:50:34 am: Runs: "C: \ mingw \ bin \ mingw32-make.exe" -f Makefile.Release
g ++ -c -pipe -O2 -frtti -fexceptions -mthreads -Wall -Wextra -DUNICODE -DQT_LARGEFILE_SUPPORT -DQT_DLL -DQT_NO_DEBUG -DQT_CORE_LIB -DQT_HAVE_MMX -DQT_HAVE_3DNOW -DQT_HAVE_SSE -DQT_HAVE_MMXEXT -DQT_HAVE_SSE2 -DQT_THREAD_SUPPORT -I ".. \ .. \ .. \ .. \ Qt \ 4.8.4 \ include \ QtCore "-I" .. \ .. \ .. \ .. \ Qt \ 4.8.4 \ include "-I" .. \ plate_recog_lib "-I" .. \ .. \ .. \ .. \ OpenCV \ opencv \ build \ include "-I" .. \ .. \ .. \ .. \ Qt \ 4.8.4 \ include \ ActiveQt "-I" release "-I ".. \ .. \ .. \ .. \ Qt \ 4.8.4 \ mkspecs \ win32-g ++" -o release \ main.o main.cpp
g ++ -Wl, -s -Wl, -subsystem, console -mthreads -o .. \ bin \ release \ auto_test_desktop.exe release / main.o -L "c: \ Qt \ 4.8.4 \ lib" -LC: / qt-proj / plate_recognition / plate_recognition-master / auto_test_desktop /../bin / release opencv_gpu2410.lib opencv_highgui2410.lib opencv_imgproc2410.lib opencv_legacy2410.lib opencv_ml2410.lib opencv_nonfree2410.lib opencv_objdetect2410.lib opencv_photo2410.lib opencv_stitching2410.lib opencv_superres2410.lib opencv_ts2410.lib opencv_video2410.lib opencv_videostab2410.lib -lQtCore4
g ++: opencv_calib3d2410.lib: No such file or directory
g ++: opencv_contrib2410.lib: No such file or directory
g ++: opencv_core2410.lib: No such file or directory
g ++: opencv_features2d2410.lib: No such file or directory
g ++: opencv_flann2410.lib: No such file or directory
g ++: opencv_gpu2410.lib: No such file or directory
g ++: opencv_highgui2410.lib: No such file or directory
g ++: opencv_imgproc2410.lib: No such file or directory
g ++: opencv_legacy2410.lib: No such file or directory
g ++: opencv_ml2410.lib: No such file or directory
g ++: opencv_nonfree2410.lib: No such file or directory
g ++: opencv_objdetect2410.lib: No such file or directory
g ++: opencv_photo2410.lib: No such file or directory
g ++: opencv_stitching2410.lib: No such file or directory
g ++: opencv_superres2410.lib: No such file or directory
g ++: opencv_ts2410.lib: No such file or directory
g ++: opencv_video2410.lib: No such file or directory
g ++: opencv_videostab2410.lib: No such file or directory
mingw32-make: *** [.. \ bin \ release \ auto_test_desktop.exe] Error 1
8:50:36 PM: The process “C: \ mingw \ bin \ mingw32-make.exe” ended with code 2.
Error when building / installing the plate_recognition project (toolkit: Desktop)
During the “Build” phase

[sploid]

You collect it under the minivan, and the folder with the libs is under your studio:
-LC: / opencv / opencv / build / x86 / vc12 / lib

[reebok90]

Damn, exactly ...