Skip to main content

Hello everyone,


Does anyone know what’s wrong with Sentinel 5P configurations? No one layer in my configuration utility dashboard visualizes correctly. Each pixel in the output coverage has the same value and color (usually black). Other data sources work well. What it could be and how to resolve this problem?

Hey,


there may be one of several causes for this.



  • If the layers used evalscripts version 1 or 2, the evalscripts were automatically converted to version 3. More about evalscripts version 3 here and about the conversion here.




  • The possible values for visualization are between 0 and 1 (black and white colors, if grayscale is used). Most Sentinel-5P bands have different value ranges (between 0 and 0.1 / between 1,000 and 110,000) - possible values available here. So, in this case, using raw band values for visualization is not particularly useful. If used, there are big patches of black of really dark grey color and just individual pixels of lighter colors or vice versa (example in EO Browser).


    Default visualizations in EO Browser seem to work ok (example). We use colorBlend function (docs). The function receives the pixel value, array of values and array of colors. The arrays function as a mapping - value at a selected index has the color at a selected index. The length of the arrays determines the number of steps - values with “fixed” color. For values between the values in the array of values, the function creates a blend of the color for the first value above and first value below the input value.


    Example:


    var val = // selected band
    var minVal = // min possible value of the selected band;
    var maxVal = // max possible value of the selected band;
    var limits = sminVal, ..., maxVal]; // needs to be in ascending order
    var colors = s 1, 0, 0], 00.5, 0, 0], 00, 1, 0], ...];

    var returnValue = colorBlend(val, limits, colors);
    // returnValue.push(dataMask); // adding dataMask if needed
    return returnValue;



  • There might simply be some small mistake in the evalscript used. If steps above don’t help, please paste here the evalscript of one of the layers and we will help to find the best solution.



Hope it helps.

Cheers!


Reply