Automated Transmission Tower Inspections (AI + Drone Integration)

Developed an AI-powered system for autonomous inspection of electrical transmission towers using drone-captured imagery, with a focus on enabling real-time component detection and reducing reliance on manual inspection processes. The project involved end-to-end development across data processing, model training, and hardware integration in a real industry setting.

Key Contributions & Technologies:

● Designed and implemented deep learning-based object detection models to identify and classify transmission tower components from aerial images.
● Worked on data preparation, annotation, and preprocessing pipelines to support robust model training and evaluation.
● Integrated the trained models with NVIDIA Jetson Orin Nano for edge-based inference, enabling real-time detection capabilities.
● Contributed to DJI drone hardware integration, supporting live video data acquisition and system-level deployment.
● Led coordination of development efforts, including SDK setup, hardware configuration, and system integration tasks.

Research & Development Approach:

● Executed the project through multiple development iterations, focusing on improving model accuracy, reliability, and deployment readiness.
● Collaborated directly with industry professionals including Tyler Vey and Eric Aylward (SEM Ltd.), gaining exposure to real-world engineering workflows.
● Presented the final solution to Stephen Rowe (Executive Vice President, SEM), demonstrating practical application and system performance.

Tools & Technologies:

● Python, TensorFlow / PyTorch
● Computer Vision & Object Detection
● NVIDIA Jetson Orin Nano (Edge AI)
● DJI SDK & Drone Systems
● Data Annotation & Model Evaluation

Outcome & Future Plans:

Successfully developed a functional AI-driven inspection system capable of detecting structural components from drone imagery, demonstrating strong potential for real-world deployment.