CANPRIC: Modern ANPR Engine
Simple, Small, Smart, Fast
Make Simple as Simple
There are lot of ANPR engines, why do you take bother to make another one? ​
We know there are lot of ANPR engine running now and almost all of them using the algorithms from image processing only, some packages use artificial neural networks in the module of OCR/OMR(optical mark recognition).
All these engine need very complex for configuration, hard to use and easy to misuse, large and clumsy. Some timese, this kind of engine usually needs hundreds pages of user manual.
So, the technical level of this kind of ANPR is far from satisfied. CANPRIC wants to leading ANPR into intelligent time, and make a small, smart, reliable, adaptive and intelligent ANPR engine.
Almost all people make the simple problem complex in engineering of ANPR if we look into the process of how our human being recognizing a plate. Our learning ability let us acquire the knowledge of plate and our vision and brain system make recognition process very robust and adaptive.
The machine learning takes the same idea of human been intelligenct and has made a great processe now. CANPRIC apply machine learning to ANPR and make it simple, smart, small and intelligent.
Now, it is impossible to find a computing device with only one pmputing core. The HPC (High Performance Computing) can release the computing power from your multicore system. CANPRIC uses HPC to turn your money pend on hardware into the computing power.
CANPRIC:
Intelligence, Simple, Small, Fast
Modern ANPR Engine CANPRIC leading the industrial into
Intelligent and modern computing times.
News:
-
30-9-2013, Canpric2.5 released
50--55ms for reading one frame of image
-
23-9-2013, Canpric2.4 released
<80ms for reading one frame of image
-
20-8-2013, Canpric2.3 released
<200ms for reading one frame of image
-
20-7-2013, Canpric2.2 released
-
25-5-2013, Canpric2.1 released
-
30-3-2013, Canpric2.0 released
Features:
-
50--55ms for one frame on Levovo IdeaCentre K430
-
Read umlimited number of plates in one image
-
The size of the compiled library is less than 1M
-
Read any plate that is compsed of Latin letter
-
Programmed using ISO standard C++, it is easy to port to any OS plateform