Earth Engine: Error: Image.select: Pattern 'test' did not match any bands

Question

I am trying to combine two images one containing various bands of the sentinel 2 mission and the other one is created from a feature collection containing two types of polygons that have the values "test" of 0 and 1 as can be seen here:

The combined image should then be exported following the tutorial here (https://colab.research.google.com/github/google/earthengine-api/blob/master/python/examples/ipynb/UNET_regression_demo.ipynb#scrollTo=FyRpvwENxE-A)

Here is my code:

RESPONSE = 'test'

trainingMap = ee.FeatureCollection('users/xxx/xxx')
trainingMapReduced = trainingMap.reduceToImage([RESPONSE], 'max');

# combine the images
featureStack = ee.Image.cat([
  image.select(BANDS),
  trainingMapReduced.select([RESPONSE])
]).float()

# consider a kernel of KERNEL_SIZE x KERNEL_SIZE around each pixel
list = ee.List.repeat(1, KERNEL_SIZE)
lists = ee.List.repeat(list, KERNEL_SIZE)
kernel = ee.Kernel.fixed(KERNEL_SIZE, KERNEL_SIZE, lists)

arrays = featureStack.neighborhoodToArray(kernel)

# load training and evaluation areas

trainingPolys = ee.FeatureCollection('users/xxx/trainingRegion')
evalPolys = ee.FeatureCollection('users/xxx/testingRegion')

trainingPolysList = trainingPolys.toList(trainingPolys.size())
evalPolysList = trainingPolys.toList(trainingPolys.size())

# These numbers determined experimentally.
n = 100 # Number of shards in each polygon.
N = 2000 # Total sample size in each polygon.

# Export all the training data (in many pieces), with one task 
# per geometry.
for g in range(trainingPolys.size().getInfo()):
  geomSample = ee.FeatureCollection([])
  for i in range(n):
    sample = arrays.sample(
      region = ee.Feature(trainingPolysList.get(g)).geometry(), 
      scale = 30, 
      numPixels = N / n, # Size of the shard.
      seed = i,
      tileScale = 8
    )
    geomSample = geomSample.merge(sample)

  desc = TRAINING_BASE + '_g' + str(g)
  task = ee.batch.Export.table.toDrive(
    collection = geomSample,
    description = desc, 
    folder = BUCKET, 
    fileNamePrefix = FOLDER + '_' + desc,
    fileFormat = 'TFRecord',
    selectors = BANDS + [RESPONSE]
  )
  task.start()

The task starts but it fails with the following error message "Error: Image.select: Pattern 'test' did not match any bands."

When I use

featureStack = image.addBands(trainingMapReduced, [RESPONSE])

instead of

featureStack = ee.Image.cat([
  image.select(BANDS),
  trainingMapReduced.select([RESPONSE])
]).float()

I get the error message "Error: Image.addBands: Cannot add band 'test' because it doesn't appear in srcImg. "

The print(trainingMapReduced) looks like this:

ee.Image({ "type": "Invocation", "arguments": { "input": { "type": "Invocation", "arguments": { "collection": { "type": "Invocation", "arguments": { "tableId": "users/xxx/xxx" }, "functionName": "Collection.loadTable" }, "properties": [ "test" ], "reducer": { "type": "Invocation", "arguments": {}, "functionName": "Reducer.max" } }, "functionName": "Collection.reduceToImage" }, "bandSelectors": [ "test" ] }, "functionName": "Image.select" })

Does anybody have an idea how to solve this error?

Answer 1

The problem was that the max reducer renames the bandname to "max" It had to be renamed again.

trainingMap = ee.FeatureCollection('users/xxx/xxx')
trainingMapReduced = trainingMap.reduceToImage([RESPONSE], 'max')
trainingMapReduced = trainingMapReduced.rename(RESPONSE)

Answer 2

I don't think ee.FeatureCollection.reduceToImage() names the bands based on the property/field they were reduced from. My guess is that trainingMapReduced only has one band named 0. Maybe try:

  image.select(BANDS),
  trainingMapReduced.select([0])
]).float()

You could also just do print(trainingMapReduced) to see what the names of the bands are.