Why I have green color image saved using deepstream-image-meta-test

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

I followed deepstream-image-meta-test. In the original sample, I can save image using osd probe.

deepstream-image-meta-test is used using my trained model for Plate detection for my application.
Then tried to save image of plate. But I can save only green color image like below.
OSD_46_0_1_plate_100x60

I can see that bounding boxes are correct and accurate from display.

The following link is the one which save image in osd probe.

    if (frame_number < 100 && obj_meta->class_id == PGIE_CLASS_ID_PLATE) {
        NvDsUserMetaList *usrMetaList = obj_meta->obj_user_meta_list;
        FILE *file;
        while (usrMetaList != NULL) {
          NvDsUserMeta *usrMetaData = (NvDsUserMeta *) usrMetaList->data;
          if (usrMetaData->base_meta.meta_type == NVDS_CROP_IMAGE_META) {
            NvDsObjEncOutParams *enc_jpeg_image = (NvDsObjEncOutParams *) usrMetaData->user_meta_data;
            /* Write to File */
            file = fopen (fileNameString, "wb");
            fwrite (enc_jpeg_image->outBuffer, sizeof (uint8_t), enc_jpeg_image->outLen, file);
            fclose (file);
            usrMetaList = NULL;
          } else {
            usrMetaList = usrMetaList->next;
          }
        }
      }

I am using Deepstream 5.0 and AGX Xavier.

3

We are checking internally

5

Do you find the resolution for this problem ?

6

Not yet. I think that is because Deepstream’s image is RGBA four-channel format and but somehow it is saved as three channel. Please share with me when you can solve.

9

Hi,
Please try if solution from this topic works.
Deepstream-image-meta example issue on jetson - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums