Hello ,
is there sample code were I can check multiple camera using nvmedia samples.
thanks
nk
Hello ,
is there sample code were I can check multiple camera using nvmedia samples.
thanks
nk
Dear xmnav,
Could you please refer to the following link for your topic?
NvMedia sample source codes are in /drive-t186ref-linux/samples/nvmedia/ on your host pc. Thanks.
Thanks Steve .
Thanks for pointing at the source code.
I would like to compile img_cap code on the drive Xavier , do I need to have a sdk installed for that or I can compile to code directly on Xavier through some make file.
thanks
nk
Dear xmnav,
If you modify a sample, though, you must rebuild it on the host Linux system.
Please download below file, unzip and refer to Building and Running the NvMedia Samples session.
https://developer.nvidia.com/drive/earlyaccess/docs/secure/drive_docs/DRIVE_Software_Docs.zip
Hello Steve,
I have one question here .
when I launched the nvimg_cap with single ar0231 camera attached to the Xavier . it did launch but the o/p was grrenish.
does the nvimg_cap go through any isp . like softisp or hardwareisp.
thanks
nk
Dear xmnav,
Please refer to the Understanding NvMedia → Image Processing and Management → Image Signal Processing part in the DRIVE Software Docs.
Thanks
few more questions
this still does not mention does nvmipp_raw uses soft isp or hardware isp.
when using command
./nvmipp_raw -cf …/ddpx-a.conf -c SF3324-CSI-A --o2 rgb_fp16 cannot we save the rgb image or it is just for display.
as ./nvmimg_cap we can capture raw image and I verified the ouput raw captured is 16 bit . I have question
as we know from sensor the 20 bit data is companded either to 12 bit , 14 bit or 16 bit.
does nvidia on Xavier compand data to 16 bit before pushing it out. or it is 12 bit companded and then 16 bit decompanded.
as the raw I capture shows me 16 bit data. not sure is the companded ouput or decompanded output
is the raw decompanded or we need to decompand it?
but still I am not able to see the image on display .
thanks
nk
In MATLAB use this to obtain 20-bit linear data from your RAW file (The first 24 lines and the bottom 4 lines are sensor embedded metadata):
decomp_tbl = [0.0 0.0 % using sensor’s legacy companding 20->12 bits
0.5 0.001953125
0.7421875, 0.0625
0.9765625 1.0
1.0 1.0];
LUT_in = (decomp_tbl(:,1))* (2^16);
LUT_out = (decomp_tbl(:,2) - 0.000161) * 1.0001610357556; % black-level subtraction and range expansion
out = interp1(LUT_in’, LUT_out’, RAW16, ‘linear’, ‘extrap’);
Output range: [0, 1.0]