In Google Earth Engine, How do I select pixels from one image collection which correspond to a selected pixel value from another image collection? - mask

Question

I want to plot the count of burn pixels for modis burned area product within my geometry regions called "table" for only agricultural pixels (obtained from 'lc' image collection). I couldn't find anything in the docs to indicate you can do such a query between 2 image collections. Anyone have any suggestions?
I have tried using a mask, but it seems that this might only work on individual ee.Image not between different image collections. The code is shown below:
var modba = ee.ImageCollection('MODIS/006/MCD64A1').filterDate('2017-01-
01', '2017-12-31').select('BurnDate')
var modbaN = ee.ImageCollection('MODIS/006/MCD64A1').filterDate('2017-01-
01', '2017-12-31').select('Uncertainty')
var lc = ee.ImageCollection('MODIS/006/MCD12Q1').filterDate('2017-01-01',
'2017-12-31').select('LC_Type1')
var AgOnly = lc.map(function(img) {
var ag = img.select('LC_Type1');
return ag.eq(12);
//Would also like to maybe have 2 or 3 LC types to select here
});
var mask_ba = modba.map(function(img){
return img.updateMask(AgOnly);
});
var bats =
//ui.Chart.image.seriesByRegion(modba, table, ee.Reducer.count());
ui.Chart.image.seriesByRegion(mask_ba, table, ee.Reducer.count());
print(bats);
var unts =
ui.Chart.image.seriesByRegion(modbaN, table, ee.Reducer.mean());
print(unts);

Answer 1

It's still doable with a wider date range and several land cover types.
In that case, just keep your old code that calculates AgOnly, and modify the code that calculates mask_ba as below:
var mask_ba = modba.map(function(img){
var img_year = img.date().format('YYYY');
var start_date = ee.Date(img_year.cat('-01-01'));
var end_date = start_day.advance(1, 'year');
var Agri_this_year = AgOnly.filterDate(start_date, end_date).max();
return img.updateMask(Agri_this_year);
});
Basically, the above code just extracts the year of the current img, then use filterDate method to select the land type cover of that year from AgOnly image collection, and finally apply updateMask.
The same idea could be applied to other land cover types.
Hope this helps.

Answer 2

As I understand, what you're trying to do is to mask each image in modba image collection (which has 12 images or one per month) by the corresponding image in AgOnly image collection (which has only 1 image for the whole year). That's totally doable.
In your provided code, you're updateMask using AgOnly (an image collection) which is not allowed by GEE.
All you need to do is just make AgOnly an image before using it for updateMask.
Try this:
var AgOnly = lc.map(function(img) {
var ag = img.select('LC_Type1');
return ag.eq(12);
//Would also like to maybe have 2 or 3 LC types to select here
}).max();
The max() method will convert your image collection into an image. You can also use min() or mean() if you like, which will all give the same result as there's only one image in AgOnl anyway.