Eval script color assignation based on statistics criteria

EO Products / Scripts
1

Hi I am new in this ecosystem, i am making a crop monitoring app.

working with NDVI need to assign colors based on terciles, quartiles, or in Z-score.

Need to create a https://docs.sentinel-hub.com/api/latest/evalscript/functions/#createdefaultcolormap

But the first column values come from the breaks of terciles, quartiles, or in Z-score.

My python function for quartiles:

def ColorLutFull(img):
    breaks = np.quantile(img, [0.25,0.5,0.75])      # Get breaks of quantiles    
    imgColorsIndex = np.searchsorted(breaks,img)    # imagen of colors index
    pallete= np.array([[255, 0, 0], [255, 215, 0], [0, 255, 0], [0, 100, 0],[0, 100, 100]],np.uint8)        
    return  pallete[imgColorsIndex] # Generate image with fancy indexing

just only 4 lines, How can do that with eval scripts ?.
In the samples script see that only it is possible to have access one pixel at time.!
The next picture show the NDVI divided in quartiles
paintedQ4Green|300x300

2

Hi @LionelMarquez,

your assumption that Evalscript is processsing one pixel at a time, is correct.
To do what you want to do, we propose a two step process:

  1. You first use our Statistical API or the updateOutputMetadata part (see Tip 6 here for an example) of the Process API to get statistics of your parcel.
  2. Then use these values to dynamically construct the EVALSCRIPT and pass it on to our system.

Both steps can be done in real-time.