Leopard, Sony AI imaging for smart cities

  • August 6, 2024
  • William Payne

Intelligent vision specialist Leopard Imaging has collaborated with Sony Semiconductor (SSS) on an AI vision system for smart cities and industrial automation. The GS500 AI vision system is powered by the Sony edge AI sensing platform AITRIOS.

The system also includes the GS500-MIPI camera equipped with the Sony IMX500 intelligent vision sensor. With a 12-megapixel sensor, the GS500 AI vision system is designed to decipher scenes and handle machine-learning tasks directly on the camera. The system incorporates edge AI and a scalable architecture for speed and security.

Installation and maintenance are supported with network options such as PoE, WiFi, and LTE-CatM1, along with OTA AI and app updates.

Data minimisation is a feature with Sony’s IMX500 sensor. This helps ensure data anonymisation at the front edge and implements top-level data protection.

Images are processed to derive metadata at capture, and image capture occurs only when necessary. This mechanism reduces data storage and transfer, making the GS500 AI vision system a secure and reliable choice for various applications, including smart city infrastructure, robotic vision, and industrial automation.

“Leopard’s new solution, leveraging Sony Semiconductor Solutions’ IMX500, creates a versatile, outdoor, edge AI camera for smart city use cases like traffic counting, curb management, parking occupancy, pedestrian safety, and more. We are pleased to welcome another camera partner into Sony’s growing AITRIOS ecosystem, expanding offerings for our customers that work under a variety of IoT environments.” – Yu Kitamura, Program Manager for Smart Cities, Sony Semiconductor Solutions

In 2023, Leopard Imaging won first place in the Pedestrian Safety Challenge Hackathon Competition co-sponsored by the tinyML Foundation, and the City of San José, with technology partner, Sony Semiconductor Solutions. This competition laid the foundation for the GS500 AI system research and development.