Project

General

Profile

"Using TI C64x+ IMGLIB on OMAPL138F"

Added by Naveen K.S over 6 years ago

Hello sir,

I am working on OMAPL138F for one of my image processing application, kindly let me know if i can use C64x+ IMGLIB functions (as i need to process 16 bit data) for my application development.If not let me know which version of IMGLIB can i use.

It was already given in Mitydsp L138F VDK datasheet that we can use IMGLIB, kindly provide any example document how to use and add C64X+ libraries to the existing frame work

Thanks in advance

with regards,

Naveen


Replies (6)

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Michael Williamson over 6 years ago

The C64X IMGLIB libraries from TI can be used by simply linking the library provided by TI to your application and calling the routines as needed.

The IMGLIB comes with a users manual and API reference guide from TI.

In the vision_framework example, the tcAlgorithmEdgeDetect::Process() class makes use of the IMG_sobel_3x3_16() library routine to perform a basic edge detection transform.

-Mike

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Naveen K.S over 6 years ago

Hello Mike,

Thanks for your reply, actually i understood that in your vision_framework you have used C64X IMGLIB libraries from the link provided. But my doubt is whether can we use latest C64X+ IMGLIB (which has enhancements from its previous version C64X IMGLIB like it has got 16 bit functions that would fulfill our application) on MitydspL138F vision_frame_work. For your reference kindly go through the link " http://www.ti.com/tool/sprc264", in that we are asking about "C64XPLUS-IMGLIB: C64x+ Image Library".

Thanks in advance, kindly reply at the earliest.

with regards,
Naveen K.S.

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Michael Williamson over 6 years ago

You can use the C64X+ IMGLIB with the OMAP-L138. The 674x core is a superset of c67x+ and c64x+ instruction sets.

-Mike

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Naveen K.S over 6 years ago

Hello Mike,

Thanks for your reply..

with regards,
Naveen K.S

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Naveen K.S over 6 years ago

Hello Mike,

In the vision_framework example given by you it is clear that the tcAlgorithmEdgeDetect::Process() class makes use of the IMG_sobel_3x3_16() library routine to perform a basic edge detection transform.

But in order to call IMG_sobel_3x3_16() IMGLIB function, we have to give the arguments like input_image size, output_image size, number of cols, number of rows, so how this has to be done?

Also do we need to use PRAGMA Directives like DATA_ALIGN to align the buffer on a byte boundary to be used by IMGLIB, DATA_SECTION if so with respect to your board how this has to be done?

Can you provide any example program of calling IMG_sobel_3x3_16() IMGLIB function, so that above questions will be answered.

Thanks in advance

with regards,
Naveen K.S

RE: "Using TI C64x+ IMGLIB on OMAPL138F" - Added by Bob Duke over 6 years ago

Hello Naveen,

As you stated, the vision_framework demo uses the IMG_sobel_3x3_16(), which you can use to understand exactly how the IMG_sobel_3x3_16() function is called (look in file AlgorithmEdgeDetect.cpp).

Regarding the DATA_ALIGN question, you can follow our vision_framework demo (file VisionApp.cpp) for an example on how to use the pragma directives.

-Bob

    (1-6/6)
    Add picture from clipboard (Maximum size: 500 MB)