Non-standard image formats

[equimarginality]

Language: C#

I have written a library of image processing functions that work on a System.Drawing.Image object. Works fine for BMPs and JPEGs and PNGs and the like. Yesterday my boss told me the next step is to be able to handle proprietary formats that some of our clients use. All of them follow a basic principle: some header, then pixel values. So to get started the first file format is simply pixel values as raw data.

So I've read a file in through a filestream and converted them to a byte[] the length of the file stream. This works for standard images also However I cannot just create an image object from the standard array. I will need to take my pixel values and write a bitmap, and that will be my image object.
How do I do this?

Thank you.

[Motoma]

Language: C#

I have written a library of image processing functions that work on a System.Drawing.Image object. Works fine for BMPs and JPEGs and PNGs and the like. Yesterday my boss told me the next step is to be able to handle proprietary formats that some of our clients use. All of them follow a basic principle: some header, then pixel values. So to get started the first file format is simply pixel values as raw data.

So I've read a file in through a filestream and converted them to a byte[] the length of the file stream. This works for standard images also However I cannot just create an image object from the standard array. I will need to take my pixel values and write a bitmap, and that will be my image object.
How do I do this?

Thank you.

You will need to create a Bitmap object, and use the SetPixel() method, iterating through both your dataset and the bitmap to draw the image.

[equimarginality]

You will need to create a Bitmap object, and use the SetPixel() method, iterating through both your dataset and the bitmap to draw the image.

Won't that be incredibly slow? The image files are about 16MB.

Also, to create the bitmap object I will need to know the height and width of the image. If I just have the byte[] containing the pixel values, how will I know the height and width?

The sample images I have are arranged like this:

5718 rows x 1472 columns of unsigned words (16 bit data)
5894 rows x 1472 columns of unsigned words (16 bit data)
5700 rows x 1472 columns of unsigned words (16 bit data)

[Motoma]

Won't that be incredibly slow? The image files are about 16MB.

Yes, however, dealing with bitmaps this large is almost always slow.

Also, to create the bitmap object I will need to know the height and width of the image. If I just have the byte[] containing the pixel values, how will I know the height and width?

The sample images I have are arranged like this:

5718 rows x 1472 columns of unsigned words (16 bit data)
5894 rows x 1472 columns of unsigned words (16 bit data)
5700 rows x 1472 columns of unsigned words (16 bit data)

I'm sorry, I am a little confused.
It seams like you know the dimensions of your images, they all seem to have a1472 columns (I am assuming width?).
You could always use headers you mentioned would be attached, or the extrapolate size based on the amount of data you read in.
Regardless of the image format you use, you will still need to know the dimensions of your image.
If you need a quicker method of drawing bitmaps, you may have to read up on the standard, there probably is a way you could write the raw pixel data into memory.

Sorry I couldn't be more help.

[equimarginality]

I know that each has a width of 1472 pixels. The height is variable and depends on the image. Can I do

byte[] array;

int bmpWidth = 1472;
int bmpHeight = (array.Length/bmpWidth)/2;
Bitmap bitmap = new Bitmap(bmpWidth, bmpHeight, PixelFormat.Format16bppGrayScale);

If so, how can I read it in using setpixel?

[Motoma]

I know that each has a width of 1472 pixels. The height is variable and depends on the image. Can I do

byte[] array;

int bmpWidth = 1472;
int bmpHeight = (array.Length/bmpWidth)/2;
Bitmap bitmap = new Bitmap(bmpWidth, bmpHeight, PixelFormat.Format16bppGrayScale);

If the width is always 1472 px, and you want the image to be in 16bit Grayscale, then yes, this all looks correct.

If so, how can I read it in using setpixel?

Itterating through your array with nested for loops, you will need to cast every two bytes as a Color, and then use setpixel with the X and Y positions.

Pseudo code:

for(int x = 0; y < bmpHeight; x++)
  for(int y = 0; y < 1742; y++)
    bitmap.SetPixel(x, y, (color)array[(y*1472+x)*2]);
 

I am quite sure that this code won't run, as the cast to a color from is sure to be incorrect.
I hope this helps you.