SECURITY MEASUREMENTS IN NIGHT VISION FOR AUTONOMOUS AI SURVEILLANCE
In last few years, surveillance systems are becoming centre of attention to the government, many organisations and different societies. They are using surveillance systems to keep an eye on different activities of living creations in intention of protection and security. But detecting living object
2025-06-28 16:29:02 - Adil Khan
SECURITY MEASUREMENTS IN NIGHT VISION FOR AUTONOMOUS AI SURVEILLANCE
Project Area of Specialization Artificial IntelligenceProject SummaryIn last few years, surveillance systems are becoming centre of attention to the government, many organisations and different societies. They are using surveillance systems to keep an eye on different activities of living creations in intention of protection and security. But detecting living object and tracking their behaviours at night is still a challenging issue to security departments due to low light level. An Intelligent Surveillance system at night-time basically censored the performance, happenings or changing information usually in terms of human beings, vehicles or any other objects from a distance by means of some electronic equipment (usually digital camera) and report the authorities about any uncertain activity. The scopes like prevention, detection, and intervention which have led to the development of real and consistent video surveillance systems are capable of intelligent video and image processing competencies. Our system will be designed in such a way that it will not only detect objects but will identify and track anomalous behaviour of the objects as well. Previously, surveillance systems were dependant on human operators which cause inadequacies. Our goal is to replace such systems with an intelligent surveillance system which will identify abnormal activities by using artificial intelligence technologies. This establishment will bestow a new category of features which will have a great impact on marketplace security.
Project Objectives- Detecting object in night-time is challenging and it is substantially dependant on enhanced quality of images. So, we will first enhance the quality of our input data and then detect objects mainly humans, animals, and dangerous equipment in night vision by applying different background subtraction and object detection techniques like histograms of oriented gradient (HOG).
- Analyse the behaviour of motion of detected object this will be beneficial in classifying objects unmistakably.
- Identifying the object by implementing object recognition algorithms which basically rely on matching, learning or pattern recognition algorithms using featured-based techniques.
- Optimize the perceived images by correcting image dimensions and image format. After that we will compress the image in order to have optimize image delivery.
- Track the behavior of living ones. There can be single or multi-objects in the proposed scene so we must detect all.
- Determine geometric transformation for registering the image to support the visual identification of certain features expressed by characteristics, shapes and patterns.
- Classify the object on feature-based or motion-based. We will be classifying multiple objects using support vector machine (SVM).
- Disclose any irregular act that can threaten security and safety. This can be done by training our system with predefined anomalous acts.
- At first, video/image sequence will be inputted to our system on which pre-processing algorithms will be applied
- From that enhanced video/image sequence our system will detect objects. Objects can be human, non-human or any non-living thing. But our major goal is to detect living or more precisely humans
- Then from that detected objects, our system will track the living ones
- When the objects will be tracked successfully then on the basis of their behaviour of motion system will detect and identify anomaly
- For which system will bring an alarm on and it would alarm about what kind of anomaly has happened as well
The system will be use by any security department such as Military, Government and for smart cities/homes as well. It will be beneficial for any person whose aim is to protect himself or surroundings.
Technical Details of Final DeliverableFinal deliverable will be an intelligent surveillance system which will be able to detect and identify anomalous behaviour of person.
Final Deliverable of the Project Hardware SystemCore Industry SecurityOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Partnerships to achieve the GoalRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 30000 | |||
| IR Lens | Equipment | 1 | 10000 | 10000 |
| Drone | Equipment | 1 | 20000 | 20000 |