Add WRI data

data_prep/WRI.py

@@ -0,0 +1,46 @@
+import os
+import pandas as pd
+from deepforest import utilities
+from PIL import Image
+
+# Define the root directory
+root_dir = "/orange/ewhite/DeepForest/WRI/WRI-labels-opensource"
+
+# Initialize an empty list to store dataframes
+dataframes = []
+
+# Iterate through each file in the directory
+for filename in os.listdir(root_dir):
+    if filename.endswith(".txt"):
+        file_path = os.path.join(root_dir, filename)
+
+        # Load the text file into a dataframe
+        df = pd.read_csv(file_path, delim_whitespace=True, header=None, names=["index", "x", "y", "width", "height"])
+
+        # Load the corresponding image to get its shape
+        image_path = os.path.join(root_dir, os.path.splitext(filename)[0] + ".png")
+        image = Image.open(image_path)
+        original_width, original_height = image.size
+
+        # Apply the coordinate transformation to each row
+        df["xmin"] = (df["x"]  * original_width).astype(int)
+        df["ymin"] = (df["y"]  * original_height).astype(int)
+        df["xmax"] = (df["xmin"] + df["width"] * original_width).astype(int)
+        df["ymax"] = (df["ymin"] + df["height"] * original_height).astype(int)
+
+        # Add the image_path column
+        df["image_path"] = image_path
+
+        # Append the dataframe to the list
+        df = utilities.read_file(df, root_dir)
+        dataframes.append(df)
+
+# Concatenate all dataframes
+annotations_df = pd.concat(dataframes, ignore_index=True)
+
+# Add the 'label' and 'source' columns
+annotations_df["label"] = "Tree"
+annotations_df["source"] = "World Resources Institute"
+
+# Save the concatenated dataframe to annotations.csv
+annotations_df.to_csv(os.path.join(root_dir, "annotations.csv"), index=False)

data_prep/collect_tasks.py

@@ -13,6 +13,7 @@
     "/orange/ewhite/DeepForest/Velasquez_urban_trees/tree_canopies/nueva_carpeta/annotations.csv",
     '/orange/ewhite/DeepForest/individual_urban_tree_crown_detection/annotations.csv',
     '/orange/ewhite/DeepForest/Radogoshi_Sweden/annotations.csv',
+    "/orange/ewhite/DeepForest/WRI/WRI-labels-opensource/annotations.csv",
     '/orange/ewhite/DeepForest/ReForestTree/images/train.csv']
 
 TreePoints = [

docs/datasets.md

@@ -37,6 +37,12 @@ ISPRS Journal of Photogrammetry and Remote Sensing, Volume 206, 2023
 
 **Location:** Mato Grosso do Sul, Brazil
 
+### World Resources Institute 
+
+NAIP Imagery from across the United States
+
+![sample_image](public/World_Resources_Institute.png)
+
 ## Points
 
 ### Amirkolaee et al. 2023