Dr. Sumaira Kausar

PhD Theme/Topic: AI-Powered Visual Monitoring for Naval Threat Detection and Coastal Defense

Supervisor: Dr. Sumaira Kausar
Contact #: 03335721300
Email: sumairakausar.buic@bahria.edu.pk
Campus/School/Dept: BUIC/BSEAS/CS & CoE-AI
RAC Approved Supervisor for Research Areas: Machine Learning, computer vision, image analysis

Supervisory Record:
PhD Produced: 0
PhD Enrolled: 2
MS/MPhil Produced: 26
MS/MPhil Enrolled: 2

Topic Brief Description:

Maritime security is increasingly challenged by asymmetric threats, illegal activities, and the strategic significance of coastal and offshore assets. For modern naval forces, effective surveillance is critical to maintaining situational awareness, enabling early threat detection, and supporting rapid operational decisions. This research proposes a machine-vision–driven marine surveillance framework that utilizes optical and infrared imagery from UAVs, USVs, and coastal monitoring systems to automatically detect, classify, and track vessels and surface-level anomalies. The system employs deep learning models tailored to maritime environments to address real-world challenges such as dynamic sea states, low-resolution imagery, low-light conditions, and visually cluttered backgrounds. By focusing on robust visual analytics, the proposed approach aims to enhance real-time detection of suspicious activities, dark vessels, and small or fast-moving intrusions. The expected outcomes will strengthen maritime domain awareness, support naval coastal defense missions, and improve the effectiveness of both autonomous and human-in-the-loop decision-making in complex maritime environments.

Research Objectives/Deliverables:

1. To develop a machine-vision–based framework for real-time detection, classification, and tracking of vessels and surface anomalies in maritime environments.
2. To design and optimize deep learning models specifically tailored to the challenges of marine imagery, including low-light conditions, sea clutter, dynamic waves, and low-resolution inputs.
3. To propose a scalable and operationally deployable surveillance solution that can support naval decision-making and enhance maritime domain awareness.

Research Questions: 

1. How effectively can machine vision and deep learning models detect and classify vessels and maritime anomalies using optical and infrared imagery?
2. What preprocessing and enhancement techniques are needed to overcome the challenges of underwater haze, reflections, glare, and low-light maritime environments?
3. Which AI architectures (e.g., CNNs, transformers) perform best for real-time marine surveillance tasks under varying sea-state conditions?
4. What level of accuracy and response time can be achieved to support real-time naval situational awareness and operational decision-making?

Candidate’s Eligibility Profile:

1. MS/MPhil in Computer Science, AI, Data Science, Electrical/Software Engineering, or a related field with a strong background in machine learning or computer vision.
2. Proficiency in Python and deep learning frameworks with prior experience in image analysis or AI-based research.
3. Strong motivation to conduct applied research in maritime or defense-related domains.