Upload 3 files

README.md

@@ -1,12 +1,39 @@
----
-title: Model1 Object Distance Claculator
-emoji: 🏢
-colorFrom: green
-colorTo: gray
-sdk: gradio
-sdk_version: 5.35.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# 📐 Real-World Object Distance Estimator
+
+This app lets you measure **real-world distances** between any two objects in an image using:
+
+- 🧠 YOLOv8 object detection
+- 📏 Reference object with known size
+- 🧮 Pixel and real-world (cm) distance calculation
+
+---
+
+## 🖼️ How to Use
+
+1. **Upload** a real-world image (e.g., photo taken with your phone)
+2. **Enter the real-world width** of a known object (e.g., "TV = 100 cm")
+3. Click **Detect & Get Object List**
+4. Use the indexes to:
+   - Select the **reference object**
+   - Select the **two target objects** you want to measure
+5. Click **Calculate Distance**
+
+---
+
+## 🛠️ Powered By
+
+- [Ultralytics YOLOv8](https://github.com/ultralytics/ultralytics)
+- [Gradio](https://gradio.app/)
+- OpenCV & NumPy
+
+---
+
+## 💡 Coming Soon
+
+- Monocular depth estimation (MiDaS)
+- Open-vocabulary object selection
+- Automatic reference detection
+
+---
+
+Try it out and let us know your feedback!

app.py

@@ -0,0 +1,101 @@
+# app.py
+
+from ultralytics import YOLO
+import gradio as gr
+import cv2
+import numpy as np
+from math import sqrt
+
+# Load YOLOv8 model
+model = YOLO("yolov8n.pt")
+
+def get_center(bbox):
+    x1, y1, x2, y2 = bbox
+    return ((x1 + x2) / 2, (y1 + y2) / 2)
+
+def calculate_distance(image, ref_index, ref_width_cm, obj1_index, obj2_index):
+    results = model.predict(image, verbose=False)
+    detections = results[0].boxes
+    class_names = model.names
+
+    if not detections:
+        return image, "No objects detected"
+
+    # Parse detected objects
+    objects = []
+    for box in detections:
+        cls_id = int(box.cls[0])
+        class_name = class_names[cls_id]
+        bbox = box.xyxy[0].tolist()  # x1, y1, x2, y2
+        objects.append((class_name, bbox))
+
+    # Error handling
+    if ref_index >= len(objects) or obj1_index >= len(objects) or obj2_index >= len(objects):
+        return image, "Invalid object index selection"
+
+    # Reference object
+    _, ref_bbox = objects[ref_index]
+    ref_pixel_width = abs(ref_bbox[2] - ref_bbox[0])
+    pixels_per_cm = ref_pixel_width / ref_width_cm
+
+    # Target objects
+    name1, bbox1 = objects[obj1_index]
+    name2, bbox2 = objects[obj2_index]
+    center1 = get_center(bbox1)
+    center2 = get_center(bbox2)
+
+    # Pixel and real-world distances
+    pixel_distance = sqrt((center1[0] - center2[0])**2 + (center1[1] - center2[1])**2)
+    real_distance = pixel_distance / pixels_per_cm
+
+    # Annotate image
+    annotated = results[0].plot()
+
+    # Draw line between objects
+    cv2.line(annotated, (int(center1[0]), int(center1[1])), (int(center2[0]), int(center2[1])), (0, 255, 0), 2)
+    cv2.putText(annotated, f"{real_distance:.2f} cm", (int((center1[0] + center2[0]) / 2), int((center1[1] + center2[1]) / 2) - 10),
+                cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
+
+    # Convert to RGB for display
+    annotated_rgb = cv2.cvtColor(annotated, cv2.COLOR_BGR2RGB)
+
+    info = f"Pixel Distance: {pixel_distance:.2f}px\nReal-World Distance: {real_distance:.2f} cm"
+    return annotated_rgb, info
+
+# Gradio Interface
+def get_object_list(image):
+    results = model.predict(image, verbose=False)
+    detections = results[0].boxes
+    class_names = model.names
+
+    if not detections:
+        return []
+
+    return [model.names[int(box.cls[0])] for box in detections]
+
+with gr.Blocks() as demo:
+    gr.Markdown("## 📐 Object Distance Estimator (YOLOv8 + Reference Object)")
+
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(type="numpy", label="Upload Image")
+            ref_width = gr.Number(label="Reference Object Width (cm)", value=30)
+            detect_btn = gr.Button("Detect & Get Object List")
+
+            object_list = gr.Textbox(label="Detected Objects (index: name)", interactive=False)
+
+            ref_index = gr.Number(label="Reference Object Index", value=0)
+            obj1_index = gr.Number(label="Object 1 Index", value=1)
+            obj2_index = gr.Number(label="Object 2 Index", value=2)
+            calculate_btn = gr.Button("Calculate Distance")
+
+        with gr.Column():
+            output_image = gr.Image(label="Annotated Image")
+            output_text = gr.Textbox(label="Distance Info")
+
+    detect_btn.click(fn=get_object_list, inputs=image_input, outputs=object_list)
+    calculate_btn.click(fn=calculate_distance,
+                        inputs=[image_input, ref_index, ref_width, obj1_index, obj2_index],
+                        outputs=[output_image, output_text])
+
+demo.launch()

requirements.txt

@@ -0,0 +1,4 @@
+ultralytics>=8.0.20
+gradio>=4.16
+opencv-python
+numpy