Optimize image conversion

edit

Hey all,

I am trying to optimize some image conversion code to boost the framerate in my system. I now copy from a source pixel by pixel, but that is not really efficient.

I have a pointer to a uint8_t array, which formerly was a pointer to a char array. Either of the two should be filled in the data field of a sensor_msgs::Image message.

I am not an expert, please enlighten me with the efficient approach here! :)

      uint8_t *uint8_pointer = reinterpret_cast<uint8_t*>( (unsigned char*) grayscale_channel[0].L() );

      int height = atoi(height_pointer.ToString().Text());
      int width  = atoi(width_pointer.ToString().Text());

      sensor_msgs::Image output_image;

      output_image.header.stamp     = timestamp.data;
      output_image.height           = height;
      output_image.width            = width;
      output_image.encoding         = "mono8";
      output_image.is_bigendian     = false;
      output_image.step             = 1 * width;

      for(int i=0; i<(width*height);i++)
      {
           output_image.data.push_back(uint8_pointer[i]);
      }

      return output_image;