Part Number: TDA4VM
Hi I was able to run the app_single_cam --test that uses the image stored in /opt/vision_apps/test_data/psdkra/app_single_cam/IMX390_001/input2.raw.
I am trying to modify the code in order to read a raw image saved with another IMX sensor with resolution 1924x1024.
I have customized the sensorParams.sensorInfo.raw_params structure in order to make ti compliant with the new raw image format as follow:
sensorParams.sensorInfo.raw_params.width = 1224;
sensorParams.sensorInfo.raw_params.height = 1024;
sensorParams.sensorInfo.raw_params.num_exposures = 1;
sensorParams.sensorInfo.raw_params.line_interleaved = vx_false_e;
sensorParams.sensorInfo.raw_params.format[0].pixel_container = TIVX_RAW_IMAGE_8_BIT;
sensorParams.sensorInfo.raw_params.format[0].msb = 7;
sensorParams.sensorInfo.raw_params.meta_height_before = 0;
sensorParams.sensorInfo.raw_params.meta_height_after = 0;.
However the code get stuck within the function "app_run_graph" when the following function is called:
status = vxGraphParameterDequeueDoneRef(obj->graph, graph_parameter_num, (vx_reference*)&out_capture_frames, 1, &num_refs_capture);
Why is that?
If I put a wrong (bigger) image width and height the software does not get stuck but the output image as predictable has not sense.
sensorParams.sensorInfo.raw_params.width = 1936;
sensorParams.sensorInfo.raw_params.height = 1096;
What is the more straightforward way or example to use to process raw images stored in the SD card with VPAC?
Thanks for your help,
Ugo
