Create inference.py

inference.py

@@ -0,0 +1,103 @@
+from deepforest import main
+from deepforest import get_data
+import matplotlib.pyplot as plt
+from predict import extract_features, predict_similarity, compare_features, extract_features_cp
+import os
+import streamlit as st
+model = main.deepforest()
+model.use_release()
+
+# quit()
+# print(img.head())
+import pandas as pd
+from PIL import Image
+
+
+def split_image_from_dataframe(dataframe, panoramic_image, output_folder_name):
+    """
+    Splits an image into multiple images based on coordinates provided in a dataframe.
+
+    Parameters:
+    dataframe (pd.DataFrame): DataFrame containing image path and coordinates.
+    """
+    cropped_images_info = []
+    cropped_image_paths = []
+    for i, row in dataframe.iterrows():
+        image_path = row['image_path']
+        left, top, right, bottom = row['xmin'], row['ymin'], row['xmax'], row['ymax']
+
+        image = Image.open(panoramic_image)
+        
+        cropped_image = image.crop((left, top, right, bottom))
+        cropped_image_dict = {
+            f'image': cropped_image,
+            f'position': (left, top, right, bottom)
+        }
+        cropped_images_info.append(cropped_image_dict)
+        cropped_image_paths.append(f'{output_folder_name}/cropped_image_{i}.png')
+        cropped_image.save(f'{output_folder_name}/cropped_image_{i}.png')  # Save each cropped image
+    
+    dataframe['cropped_image_path'] = cropped_image_paths
+    return cropped_images_info
+
+
+# print(images_list)
+# quit()
+# Load images from folder
+def extract_treespecies_features(folder_path):
+    image_files = [os.path.join(folder_path, f) for f in os.listdir(folder_path) if f.endswith(('png', 'jpg', 'jpeg', '.JPG'))]
+
+    species_feature_list = [{"feature": extract_features(file), "file_name": file} for file in image_files]
+    return species_feature_list
+
+
+# print(species_feature_list[:2])
+def perform_inference(images_list, species_feature_list):
+    for idx, item in enumerate(images_list):
+        image = item["image"]
+        feature_cp = extract_features_cp(image)
+        for idx, species in enumerate(species_feature_list):
+            euclidean_dist, cos_sim = compare_features(feature_cp, species["feature"])
+        # print(f'Euclidean Distance: {euclidean_dist}')
+        # print(f'Cosine Similarity: {cos_sim}')
+
+        # Predict similarity
+            is_similar = predict_similarity(feature_cp, species["feature"], threshold=0.8)
+        # print(species)
+        # print(f'Are the images similar? {"Yes" if is_similar else "No"}')
+
+            result = "Yes" if is_similar else "No"
+            if result == "Yes":
+                item[f"result_{idx}"] = result
+                item[f"file_name_{idx}"] = species["file_name"]
+
+    return images_list
+
+
+
+
+if __name__ == '__main__':
+    pan_image = "D:/Downloads/image/plant_images/plant_images/drone_igapo_flooded_forest/DJI_20240504124024_0037_D.JPG"
+
+    sample_image_path = get_data(pan_image)
+    # img = model.predict_image(path=sample_image_path, return_plot=False)
+    # from PIL import Image
+    # print(img)
+    img_df = ""
+    # img_actual = model.predict_image(path=sample_image_path, return_plot=True, color=(0, 165, 255), thickness=9)
+    img_actual = model.predict_tile(raster_path=sample_image_path, return_plot=True, patch_size=100,patch_overlap=0.25)
+    # im = Image.open('Foto.jpg')
+    # im.save('Foto.png')
+    #predict_image returns plot in BlueGreenRed (opencv style), but matplotlib likes RedGreenBlue, switch the channel order. Many functions in deepforest will automatically perform this flip for you and give a warning.
+    plt.imshow(img_actual[:,:,::-1])
+    # plt.show(img[:,:,::-1])
+    plt.savefig("cropped_test3/panoramic_2.png")
+    quit()
+    images_list = split_image_from_dataframe(img_df, pan_image)
+    folder_path = 'D:/Downloads/image/plant_images/plant_images/drone_igapo_flooded_forest/identified_species'
+
+    species_feature_list = extract_treespecies_features()
+    final_result = perform_inference(images_list, species_feature_list)
+    print(final_result)
+
+