Normalizing Bands in All Images in Image Collection

Question

I am trying to normalize the values in each band across images in the imageCollection I have filtered.

I tried running this here

var normalized_collection = l1c_collection
  .filterDate(start_scene1, end_scene1)
  .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 5))
  .map(function(my_image){
    var normalized_image = ee.ImageCollection.fromImages(
       my_image.bandNames().map(function(name){
         name = ee.String(name);
         var band = my_image.select(name);
         return band.expression(
           '(x * slope) + intercept', {
             'x': band,
             'slope': 0.0342,
             'intercept': 0.6494});
       })).toBands().rename(my_image.bandNames());
       return normalized_image;
});

However, the process returned an Error: User memory limit exceeded.

Is there a workaround to this?

Answer 1

There is no real need to first map across images, then across bands. Doing so can result in a large computational load on one node at Google Earth Engine's server-side.

It is better to use inbuilt functions which are innately parallelized on the server-side. In your case, you can use .multiply() and .add() since the slopes and intercepts are constant across images and bands.

This will look like the following:

var slope = 0.0342;
var intercept = 0.6494;
var normalized_collection = l1c_collection
  .filterDate(start_scene1, end_scene1)
  .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 5))
  .map(function(image){
    return image.multiply(slope).add(intercept).copyProperties(image)
  })

The above code does not run into memory errors.

Link to complete code.