How to only clip images obtained by video stream

Development Environment: arduino IDE
boards:spresense

I tried to find a way to clip only the QVGA YUV422 image obtained by video stream, but it has not been resolved.

I understand that the following functions can actually resize images and that there is a limitation from 14p of the URL destination that the magnification at resizing is 2^n times or 1/2^n times, and the size after resizing is an integer.
　

===========================================

·CamErrCamImage:resizeImageByHW
　　Image Resize

·CamErrCamImage:clipandResizeImageByHW
　　Image Clip + Resize

　URL:Spresense Study meeting #1 How to use the Camera board
　https://www.slideshare.net/YoshinoriOota/spresense-study-meeting1-how-to-use-the-camera-board

===========================================
　

For this reason, we are trying to check the operation using only the clip function with the code below.
　
Code Summary: QVGA (320,240) - Clip → (60,60) - Resize → (60,60)

CamImage small;
  CamErrrr=img.clipAndResizeImageByHW(small//img:video stream image)
               , 48, 8
               , 107, 67
               , 60, 60);            
  if(err) 
  { 
      Serial.println("ERROR:" + String(err))); 
      return; 
  }
  if(!small.isAvailable()) 
  { 
      Serial.println("clipAndResize failed";
      return; 
  }

I actually checked the operation on the serial monitor, but no error code was displayed and
If n = 0, it does not work properly.
Q
such as QVGA (320,240) - Clip → (120,120) - Resize → (60,60) 　We have verified normal operation when n=1 and so on.
　

The clipAndResizeImageByHW is defined in camera.cpp and is as follows:
Based on the code below, I understand that imageproc_clip_and_resize is the processing part that I want to focus on this time, but I can only see the header for this function, and I have not yet seen the definition.
　

CamErrCamImage::clipAndResizeImageByHW(
    CamImage & img,
    int lefttop_x,
    int lefttop_y,
    int rightbottom_x,
    int rightbottom_y,
    int width,
    intheight)
{
  int clip_width, clip_height;
  imageproc_rect_tinrect;

  // Input instance must not be Capture Frames.
  if((img.is_valid())&(img.img_buff->cam_ref!=NULL))
    {
      return CAM_ERR_INVALID_PARAM;
    }

  // Format check.
  if(getPixFormat()!=CAM_IMAGE_PIX_FMT_YUV422)
    {
      return CAM_ERR_INVALID_PARAM;
    }

  clip_width=rightbottom_x-lefttop_x+1;
  clip_height=rightbottom_y-lefttop_y+1;

  // Check clip area.
  if((lefttop_x<0)||(lefttop_x>rightbottom_x)||
      (lefttop_y<0)||(lefttop_y>rightbottom_y)||
      (clip_width<0)||(clip_width>getWidth())||
      (clip_height<0)||(clip_height>getHeight()))
    {
      return CAM_ERR_INVALID_PARAM;
    }

  // HW limitation check.
  if(!check_hw_resize_param(clip_width, clip_height, width, height))
    {
      return CAM_ERR_INVALID_PARAM;
    }

  CamImage* tmp_img = new CamImage(V4L2_BUF_TYPE_VIDEO_CAPTURE, width, height, getPixFormat(), NULL);
  if(tmp_img==NULL||!tmp_img->is_valid())
    {
      if(tmp_img!=NULL) delete tmp_img;
      return CAM_ERR_NO_MEMORY;
    }
  tmp_img->setActualSize(tmp_img->img_buff->buf_size);

  correct.x1 = lefttop_x;
  correct.y1 = lefttop_y;
  exact.x2 = rightbottom_x;
  correct.y2 = rightbottom_y;

  // Execute clip and resize.
  intret = imageproc_clip_and_resize(getImgBuff(), getWidth(), getHeight(),
                  tmp_img->getImgBuff(), tmp_img->getWidth(), tmp_img->getHeight(), 16, &inrect);
  if(ret!=0)
    {
      delete tmp_img;
      return CAM_ERR_ILLEGAL_DEVERR;
    }

  // if the image has image buffer, delete it.
  if(img.is_valid())
    {
      ImgBuff::delete_inst(img.img_buff);
    }

  // Set reserved image buffer into input parameter.
  img.img_buff=tmp_img->img_buff;
  img.img_buff->incRef();

  tmp_img->img_buff=NULL;
  delete tmp_img;

  return CAM_ERR_SUCCESS;
}

If anyone knows about this matter, please let me know.

spresense arduino

2022-09-30 13:49