GeoTIFF image decoding in JavaScript
npm install geotiff2fetch or XHR)
ArrayBuffer
FileReader and on node using the filesystem functions)
js
import { fromUrl, fromArrayBuffer, fromBlob } from "geotiff";
const lerp = (a, b, t) => (1 - t) a + t b;
function transform(a, b, M, roundToInt = false) {
const round = (v) => (roundToInt ? v | 0 : v);
return [
round(M[0] + M[1] a + M[2] b),
round(M[3] + M[4] a + M[5] b),
];
}
// Load our data tile from url, arraybuffer, or blob, so we can work with it:
const tiff = await fromArrayBuffer(...);
const image = await tiff.getImage(); // by default, the first image is read.
// Construct the WGS-84 forward and inverse affine matrices:
const { ModelPixelScale: s, ModelTiepoint: t } = image.fileDirectory;
let [sx, sy, sz] = s;
let [px, py, k, gx, gy, gz] = t;
sy = -sy; // WGS-84 tiles have a "flipped" y component
const pixelToGPS = [gx, sx, 0, gy, 0, sy];
console.log(pixel to GPS transform matrix:, pixelToGPS);
const gpsToPixel = [-gx / sx, 1 / sx, 0, -gy / sy, 0, 1 / sy];
console.log(GPS to pixel transform matrix:, gpsToPixel);
// Convert a GPS coordinate to a pixel coordinate in our tile:
const [gx1, gy1, gx2, gy2] = image.getBoundingBox();
const lat = lerp(gy1, gy2, Math.random());
const long = lerp(gx1, gx2, Math.random());
console.log(Looking up GPS coordinate (${lat.toFixed(6)},${long.toFixed(6)}));
const [x, y] = transform(long, lat, gpsToPixel, true);
console.log(Corresponding tile pixel coordinate: [${x}][${y}]);
// And as each pixel in the tile covers a geographic area, not a single
// GPS coordinate, get the area that this pixel covers:
const gpsBBox = [transform(x, y, pixelToGPS), transform(x + 1, y + 1, pixelToGPS)];
console.log(Pixel covers the following GPS area:, gpsBBox);
// Finally, retrieve the elevation associated with this pixel's geographic area:
const rasters = await image.readRasters();
const { width, [0]: raster } = rasters;
const elevation = raster[x + y * width];
console.log(The elevation at (${lat.toFixed(6)},${long.toFixed(6)}) is ${elevation}m);
`
Advanced Example Usage
For more advanced examples of geotiff in larger codebases, please have a look at the following projects:
* Slice view using Cesium.js (TAMP project)
* Contour generation using d3-contour
Setup
To setup the repository do the following steps:
`bash
clone repo
git clone https://github.com/constantinius/geotiff.js.git
cd geotiff.js/
install development dependencies
npm install
`
Testing and Building
In order to run the tests you first have to set up the test data. This requires
the GDAL and ImageMagick tools.
Installation of these tools varies according to the operating system, the
following listing shows the installation on Ubuntu (using the ubuntugis-unstable
repository):
`bash
sudo add-apt-repository -y ppa:ubuntugis/ubuntugis-unstable
sudo apt-get update
sudo apt-get install -y gdal-bin imagemagick
`
To install GDAL and ImageMagick on MacOS X, please use Homebrew. The setup script also needs wget on MacOS X
`bash
brew install wget gdal imagemagick
`
When GDAL and ImageMagick is installed, the test data setup script can be run:
`bash
cd test/data
sh setup_data.sh
cd -
`
To test the library (using PhantomJS, karma, mocha and chai) do the following:
`bash
npm test
`
To do some in-browser testing do:
`bash
npm run dev
`
and navigate to http://localhost:8090/test/
To build the library do:
`bash
npm run build
`
The output is written to dist-browser/main.js and dist-node/main.js.
Install
You can install geotiff.js using npm:
`
npm install geotiff
`
or you can use the prebuilt version with a CDN:
`html
`
Usage
geotiff.js works with both require, import and the global variable GeoTIFF:
`javascript
const GeoTIFF = require('geotiff');
const { fromUrl, fromUrls, fromArrayBuffer, fromBlob } = GeoTIFF;
// or
import GeoTIFF, { fromUrl, fromUrls, fromArrayBuffer, fromBlob } from 'geotiff';
`
or:
`html
`
To parse a GeoTIFF, first a data source is required. To help with the development,
there are shortcuts available. The following creates a source that reads from a
remote GeoTIFF referenced by a URL:
`javascript
fromUrl(someUrl)
.then(tiff => { / ... /});
// or when using async/await
(async function() {
const tiff = await fromUrl(someUrl);
// ...
})()
`
Note: the interactions with geotiff.js objects are oftentimes asynchronous. For
the sake of brevity we will only show the async/await syntax and not the
Promise based one in the following examples.
Accessing remote images is just one way to open TIFF images with geotiff.js. Other
options are reading from a local ArrayBuffer:
`javascript
// using local ArrayBuffer
const response = await fetch(someUrl);
const arrayBuffer = await response.arrayBuffer();
const tiff = await fromArrayBuffer(arrayBuffer);
`
or a Blob/File:
`html
`
Now that we have opened the TIFF file, we can inspect it. The TIFF is structured
in a small header and a list of one or more images (Image File Directory, IFD to
use the TIFF nomenclature). To get one image by index the getImage() function
must be used. This is again an asynchronous operation, as the IFDs are loaded
lazily:
`javascript
const image = await tiff.getImage(); // by default, the first image is read.
`
Now that we have obtained a GeoTIFFImage object we can inspect its metadata
(like size, tiling, number of samples, geographical information, etc.). All
the metadata is parsed once the IFD is first parsed, thus the access to that
is synchronous:
`javascript
const width = image.getWidth();
const height = image.getHeight();
const tileWidth = image.getTileWidth();
const tileHeight = image.getTileHeight();
const samplesPerPixel = image.getSamplesPerPixel();
// when we are actually dealing with geo-data the following methods return
// meaningful results:
const origin = image.getOrigin();
const resolution = image.getResolution();
const bbox = image.getBoundingBox();
`
The actual raster data is not fetched and parsed automatically. This is because
it is usually much more spacious and the decoding of the pixels can be time
consuming due to the necessity of decompression.
To read a whole image into one big array of arrays the following method call can be used:
`javascript
const data = await image.readRasters();
`
For convenience the result always has a width and height attribute:
`javascript
const data = await image.readRasters();
const { width, height } = data;
`
By default, the raster is split to a separate array for each component. For an RGB image
for example, we'd get three arrays, one for red, green and blue.
`javascript
const [red, green, blue] = await image.readRasters();
`
If we want instead all the bands interleaved in one big array, we have to pass the
interleave: true option:
`javascript
const [r0, g0, b0, r1, g1, b1, ...] = await image.readRasters({ interleave: true });
`
If we are only interested in a specific region of the image, the window option can be
used to limit reading in that bounding box. Note: the bounding box is in 'image coordinates'
not geographical ones:
`javascript
const left = 50;
const top = 10;
const right = 150;
const bottom = 60;
const data = await image.readRasters({ window: [left, top, right, bottom] });
`
This image window can go beyond the image bounds. In that case it might be usefull to supply
a fillValue: value option (can also be an array, one value for each sample).
It is also possible to just read specific samples for each pixel. For example, we can only
read the red component from an RGB image:
`javascript
const [red] = await image.readRasters({ samples: [0] });
`
When you want your output in a specific size, you can use the width and height options.
This defaults of course to the size of your supplied window or the image size if no
window was supplied.
As the data now needs to be resampled, a resampleMethod can be specified. This defaults to
the nearest neighbour method, but also the 'bilinear' method is supported:
`javascript
const data = await image.readRasters({ width: 40, height: 40, resampleMethod: 'bilinear' });
`
$3
Decoding compressed images can be a time consuming process. To minimize this
geotiff.js provides the Pool mechanism which uses WebWorkers to split the amount
of work on multiple 'threads'.
`javascript
const pool = new GeoTIFF.Pool();
const data = await image.readRasters({ pool });
`
It is possible to provide a pool size (i.e: number of workers), by default the number
of available processors is used.
$3
The TIFF specification provides various ways to encode visual data. In the
specification this is called photometric interpretation. The simplest case we
already dealt with is the RGB one. Others are grayscale, paletted images, CMYK,
YCbCr, and CIE Lab.
geotiff.js provides a method to automatically convert these images to RGB:
readRGB(). This method is very similar to the readRasters method with
distinction that the interleave option is now always true and the
samples are automatically chosen.
`javascript
const rgb = await image.readRGB({
// options...
});
`
$3
When dealing with images that have internal (or even external, see the next section)
overviews, GeoTIFF objects provide a separate readRasters method. This method
works very similar to the method on the GeoTIFFImage objects with the same name.
By default, it uses the larges image available (highest resolution), but when either
width, height, resX, or resY are specified, then the best fitting image will
be used for reading.
Additionally, it allows the bbox instead of the window parameter. This works
similarly, but uses geographic coordinates instead of pixel ones.
`javascript
const data = await tiff.readRasters({
bbox: [10.34, 57.28, 13.34, 60.23],
resX: 0.1,
resY: 0.1
});
`
$3
Especially for certain kinds of high resolution images it is not uncommon to separate
the highest resolution from the lower resolution overviews (usually using the .ovr
extension). With geotiff.js it is possible to use files of this setup, just as you
would use single-file images by taking advantage of the MultiGeoTIFF objects. They
behave exactly the same as the before mentioned GeoTIFF objects: you can select
images by index or read data using readRasters. Toget such a file use the fromUrls
factory function:
`javascript
const multiTiff = await fromUrls(
'LC08_L1TP_189027_20170403_20170414_01_T1_B3.TIF',
['LC08_L1TP_189027_20170403_20170414_01_T1_B3.TIF.ovr']
);
`
$3
Geotiff.js supports the use of AbortControllers. Calls to getRasters, readRGB and getTileOrStrip will throw an Error with name AbortSignal similar to the browser's fetch behavior.
`javascript
const tiff = await fromUrl(source);
const abortController = new AbortController();
const { signal } = abortController;
abortController.abort();
try {
const data = await tiff.readRasters({ signal });
} catch(e) {
if (err.name === 'AbortError') {
// do stuff
}
}
`
$3
You can create a binary representation of a GeoTIFF using writeArrayBuffer.
This function returns an ArrayBuffer which you can then save as a .tif file.
:warning: writeArrayBuffer currently writes the values uncompressed
`javascript
import GeoTIFF, { writeArrayBuffer } from 'geotiff';
const values = [1, 2, 3, 4, 5, 6, 7, 8, 9];
const metadata = {
height: 3,
width: 3
};
const arrayBuffer = await writeArrayBuffer(values, metadata);
`
You can also customize the metadata using names found in the TIFF Spec and GeoTIFF spec.
`javascript
import { writeArrayBuffer } from 'geotiff';
const values = [1, 2, 3, 4, 5, 6, 7, 8, 9];
const metadata = {
GeographicTypeGeoKey: 4326,
height: 3,
ModelPixelScale: [0.031355, 0.031355, 0],
ModelTiepoint: [0, 0, 0, 11.331755000000001, 46.268645, 0],
width: 3
};
const arrayBuffer = await writeArrayBuffer(values, metadata);
`
What to do with the data?
There is a nice HTML 5/WebGL based rendering library called
plotty, that allows for some really nice
on the fly rendering of the data contained in a GeoTIFF.
`html
`
There's also a library called geotiff-geokeys-to-proj4, that allows for reprojecting pixel coordinates and, therefore, consuming geospatial data contained in GeoTIFF.
BigTIFF support
geotiff.js has a limited support for files in the BigTIFF format. The limitations
originate in the capabilities of current JavaScript implementations regarding
64 bit integer parsers and structures: there are no functions to read 64 bit
integers from a stream and no such typed arrays. As BigTIFF relies on 64 bit
offsets and also allows tag values of those types. In order to still provide
a reasonable support, the following is implemented:
* 64 bit integers are read as two 32 bit integers and then combined. As
numbers in JavaScript are typically implemented as 64 bit floats, there
might be inaccuracies for very large values.
* For 64 bit integer arrays, the default Array` type is used. This might