AI Enabled EO Target Detection and Fire Control Solution - Internship
HQ ARMY TRAINING COMMAND Verified employer
Actively hiring · Posted 1d ago · JEV2611C20336572C
Job description
About the internship:Project Requirements:
1. Dataset Preparation: A comprehensive collection of daytime and thermal images and videos of drones, with annotations identifying their locations. The dataset should also include birds, aircraft and background clutter to improve classification accuracy and minimise false detections.
2. Target Detection and Classification: An AI-based model capable of detecting drones in live camera feeds and distinguishing them from other objects in real time. The model should be optimised to operate on compact computing devices such as NVIDIA Jetson for field deployment.
3. Multi-Target Tracking: A tracking system capable of monitoring multiple drones simultaneously, maintaining their individual identities even when their flight paths intersect, and prioritising them based on proximity, speed, and movement patterns.
4. Locating the Target in the Sky: A system that converts the drone's position in the camera frame into real-world directional coordinates, accounting for camera movement. The target's distance should also be estimated using a laser rangefinder or suitable image-based techniques.
5. Movement Prediction: A system that analyses the drone's speed, direction, and trajectory to estimate its future position. It should provide movement forecasts to help operators assess potential risks and monitor changing flight paths.
6. Operator Display: An intuitive interface displaying live camera feeds, highlighted drone detections, tracked-target information and directional indicators. Operators should be able to select targets, review alerts and retain manual control at all times.
7. Testing and Validation: The system should be evaluated using recorded videos, computer simulations and real-world camera footage. Its performance should be measured based on detection range, classification accuracy, response time, tracking reliability, and positional estimation accuracy.Who can apply:Only those candidates can apply who:are available for full time (in-office) internshipare available for duration of 3 monthshave relevant skills and interestsare Computer Science Engineering studentsStipend:INR₹ 75,000 /monthDeadline:[hidden]:59:59Other perks:CertificateSkills required:Python, Computer Vision, C++ Programming, Deep Learning, PyTorch and Object Oriented Programming (OOP)Other Requirements:1. Must be pursuing or have completed a B.Tech./B.E. or an equivalent degree.2. Specialisation in Applied Electronics and Instrumentation Engineering, Artificial Intelligence and Machine Learning, and Robotics and Mechatronics3. Are available for a duration of 3 Months4. Have relevant skills and interestsAbout Company:HQ Army Training Command (ARTRAC) is a key formation of the Indian Army responsible for overseeing and improving military training across the organisation. The command works towards strengthening the professional skills and operational preparedness of Army personnel. It also promotes research, innovation and the adoption of emerging technologies in defence
Responsibilities
Project Requirements:
1. Dataset Preparation: A comprehensive collection of daytime and thermal images and videos of drones, with annotations identifying their locations. The dataset should also include birds, aircraft and background clutter to improve classification accuracy and minimise false detections.
2. Target Detection and Classification: An AI-based model capable of detecting drones in live camera feeds and distinguishing them from other objects in real time. The model should be optimised to operate on compact computing devices such as NVIDIA Jetson for field deployment.
3. Multi-Target Tracking: A tracking system capable of monitoring multiple drones simultaneously, maintaining their individual identities even when their flight paths intersect, and prioritising them based on proximity, speed, and movement patterns.
4. Locating the Target in the Sky: A system that converts the drone's position in the camera frame into real-world directional coordinates, accounting for camera movement. The target's distance should also be estimated using a laser rangefinder or suitable image-based techniques.
5. Movement Prediction: A system that analyses the drone's speed, direction, and trajectory to estimate its future position. It should provide movement forecasts to help operators assess potential risks and monitor changing flight paths.
6. Operator Display: An intuitive interface displaying live camera feeds, highlighted drone detections, tracked-target information and directional indicators. Operators should be able to select targets, review alerts and retain manual control at all times.
7. Testing and Validation: The system should be evaluated using recorded videos, computer simulations and real-world camera footage. Its performance should be measured based on detection range, classification accuracy, response time, tracking reliability, and positional estimation accuracy.
Key skills
How to apply
- Sign in or create your free profile — takes about two minutes.
- Apply with one tap — your profile is shared directly with HQ ARMY TRAINING COMMAND.
- Track your application and get updates by email and WhatsApp.