OptiVision is a real-time object detection and analysis system capable of processing images, videos, and RTSP streams. It uses TensorFlow, OpenCV, and a sleek Flask web interface to provide real-time detection results, visualizations, and analytics for any object detection task.
Welcome to OptiVision! This is the central page of the application where you can navigate to different sections for processing images, videos, and streams. The homepage provides quick access to each detection feature.
Stream live feeds from RTSP cameras and detect objects in real-time with precision.
- Seamless connection to RTSP camera streams.
- Real-time object detection and tracking.
- Display detected objects live as the stream plays.
- Enter the RTSP stream URL or channel ID in the provided input field.
- Click Start Stream to begin receiving and analyzing the feed.
- Detected objects will be displayed live with bounding boxes.
Upload static images for detailed object detection and analysis.
- Upload any image to detect objects.
- View detected objects with bounding boxes.
- Save and download the processed image with detected objects.
- Click Upload Image and select an image from your computer.
- Set the detection threshold to control sensitivity.
- Click Process Image to view the results.
- Download the processed image by clicking the Download Image button.
Analyze uploaded videos to identify and track objects frame by frame.
- Upload video files for object detection.
- Track objects across frames with bounding boxes.
- Video playback with detected objects.
- Upload a video file by clicking Upload Video.
- Select the detection options (e.g., processing every nth frame).
- Click Start Processing to view the video with detected objects.
- Watch the video in real-time or download the processed video.
Connect and analyze streams from various sources for seamless detection.
- Support for RTSP and HTTP-based stream sources.
- Real-time frame-by-frame analysis.
- View detected objects on a live video feed.
- Enter the stream URL (RTSP or HTTP).
- Click Start Stream to connect to the stream.
- View real-time object detection results.
- Optionally, download detection results.
Visualize detection performance and trends with interactive charts.
- View detection statistics like object counts and accuracy over time.
- Generate trend graphs for object detection performance.
- Interactive and dynamic charts for better analysis.
- Select a date range to visualize the detection statistics.
- View the generated interactive graphs showing trends and performance.
- Analyze the data for further insights into detection accuracy.
Get instant help and insights from our intelligent assistant.
- Chat with an AI-powered assistant for help.
- Get real-time insights and troubleshooting for detection tasks.
- Open the chatbot by clicking the Chat with Assistant button.
- Type your question or request for help.
- Get responses to help guide you through any detection task or issue.
- Ganesh Patidar (20214061)
- Hardik Kumar Singh (20214249)
- Divyanshu (20214317)
- Harsh Dave (20214534)
- Problem Statement
- Introduction
- Motivation
- Applications
- Proposed Work
- Experimental Setup
- Result Analysis
- Challenges
- Future Work
- References
- Livestream Camera Integration with Jetson Nano Hardware
- Object Detection on Images, Videos, and Livestream Feeds
This project implements a real-time object detection system using Jetson Nano, leveraging deep learning algorithms for accurate and efficient object classification. It enhances surveillance, security, and operational efficiency in various applications.
The inspiration for this project stems from the critical need to improve security measures in public transport systems. By leveraging real-time CCTV feeds, we aim to provide an automated surveillance system that ensures passenger safety, particularly for vulnerable groups. Our goal is to enable authorities to detect potential security threats proactively.
- Surveillance and Security Systems
- Traffic Management
- Retail Analytics
- Industrial Automation
- Smart Cities
- Environmental Monitoring
- Jetson Nano Setup
- Live Stream Implementation
- Data Collection & Model Training
- Evaluation of Object Detection Models
- Performance Analysis of Different Models
- Flashed the NVIDIA OS using Balena Etcher.
- Installed JetPack SDK 4.4.0 for development.
- Booted Jetson Nano and configured the environment.
- Utilized OpenCV with CUDA for optimized real-time video processing.
- Enabled efficient video capture and frame-by-frame object detection.
- Collected data using
simple_image_download. - Labeled images using
labelImg. - Trained a YOLOv7 model using Google Colab for improved computational performance.
- Compared TensorFlow Model Zoo models:
- SSD ResNet50 640x640
- CenterNet ResNet101 FPNv1 512x512
- Evaluated based on mean Average Precision (mAP) and inference time.
- Precision = TP / (TP + FP)
- Recall = TP / (TP + FN)
- mAP = Average of AP across all classes
| Model | mAP (Accuracy) |
|---|---|
| CenterNet ResNet-101 | Low |
| SSD ResNet-50 | Moderate |
| YOLOv7 (Custom) | High |
- Fastest: CenterNet ResNet-101 (Low accuracy, high speed)
- Balanced: SSD ResNet-50 (Moderate speed & accuracy)
- Most Accurate: YOLOv7 (High accuracy, slower inference)
- Proxy Configuration Issues
- Package Installation Errors
- SSL Wrong Version Number
- Python Version Conflicts
- Extended Training Time
- Jetson Nano Compatibility Issues
- Unexpected Shutdowns During Execution
- Performance Optimization
- Cloud Integration
- Real-time Alerts & Notifications
- Enhanced User Interface
- IoT Device Integration
- Abadi, M. et al. TensorFlow Model Zoo
- Liu, W., Anguelov, D., et al. SSD: Single Shot Multibox Detector, ECCV (2016)
- Redmon, J., et al. YOLO: Unified, Real-Time Object Detection, IEEE TPAMI (2016)
- Wang, J., et al. YOLOv7: Trainable Bag of Freebies, IEEE TPAMI (2021)
- PyTorch for Jetson
🚀 Thank you! We appreciate your time in reviewing our project! 🎯