In HAS 2018, Yujun (Monica) Wang, APAC Leader of Safe Cities & Video Surveillance Research at IHS Markit, shared her opinions on the development trends of intelligent video surveillance cameras.
Q1. What do you think of the camera market’s development, from your perspective as an analyst?
A: According to IHS Markit, the global shipment of professional network cameras in 2017 was about 80 million. From 2017 to 2022, the Compound Annual Growth Rate (CAGR) will increase by 40 percent. The main powerhouse for the market growth comes from two aspects. One is emerging markets’ growing demands for new installment projects — for example, in India, South Asia, and South America. The other is the ever-increasing demands to upgrade standard cameras to smart ones. So now, security cameras are used to ensure security as well as to implement business intelligence, city management, and other IoT applications.
Q2. How will cameras evolve in the future? For example, their intelligence, computing power, and multi-algorithm capability?
A: A decade ago, the market expectations were too high on video analytics. At that time, it was predicted that video analytics would maintain triple-digit growth over many years. However, certain factors — such as the high false positive rate, installation difficulties, and the small number of success cases — affected the market growth. After that period, the market began to bounce back, and suppliers set more-realistic expectations for customers. Video analytics is used as an approach to implement automatic surveillance and has been applied throughout the market. As a next-generation product, it has restored market confidence. In cameras with excellent computing performance, their chips support diverse deep learning algorithms. Therefore, the chip is the key driver to improve cameras’ video analytics capability. It is predicted that more than 50 percent of security cameras will have basic video analytics capability by 2022.
Q3. Huawei recently launched its software-defined camera. Can you discuss your opinions on the camera from your perspective as an analyst?
A: For smart cameras, the mission-critical task is to enable sustained algorithm upgrades. With a traditional camera, if users want to switch among different vendors’ algorithms they need to replace the camera. In a world with ever-changing technologies and functionalities, this approach is impractical because it hinders end users ability to try new technologies and requires sunk costs for specific manufacturers and proprietary interfaces. This is why we need an open-ended platform and camera to integrate different deep learning algorithms fast by decoupling software from hardware. The software-defined camera provides an open-ended platform, so the upgrade or replacement of applications and algorithms is effortless.