TypeError: Image data of dtype object cannot be converted to float

[Muhammad-Waqas5012]

The main functionality of the script is to extract GeoJSON data, load it into Earth Engine, retrieve Sentinel-2 imagery, calculate the mean NDVI, and visualize the NDVI for each GeoJSON feature.

My GeoJSON file.

{"type":"FeatureCollection","name":"66","crs":{"type":"name","properties":{"name":"urn:ogc:def:crs:OGC:1.3:CRS84"}},"features":[[{"type":"Feature","properties":{"farmer_id":"66","field_id":"2527","field_crop_id":"782"},"geometry":{"type":"MultiPolygon","coordinates":[["73.807923","31.874025"],["73.807375","31.874134"],["73.807794","31.871734"],["73.811309","31.871643"],["73.811502","31.87213"],["73.811148","31.872208"],["73.811529","31.873"],["73.810622","31.873333"],["73.810574","31.873319"],["73.808062","31.874189"],["73.807987","31.874025"],["73.807923","31.874025"]]}},{"type":"Feature","properties":{"farmer_id":"66","field_id":"408","field_crop_id":"779"},"geometry":{"type":"MultiPolygon","coordinates":[["73.806423917413","31.876305483265"],["73.805073425174","31.875375605313"],["73.805913291872","31.871866128353"],["73.811334036291","31.871658562207"],["73.811508715153","31.872104159735"],["73.811180479825","31.872252786674"],["73.811554983258","31.873010438666"],["73.810586370528","31.873344418569"],["73.809291869402","31.87384382019"],["73.808021843433","31.874272893423"],["73.807903490961","31.874086971424"],["73.80733653903","31.87418007483"],["73.807264119387","31.874479599523"],["73.806831277907","31.874537112708"],["73.806423917413","31.876305483265"]]}}]]}

My Script.

#import all required libraries
import geopandas as gpd
import pandas as pd
import ee
import geemap
import os
import glob
import matplotlib.pyplot as plt
import numpy as np
import datetime
from shapely.geometry import Polygon, MultiPolygon
import json

# Authenticate Earth Engine (Note: if already not) or Authentication is expired, Remove "#" to run the next line.
# ee.Authenticate()

# Initialize Earth Engine
ee.Initialize()

# Specify the folder path containing GeoJSON files
folder_path = 'D:/R1'

# Get a list of all GeoJSON files in the folder
geojson_files = glob.glob(folder_path + '/*.geojson')

# Iterate over each GeoJSON file
for geojson_file in geojson_files:
    with open(geojson_file, 'r') as f:
        data = json.load(f)

    # Access the features in the GeoJSON file
    features = [feature for feature_list in data['features'] for feature in feature_list]

    # Iterate over each feature
    for feature in features:
        # Extract properties and geometry
        properties = feature['properties']
        coordinates = feature['geometry']['coordinates']  # Extracting the nested list of coordinates
        geometry = MultiPolygon([Polygon([(float(coord[0]), float(coord[1])) for coord in coordinates])])
        
        # Extract feature ID
        feature_id = properties.get('field_id')  # Assuming the ID property is named 'Id'
        farmer_id = properties['farmer_id']
        field_crop_id = properties['field_crop_id']
        if feature_id is None:
            raise ValueError("Field ID is missing in the GeoJSON file.")

        # Load the GeoJSON string into Earth Engine
        feature_gdf = gpd.GeoDataFrame.from_features([{'geometry': geometry, 'properties': properties}])
        feature_gdf = feature_gdf.set_crs('EPSG:4326')  
        # Load the GeoJSON string into Earth Engine
        feature_geojson = geemap.gdf_to_ee(feature_gdf, geodesic=False)
        
        # Extact the Date
        start_date = ee.Date(datetime.datetime.now()).advance(-5, 'day')
        end_date = ee.Date(datetime.datetime.now())
        
        # Load Sentinel-2 image collection
        s2_collection = ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED').filterDate(start_date, end_date).filterBounds(feature_geojson)
        
        # Check if the collection is empty
        image_count = s2_collection.size().getInfo()
        if image_count == 0:
            # No images available for this feature
            print(f"No images available for Feature ID {feature_id}. Skipping...")
            continue

        # Calculate mean NDVI for each image in the collection
        def calculate_mean_ndvi(image):
            # Calculate NDVI
            ndvi_image = image.normalizedDifference(['B8', 'B4']).rename('NDVI')
            # Calculate mean NDVI
            mean_ndvi = ndvi_image.reduceRegion(reducer=ee.Reducer.mean(), geometry=feature_geojson, scale=10)
            return ndvi_image.set('system:time_start', image.get('system:time_start'))

        # Batch processing of images
        s2_list = s2_collection.toList(image_count)

        # List to store mean NDVI images
        mean_ndvi_images = []
        
    # Iterate over each image in the collection
for i in range(image_count):
    # Get the ith image
    image = ee.Image(s2_list.get(i))
    
    # Call the function to calculate mean NDVI
    mean_ndvi_image = calculate_mean_ndvi(image)
    
    # Append mean NDVI image to the list
    mean_ndvi_images.append(mean_ndvi_image)
    # Check the data type of the first image in the list
    print(type(mean_ndvi_images[0]))


    # Convert NDVI image to a NumPy array for plotting
    ndvi_array = mean_ndvi_image.select('NDVI').clip(feature_geojson.geometry()).getInfo()

    # Plot NDVI
    plt.figure(figsize=(5.932, 5.932))  # Adjust plot size
    plt.imshow(ndvi_array, cmap='RdYlGn')
    plt.colorbar(aspect=40)
    plt.title(f"NDVI Image for {date_str}")
    plt.show()

the error this script is producing.

TypeError                                 Traceback (most recent call last)
<ipython-input-1-eed1f97587dc> in <module>
    100     # Plot NDVI
    101     plt.figure(figsize=(5.932, 5.932))  # Adjust plot size
--> 102     plt.imshow(ndvi_array, cmap='RdYlGn')
    103     plt.colorbar(aspect=40)
    104     plt.title(f"NDVI Image for {date_str}")

C:\Program Files\ArcGIS\Pro\bin\Python\envs\arcgispro-py3\lib\site-packages\matplotlib\pyplot.py in imshow(X, cmap, norm, aspect, interpolation, alpha, vmin, vmax, origin, extent, filternorm, filterrad, resample, url, data, **kwargs)
   2901         filternorm=True, filterrad=4.0, resample=None, url=None,
   2902         data=None, **kwargs):
-> 2903     __ret = gca().imshow(
   2904         X, cmap=cmap, norm=norm, aspect=aspect,
   2905         interpolation=interpolation, alpha=alpha, vmin=vmin,

C:\Program Files\ArcGIS\Pro\bin\Python\envs\arcgispro-py3\lib\site-packages\matplotlib\__init__.py in inner(ax, data, *args, **kwargs)
   1359     def inner(ax, *args, data=None, **kwargs):
   1360         if data is None:
-> 1361             return func(ax, *map(sanitize_sequence, args), **kwargs)
   1362 
   1363         bound = new_sig.bind(ax, *args, **kwargs)

C:\Program Files\ArcGIS\Pro\bin\Python\envs\arcgispro-py3\lib\site-packages\matplotlib\axes\_axes.py in imshow(self, X, cmap, norm, aspect, interpolation, alpha, vmin, vmax, origin, extent, filternorm, filterrad, resample, url, **kwargs)
   5607                               resample=resample, **kwargs)
   5608 
-> 5609         im.set_data(X)
   5610         im.set_alpha(alpha)
   5611         if im.get_clip_path() is None:

C:\Program Files\ArcGIS\Pro\bin\Python\envs\arcgispro-py3\lib\site-packages\matplotlib\image.py in set_data(self, A)
    698         if (self._A.dtype != np.uint8 and
    699                 not np.can_cast(self._A.dtype, float, "same_kind")):
--> 700             raise TypeError("Image data of dtype {} cannot be converted to "
    701                             "float".format(self._A.dtype))
    702 

TypeError: Image data of dtype object cannot be converted to float

What I have tried.

I added a print statement to print out the ndvi_array just before I attempted to plot it.

Print out the ndvi_array

{'type': 'Image', 'bands': [{'id': 'NDVI', 'data_type': {'type': 'PixelType', 'precision': 'float', 'min': -1, 'max': 1}, 'dimensions': [62, 53], 'origin': [8697, 7265], 'crs': 'EPSG:32643', 'crs_transform': [10, 0, 300000, 0, -10, 3600000]}], 'properties': {'system:time_start': 1713074411607, 'system:footprint': {'type': 'Polygon', 'coordinates': [[[73.806423917413, 31.876305483265007], [73.805073425174, 31.875375605313], [73.805913291872, 31.871866128353002], [73.811334036291, 31.871658562207003], [73.811508715153, 31.872104159734995], [73.811180479825, 31.872252786674], [73.811554983258, 31.873010438666004], [73.810586370528, 31.873344418569005], [73.809291869402, 31.873843820189997], [73.808021843433, 31.874272893423], [73.807903490961, 31.874086971424], [73.80733653903, 31.87418007483], [73.807264119387, 31.874479599523003], [73.806831277907, 31.874537112708], [73.806423917413, 31.876305483265007]]]}}}

The printed output shows the structure of the ndvi_array, which is a dictionary representing an Earth Engine Image object. This indicates that the conversion from Earth Engine Image to NumPy array is not working as expected.

[giswqs]

If you want others to help you, try to provide a minimal script that can reproduce the issue. Don't use absolute file paths. No one can access your local files. Avoid providing a long script with all kinds of for loops, making it very difficult to debug.