Abstract

The detection of underwater Submerged Historical Monuments and Preservation of the vital for understanding historical civilizations. This project employs a hybrid approach combining classical computer vision (CV) techniques and modern deep learning-based detection models to identify submerged historical monuments. By integrating OpenCV for preprocessing and feature detection and YOLO for real-time object detection, the project provides a robust framework for identifying and documenting underwater structures.

Motivation

Submerged historical monuments are often neglected due to the challenges of underwater detection, including poor visibility, marine growth, and high computational requirements. Advanced computer vision techniques, leveraging both classical and deep learning methods, can bridge the gap between technology and marine archaeology, facilitating.

Importance of the project

1. Cultural Heritage: Helps preserve and document underwater archaeological sites.
2. Technological Innovation: Combines classical and modern computer vision methodologies.
3. Practical Applications: Offers scalable solutions for underwater surveys in marine archaeology and environmental research.

Pipeline Overview

Image Input --> Preprocessing (Classical CV) --> Feature Detection --> YOLO Detection (if available) --> Visualization

Code Implementation

pip install opencv-python matplotlib torch ultralytics
from ultralytics import YOLO
import os

def load_yolo_model(model_path='yolo_model.pt'):
if os.path.exists(model_path):
return YOLO(model_path)
else:
print('YOLO model file not found. Proceeding with classical CV pipeline.')
return None
import cv2

def preprocess_image(image):
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
return clahe.apply(gray)

def detect_features_cv(image):
edges = cv2.Canny(image, 50, 150)
contours, _ = cv2.findContours(edges, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
return contours

Visualization

import matplotlib.pyplot as plt

def visualize_detection(image, contours, title='Detected Features'):
overlay = image.copy()
cv2.drawContours(overlay, contours, -1, (0, 255, 0), 2)
plt.figure(figsize=(10, 6))
plt.imshow(cv2.cvtColor(overlay, cv2.COLOR_BGR2RGB))
plt.title(title)
plt.axis('off')
plt.show()

Practical Implementation

def detect_monuments(image_path, yolo_model=None):
image = cv2.imread(image_path)
if yolo_model:
results = yolo_model(image)
for result in results:
boxes = result.boxes.xyxy.cpu().numpy()
for box in boxes:
x1, y1, x2, y2 = map(int, box[:4])
cv2.rectangle(image, (x1, y1), (x2, y2), (255, 0, 0), 2)
plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
plt.title('YOLO Detection')
plt.axis('off')
plt.show()
else:
processed_image = preprocess_image(image)
contours = detect_features_cv(processed_image)
visualize_detection(image, contours)

Dataset Processing

dataset_path = '/content/drive/MyDrive/DataSet'
yolo_model_path = 'yolo_model.pt'

yolo_model = load_yolo_model(yolo_model_path)

for filename in os.listdir(dataset_path):
if filename.lower().endswith(('.jpg', '.png')):
image_path = os.path.join(dataset_path, filename)
print(f'Processing: {image_path}')
detect_monuments(image_path, yolo_model)

Visualization

Scope for Future Development

1. Improved Deep Learning Models:

• Train specialized models for underwater environments.
• Enhance object detection with fine-tuned weights for low-visibility conditions.

Integration with ROVs/Drones

• Deploy models on embedded systems for real-time underwater detection.

Augmented Reality:

• Use AR to visualize detected structures for archaeological analysis.

Conclusion

This project demonstrates a hybrid approach to detect submerged historical monuments using computer vision techniques. By leveraging classical CV and deep learning, we aim to provide scalable, efficient solutions to preserve our underwater heritage.