Venkatesh Kumar Raju

About Me

I am a passionate and results-driven GIS professional who loves transforming complex geospatial data into actionable insights. My journey into GeoAI and machine learning was fueled by a desire to push the boundaries of what's possible with location-based data.

As a certified drone pilot, I capture unique perspectives of the world from above and turn that data into detailed 3D maps and analysis. When I'm not working with maps or training models, I'm exploring new technologies, contributing to open-source projects, or flying my drone.

My Expertise

Specialized expertise in geospatial, drone photogrammetry, and AI domains.

🎯

Object Detection

Custom YOLO (v8–v12) models to detect and classify objects in images and videos.

🧩

Image Segmentation

Pixel-level classification and precise object masking for fine-grained analysis.

⛰️

Topographic Mapping

Detailed maps of the Earth's surface with elevation and feature data.

🗺️

Planimetric Mapping

Maps showing horizontal positions of features without elevation.

🚁

Drone Operations

Certified aerial surveys and data acquisition with UAVs.

🛰️

Drone Image Processing

Orthomosaics, 3D models, DSM/DTM from aerial imagery.

🌐

Web App Development

Custom web apps to visualize and interact with geospatial data.

My Skills

Technologies and tools I'm proficient with.

Geospatial

PhotogrammetryDTMDSMGCPTopographyOrtho Map3D MappingImage ProcessingSpatial Data

Software

ArcGIS ProQGISAutoCAD Map 3DDrone2MapWebODMDronelink

AI & Programming

PythonMATLABOpenCVPyTorchTensorFlowKerasFlask

Languages

TamilEnglishFrench (A2)

Work Experience

My professional journey and key accomplishments.

Oct 2024 – Present

GIS Developer & ML Engineer

SeaMount Geospatial LLC (Remote) ↗

Led development of the CORAL INTEL platform ↗
Data acquisition & preparation, annotation pipeline management
YOLO model implementation (v8–v12)
Web application development with Flask

Nov 2022

GIS Digitization Intern

MAPe IT Solutions Private Limited ↗

Mapped rooftop solar panels, swimming pools, and structural footprints with topology validation

Aug 2021 – Mar 2023

Placement Representative

Anna University

Facilitated communication between students and the placement cell

Mar 2019 – Dec 2019

Junior Photogrammetrist Engineer

Geo Adithya Technologies

Digitized topographic maps and digital surface models

Featured Projects

Interactive portfolio of GeoAI, drone photogrammetry, computer vision, and geospatial engineering. Click any card to explore full architecture & details.

Production • SeaMount

🪸

Coral Intel Platform

State-of-the-art deep learning platform for automated coral reef health assessment, substrate segmentation, and bleaching detection from underwater ROV/UAV imagery.

YOLOv8–v12 •Achieved an mAP50-95 of 0.86% on the object detection task and an mAP50-95 of 80.7 % on the segmentation task

PyTorch YOLOv8-v12 Flask OpenCV Annotation Pipeline

Photogrammetry

🚁

Drone 3D Mapping & Orthomosaic

End-to-end UAV photogrammetry workflow transforming raw aerial imagery into survey-grade orthomosaics, Digital Surface Models (DSM), DTM, and contour datasets with millimeter GCP accuracy.

Sub-2cm GSD • Full 3D Point Cloud

WebODM Drone2Map ArcGIS Pro QGIS GCP Alignment

Computer Vision

🚘

Real-Time ANPR & YOLO11 Detection

High-speed object detection and Automatic Number Plate Recognition (ANPR) system powered by YOLO11 and OpenCV for real-time vehicular surveillance.

YOLO11 • 45+ FPS Inference •Achieved an mAP50 91.1% and mAp50:95 68.1% OCR

YOLO11 OpenCV PyTorch OCR Vehicle Analytics

Augmented Reality

✨

Virtual Makeup & Facial AR (Sunglasses & Lipstick)

Real-time computer vision AR application that applies virtual sunglasses, dynamic lipstick, and cosmetic overlays using 68-point facial landmark localization with OpenCV and Dlib.

68 Landmarks • Virtual Sunglasses & Lipstick

OpenCV Dlib Virtual Sunglasses Lipstick AR Mesh Warping

Internship • MAPe IT Solutions

☀️

Urban Infrastructure & Solar Digitization

Large-scale aerial & satellite feature extraction detecting solar panels, swimming pools, and structural footprints with geometric topology validation, completed during GIS internship at MAPe IT Solutions Private Limited.

10,000+ Assets Vectorized • Clean Topology

ArcGIS Pro QGIS AutoCAD Map 3D Satellite Imagery Geodatabase

Publications

Peer-reviewed research in deep learning for marine ecosystem monitoring.

📚 Intelligent Marine Technology and Systems

Jun 2026

Coral Intel: a YOLOv8 deep learning framework for monitoring Caribbean corals

This study details the development and evaluation of a YOLOv8-based deep learning model ("Coral Intel") for identifying 59 Caribbean coral taxa from underwater imagery, achieving high precision (mAP50 of 0.954 at genus level) in controlled testing but lower performance during independent field validation due to factors like image perspective differences and taxonomic similarities. The framework is currently available as a prototype web application intended to support scalable reef monitoring efforts, offering a foundation for regionally specific coral detection models.

Read on Springer →

📚 Journal of Marine Science and Engineering

Sep 2026

Automated coral detection and instance segmentation in underwater imagery: Evaluating YOLO model performance and transferability to mesophotic coral ecosystems

Automated coral image analysis is difficult due to colony resolution and taxonomy needs. We trained YOLO-based instance segmentation on 9,530 annotated Caribbean coral images (57 species, 22 genera). YOLOv8x and YOLOv11x achieved the best balance, with species mAP₅₀ = 0.854 and genus mAP₅₀ = 0.883 at 4–6 ms inference. External transfer showed higher genus-level precision.

Read Preprint →