AI-Powered Mine Detection

Case Studies

Automated detection of mines and missile remnants using drones equipped with multispectral sensors

Defense

Industry

Ukraine

Location

AI & Computer Vision Development, Drone Data Pipeline & Sensor Integration, Model Optimization & On-device Inference, Mobile Development, MLOps & Deployment

Services

$85,000 – $95,000

Budget

Challenge

Post-conflict territories face severe threats from unexploded landmines, resulting in civilian casualties and obstructing reconstruction efforts. Manual demining methods are dangerous, slow, expensive, and environmentally destructive.

See what we can do for you

Outcomes We Deliver

Solution

The solution operates offline, provides top-3 identification suggestions, captures geocoordinates, and includes a feedback loop for continuous model improvement – all within a one-minute processing window.

Let’s talk about what’s possible

Tech Stack

To develop and deployed an AI-powered aerial system for landmine and UXO detection, Blackthorn AI applied a production-grade tech stack including:

Python

Dart

PyTorch

TensorFlow Lite

Python

Dart

PyTorch

TensorFlow Lite

Roadmap

Project duration

01–04 Weeks

Research & Dataset Development

Gathered aerial imagery of landmines from public sources; analyzed limitations of existing DJI drone platforms.

05–08 Weeks

Data Augmentation & Baseline Model

Built and trained a baseline CNN on expanded, augmented datasets with varied terrains.

09–16 Weeks

Drone Integration & Data Pipeline Setup

Integrated DJI drones with custom AI system; conducted live field tests and optimized the image processing pipeline.

17–24 Weeks

Model Optimization & Testing

Trained and validated a high-performance deep learning model with real-time inference capabilities.

25–30 Weeks

Deployment & Field Evaluation

Deployed the final model with offline support, feedback loop, and real-time UXO detection across diverse environments.

01–04 Weeks

Research & Dataset Development

Gathered aerial imagery of landmines from public sources; analyzed limitations of existing DJI drone platforms.

05–08 Weeks

Data Augmentation & Baseline Model

Built and trained a baseline CNN on expanded, augmented datasets with varied terrains.

09–16 Weeks

Drone Integration & Data Pipeline Setup

Integrated DJI drones with custom AI system; conducted live field tests and optimized the image processing pipeline.

17–24 Weeks

Model Optimization & Testing

Trained and validated a high-performance deep learning model with real-time inference capabilities.

25–30 Weeks

Deployment & Field Evaluation

Deployed the final model with offline support, feedback loop, and real-time UXO detection across diverse environments.

Team Size

1 x AI Solutions Architect

1 x Deep Learning Engineer

1 x Project Manager

1 x Flutter Developer

1 x MLOps Engineer

5 team members from Blackthorn.ai

1 x AI Solutions Architect

1 x Deep Learning Engineer

1 x Project Manager

1 x Flutter Developer

1 x MLOps Engineer

Delivering Impact

Beyond the values already highlighted, there’s even more to discover. Our commitment to innovation, client success, and impactful results sets us apart

Book a Meeting

22%

Improvement in detection accuracy

Reaching 92% vs. 70% with manual methods

20×

Faster area coverage

Scanning 10,000 m²/day vs. 500 m²/day manually

90%

Reduction in accidents

Down to 0.5 per 1,000 operations

2×

Easier to scale

Needing just a 50% resource increase to double coverage

80%

Cost savings

Reducing detection cost to $1 per m²

22%

Improvement in detection accuracy

Reaching 92% vs. 70% with manual methods

20×

Faster area coverage

Scanning 10,000 m²/day vs. 500 m²/day manually

90%

Reduction in accidents

Down to 0.5 per 1,000 operations

2×

Easier to scale

Needing just a 50% resource increase to double coverage

80%

Cost savings

Reducing detection cost to $1 per m²

Armed Forces of Ukraine

Automated detection of mines and missile remnants using drones equipped with multispectral sensors

Defense

Ukraine

AI & Computer Vision Development, Drone Data Pipeline & Sensor Integration, Model Optimization & On-device Inference, Mobile Development, MLOps & Deployment

$85,000 – $95,000

Post-conflict territories face severe threats from unexploded landmines, resulting in civilian casualties and obstructing reconstruction efforts. Manual demining methods are dangerous, slow, expensive, and environmentally destructive.

Solution

To develop and deployed an AI-powered aerial system for landmine and UXO detection, Blackthorn AI applied a production-grade tech stack including:

Project duration

01–04 Weeks

05–08 Weeks

09–16 Weeks

17–24 Weeks

25–30 Weeks

01–04 Weeks

05–08 Weeks

09–16 Weeks

17–24 Weeks

25–30 Weeks

Team Size

Delivering Impact

22%

20×

90%

2×

80%

22%

20×

90%

2×

80%

Discover More Related Case Studies