Parking Space Identifier & segmenter

Automated computer vision system to identify legal and unmarked parking spaces on the side of Indian roads. Indian roads do not have marked parking spaces on every road but most roads have legal parking spaces which the driver must identify by himself. While working for Kriti Labs, an IoT company that undertook this project, I developed an automated parking space identification system, which aided the Chennai government.

🔑 Key Feature

• Dynamic parking space identification based on location (when there is a lack or marked borders)
• Customized instance segmentation for legal parking spaces
• Automated parking space segmentor to accomodate bikes
• Identification of legal/illegal parking spaces in real time

💻 Sample Output

Parking space identification:

Parking space segmentation to accomodate bikes:

🛠 Skills

Python , Computer Vision , Deep Learning , Yolo models, Instance segmentaion

📖 Libraries Used

Yolo, OpenCV, TensorFlow, PyTorch, and CUDA, Roboflow

📃 Lessons Learnt

• Real time instance segmentation
• Creating custom instance segmentation dataset
• Creating custom dataset for situation based parkign space identification