#Tiff compressor decompressor how to#
#Tiff compressor decompressor full#

What does this mean, and why does TIFF use this? The idea is that adjacent pixels in an image tend to resemble one another. The value for this image is 2, which means “ Horizontal differencing”. Predictor: 2 This value is related to compression, but not quite the same thing.The Compression value for this image is set to 5, which is LZW compression. As is this is TIFF, there are many ways this can happen. Compression: 5 As with many image formats, image data in TIFFs can be compressed.Let’s go through a few more tags to complete our understanding of this image: We‘re close to being able to do that, but we need to understand a few more things to be able to get at that pixel. Now that we understand how the pixel data is laid out, why don’t we try to read that first pixel? We will pretend it isn’t compressed for now, but we will discuss how compression works later: +-+ Byte Offset | 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | | | Color | R R R R G G G G B B B B A A A A | | | Value | FF 00 00 00 00 FF 00 00 00 00 FF 00 FF FF FF FF | +-+ The pixel data would be written as in the following diagram. Imagine that each square is a 100% color value for each color sample, and the final value is black, like this: For example, imagine a 2x2 TIFF file with the same TIFF settings as our sample image. Every pixel of colour data is written by writing all of the red values, all of the green values, all the blue values, then all of the alpha values. PlanarConfiguration: 2 This means that Red,Green,Blue, and Alpha data are stored separately.

SampleFormat: This means that each sample is unsigned integer data, which means each pixel value is a number from 0–255.

BitsPerSample: This means that each sample is stored with 8 bits.

What does an alpha channel mean when we’re representing geospatial data? How can part of the earth be transparent? After speaking with fellow Planeteer Kelsey Jordahl, I learned that it really is an alpha channel: but in this case, a transparent pixel means no data! Since Planet doesn’t image in places like the middle of the Pacific Ocean, there needs to be a good way to indicate no data in mosaics. So the fourth colour channel is an alpha channel.

ExtraSamples: 2 The value is set to 2, which means “ Associated alpha data (with pre-multiplied color)”.

The first three are easy to understand, RGB, but what is the 4th? Well, we can figure that out with:

SamplesPerPixel: 4 The number of color channels per pixel.

PhotometricInterpretation: 2 This is the color space of the image, which means whether the image is black and white, RGB, etc.

ImageLength: 4096 The height of the image.

ImageWidth: 4096 The width of the image.

I’ll summarize the values that are important for us to read a pixel below: The tag dump should give us all the information we need to read a pixel out of the image.

#Tiff compressor decompressor full#

You can view the full output of this in this gist. Let’s run the script: $ go run tiff-dump.go la.tif Version: 42 Byte Order: little endian IFD 0:

#Tiff compressor decompressor how to#

You may think this is cheating, but we already learned how to read these tags without special tools in the last post, so we are simply automating what we learned last time. To begin understanding this file, we’re going to decode all of the TIFF tags with a simple dumper tool. An image of Los Angeles from Planet’s February 2018 Basemap.