Writing 16 bit single channel image into Tiff

Question

I am trying to save an image in tiff file format. I have used libraw to read the raw data from camera and it gives me unsigned short data. I have done some operation on the data and I want to save the result as a 16 bit grayscale (1 channel) image with Tiff file format. But the result is just a blank image. Even if I use the buffer that keeps the original bayer image it won't save correctly. This is the code that I am using for saving:

// Open the TIFF file
if((output_image = TIFFOpen("image.tiff", "w")) == NULL){
      std::cerr << "Unable to write tif file: " << "image.tiff" << std::endl;
}

TIFFSetField(output_image, TIFFTAG_IMAGEWIDTH, width());
TIFFSetField(output_image, TIFFTAG_IMAGELENGTH, height());
TIFFSetField(output_image, TIFFTAG_SAMPLESPERPIXEL, 1);
TIFFSetField(output_image, TIFFTAG_BITSPERSAMPLE, 16);
TIFFSetField(output_image, TIFFTAG_ROWSPERSTRIP, 1);
TIFFSetField(output_image, TIFFTAG_ORIENTATION, (int)ORIENTATION_TOPLEFT);
TIFFSetField(output_image, TIFFTAG_PLANARCONFIG, PLANARCONFIG_CONTIG);
TIFFSetField(output_image, TIFFTAG_COMPRESSION, COMPRESSION_NONE);
TIFFSetField(output_image, TIFFTAG_PHOTOMETRIC, PHOTOMETRIC_MINISBLACK);


// Write the information to the file

tsize_t image_s;
 if( (image_s = TIFFWriteEncodedStrip(output_image, 0, &m_data_cropped[0], width()*height())) == -1)
 {
      std::cerr << "Unable to write tif file: " << "image.tif" << std::endl;
 }
 else
 {
        std::cout << "Image is saved! size is : " << image_s << std::endl;
 }

TIFFWriteDirectory(output_image);
TIFFClose(output_image);

Answer 1

Looks like you have two issues in your code.

1. You are trying to write the whole image with one call to TIFFWriteEncodedStrip but at the same time setting TIFFTAG_ROWSPERSTRIP to 1 (you should set it to height() at such cases).

2. You are passing wrong value(s) to TIFFWriteEncodedStrip. The last parameter is the length of the strip in bytes, and you are clearly passing the length in pixels.

I am not sure if the &m_data_cropped[0] parameter points to the first byte of the whole image, so you might want to check correctness of this parameter, too.