Object Recognition in Smart Cities: Lattice Semiconductor's Smart City Drive

By Len WegnerSeptember 21, 2017

At one point or another in our lives, many of us have been fascinated by robots, especially the way these machines could potentially see, hear, and do things with human-like abilities. Today, with the advancement of smart technology and the emergence of mobile-driven applications, this is increasingly becoming a reality through smart homes and even entire smart cities.

The Vision of Smart City

The concept of smart cities is primarily based on the integration of networks and sensors through a string of seamless electronic technologies. The idea is to use these networks to create an augmented reality so that cities can operate smoothly and safely while saving time, money, and energy. 

Smart city networks, besides controlling and monitoring everyday operations and services such as water consumption, waste management and traffic flow, will continuously collect data to improve the quality of life by managing energy use, regulating power generation and storage, handling pollution and providing enhanced safety and security to citizens in times of emergency.

The Object Recognition Landscape

The widespread adoption of internet-enabled smart devices such as tablets, smartphones, smartwatches, and smartglasses is pushing the concept of smart city forward, where users can actually interact with objects, machines, and cities through their mobile applications.

Depending on their field of view and position, users can also receive data on their smart devices in augmented reality. For instance, information on shops, restaurants, monuments, or hotels can be available at your fingertips by simply steering a smart device towards the subject of interest. This is made possible through augmented reality systems such as GPS and compass, which can display the relevant information by determining the field of view and position of the user.

These existing systems have limitations, however, and can only show marked places like famous archaeological sites, prominent landmarks, or large-scale infrastructure, provided the user is precisely positioned close to the subject of interest.

Technologies like compass and GPS lack the accuracy to precisely display information, and therefore they cannot be relied on as a stable solution that can leverage the full potential of augmented reality. The problem intensifies when users are positioned in random locations, or when it comes to recognizing objects that have smaller dimensions, like paintings, statues, or, in the case of smart cities, a house in a certain location. In these cases, compass and GPS malfunction due to the high inaccuracies inherent in these legacy technologies.

Today, new technologies are rapidly emerging that can precisely determine the location of a user by leveraging state-of-the-art, content-based image retrieval and object recognition techniques. With these revolutionary technologies, the point of view of a user, with respect to the object observed, can be accurately determined by aligning the scene being observed by the user with a pack of pre-aligned virtual scenes like images, videos, 3D model, or animations.

To render a virtual scene in real-time, and with respect to the correct position and perspective, it is crucial to recognize the relevant position of the user and the particular object the user is looking at. At this point, object recognition technology plays a pivotal role in enhancing the user experience with precision and fluidity.

Lattice Semiconductor's Smart City Drive

Lattice Semiconductor, a global leader in customizable smart connectivity solutions, has launched its new ECP5 FPGA family that supports low power, flexible connectivity demands of smart city applications.

Lattice Semiconductor's NEX-ADAS 360o 3-D Surround View Monitoring Technology, developed in partnership with Shenzhen Moorechip Technologies Limited, joins images from four cameras to produce an up-to-the-minute 3D view of the real-world surroundings.

The NEX-ADAS 360o, leveraging advanced, high-class image-stitching technologies, promotes the intelligent traffic systems (ITS) by monitoring the traffic flow, traffic violations, and toll collections, while enabling object recognition features like number plate detection, vehicle identification, and smart parking.

NEX-ADAS 360o performs these high-end functionalities with a variety of intelligent systems that leverage Lattice's cutting-edge embedded vision technology. The new ECP5 FPGA family, a primary driver for NEX-ADAS 360o, reinforces Lattice's commitment to building the next-generation of smart cities with flexible connectivity and acceleration at the Edge.

With the advancement of intelligence at the Edge, Lattice's Embedded Vision Development Kit will inspire more innovative object recognition solutions through mobile-influenced technology in systems like machine vision, smart surveillance cameras, drones, and robotics.

For more information about Lattice’s products and the Embedded Vision Development Kit, contact WPG America’s Director of Intelligent Connectivity Solutions Len Wegner.