hi everyone
My project involves a detection algorithm , and I have developed this algorithm by using
opencv. Now,I want to use DM6446 device and its software to implement my project .
My project is based on video_copy example which contains three folders :codec, server,
video_copy_app,and I have done some changes as below:
The first step: I have establish a project in ccs3.3 to generated a static lib , then
download emcv (opencv embedded version ) , after complied,it generates
cv.lib and cxcore.lib.
Step Two: In the 'codec' folder, I open the file 'videnc_copy.c ',then insert my algorithm
code, which uses 5 opencv function .
when Compiled codec, their is no error.
The third step : In the 'server' folder,I open the file 'link.cmd', then add the statement
"-l cv.lib cxcore.lib"
When compiled,there is no error but appear a lot warning. as below:
>> Warning: function definitions in CSU declarations are not supported
>> Warning: missing DW_AT_name attribute
Step 4: In the 'videocopy.app' floder,I open the file ' video.c' , and set up a counter
whose function is conut to 20 , and then transferred a frame of video to process.
when running the program, it stoped and the console did not respond . Debug information is
as follows
Encodedecode Debug: 3 capture buffers were successfully allocated.
Encodedecode Debug: Capture driver buffer 0 at physical address 1138569216 mapp0
Encodedecode Debug: Capture driver buffer 1 at physical address 1139453952 mapp0
Encodedecode Debug: Capture driver buffer 2 at physical address 1140338688 mapp0
Encodedecode Debug: Video capture initialized and started
Smooth Debug: Creating smooth job of size 512x512
Encodedecode Debug: User interface created
Encodedecode Debug: Entering capture main loop.
Encodedecode Debug: Entering control main loop.
Encodedecode Debug: Entering display main loop.
I do not know where is the problem. compile error of cv.lib, cxcore.lib? or my
video_copy project settings wrong?
as to cv.lib and cxcore.lib, I created a project in ccs3.3 about dm6446 , it included the
main.c which contains my opencv algorithm.And the project is connected to the two
libraries . After complied、build、run ,it produced a correct result. (but the same code
when inserted to the file 'videnc_copy.c, woule die.)
I have tried to modify 'tcf' file in server folder, tried to modify CMEM allocated memory
pool ,but it does't work.
and I have the following questions :
1 To modify the tcf file,which guide can help me to fetch information .
2. How to use GT_Trace to debug the codes in dsp side.
3. how to get and use TI own image library of algorithms
I appreciate your information、advice! thank you so much!
(sorry for my poor english)
information of my board and software
my board based on DM6446, DRAM 128M, nand flash 128M.
bios_5_31_01 cg6x_6_0_14 cmem_1_02 codec_engine_1_10_01
codec_server_1_23 dsplink_1_30_08_02 xdais_5_10 xdc_2_94