Discussion: Histogram Matching Demo, Client-side JS implementation

Hi there,

First wanted to say thanks for this great rio-hist utility. I've been tempted to compute Histogram Matching client-side for a specific project for quite some time and finally took the leap.

This ended up with the below demo app histogram-matching-js.
Runs in 20ms on FHD images. If source is geotiff, geotiffjs is used to copy all its geoMetadata.

Lets you interactively tune:

• source and target input raster images via File inputs
• maxMpx to downsample input source/target Uint8 arrays for faster histogram computation
• color_space and bands after having a look at https://github.com/mapbox/rio-hist/issues/3 and how your CLI offers optional tweaking of these params
• binCount: marginal compute time enhancement since interp on 256-length arrays is already fast with regards to extracting the RGB constituents from the input raster arrays

Start by reviewing the linked histogram-matching-js demo, its use of geotiffjs, and the parameter discussion in issue #3. The issue does not identify a repository file or a requested change; clarify whether maintainers want an integrated client-side implementation, documentation, or only a reference to the demo before defining what done means.