Dirt Detection on floor usage in robotic vaccum cleanerEdgeSync Technologies
Developed a computer vision-based dirt detection system using YOLOv8 (You Only Look Once v8) to identify and classify dirt patches on floor surfaces in real time. The model was trained to detect the presence of dirt and further classify it as wet or dry using labeled image datasets captured under diverse lighting and surface conditions.
Based on the classification output, the robotic vacuum cleaner dynamically adjusts its cleaning mode—activating scrubbing/wet mopping for wet dirt and suction mode for dry debris—optimizing cleaning efficiency and resource usage.
Key Technical Skills & Technologies:
Deep Learning / Object Detection
YOLOv8 architecture (Ultralytics)
Custom dataset curation, annotation (COCO/Pascal VOC format)
Transfer learning and model fine-tuning
Data augmentation (brightness, blur, noise)
Python, PyTorch, OpenCV
Model deployment on edge devices (NVIDIA Jetson/ROS environment)
Real-time inference and decision logic integration
This intelligent classification enables adaptive cleaning behavior, improving autonomy and performance in home and commercial environments.
What's included
Dirt Detection for Robotic Vaccum Cleaner
Usages YOLOv8 and Computer vision to develop the model to detect the dirt and predict the dirt is wet or dry.
Contact for pricing
Service provided by
EdgeSync Technologies Kolkata, India
Dirt Detection on floor usage in robotic vaccum cleanerEdgeSync Technologies
Contact for pricing
Developed a computer vision-based dirt detection system using YOLOv8 (You Only Look Once v8) to identify and classify dirt patches on floor surfaces in real time. The model was trained to detect the presence of dirt and further classify it as wet or dry using labeled image datasets captured under diverse lighting and surface conditions.
Based on the classification output, the robotic vacuum cleaner dynamically adjusts its cleaning mode—activating scrubbing/wet mopping for wet dirt and suction mode for dry debris—optimizing cleaning efficiency and resource usage.
Key Technical Skills & Technologies:
Deep Learning / Object Detection
YOLOv8 architecture (Ultralytics)
Custom dataset curation, annotation (COCO/Pascal VOC format)
Transfer learning and model fine-tuning
Data augmentation (brightness, blur, noise)
Python, PyTorch, OpenCV
Model deployment on edge devices (NVIDIA Jetson/ROS environment)
Real-time inference and decision logic integration
This intelligent classification enables adaptive cleaning behavior, improving autonomy and performance in home and commercial environments.
What's included
Dirt Detection for Robotic Vaccum Cleaner
Usages YOLOv8 and Computer vision to develop the model to detect the dirt and predict the dirt is wet or dry.
Contact for pricing