Founder and CEO of Flash Ma Intelligent: "Camera+Radar" fusion to achieve multi -mode visual perception, making traffic management more accurate

Cover reporter Wu Yujia

On September 1, the 2022 World Artificial Intelligence Conference was held. Top scientists and scholars from the artificial intelligence fields from various countries, as well as well -known entrepreneurs from all walks of life gathered at this highly international AI industry event, surrounding "human, technology, industry , Cities, and the Five "five elements, embrace the future urban maps of the Yuan universe era, in -depth interpretation of the theme of this conference," Zhilian World, Yuansheng has no boundaries. "

Focusing on the application and practice of AI in smart transportation, the founder and CEO of Flash Ma Smart shared their views. Peng Yan believes that "AI itself is not an industry. It needs to be combined with all walks of life to transform technology from productive forces, with technical service scenarios, from the demand, the application of the technical service industry is particularly significant at the moment."

It is understood that the use of cameras to identify traffic events in the past did not apply to special circumstances such as rainy days, nights, and fog. Some cities will use radar instead of cameras. Although better recognition effects are achieved under low energy visibility, it is usually impossible to display the rich information contained in the real scenes like photos. In addition, traditional radar can provide exercise and location information for all close entities, but it is difficult to reliably distinguish the differences between different modes.

How to solve the pain point? Peng Yan said that using the fusion of "camera+radar" to achieve multi -mode visual perception, it can accurately solve the problems encountered by traditional camera monitoring traffic modes and achieve more reliable traffic management. Adopting the overall fusion+target level integration scheme, based on the "cloud+edge" integration computing mode to create a cloud -edged fusion structure, it can be applied to the intersection, road section, tunnel, elevated and other electronic police cameras, bayonet cameras, monitoring balls and monitoring HD videos such as gun machines connect to the edge wisdom box equipment, obtain multi -dimensional information in real time, and break through traditional barriers, making traffic management more accurate and efficient.

Multi -mode state is widely used in smart transportation. Whether it is past, now or in the future, Peng Yan said that they will continue to iterate multi -modal AI traffic algorithms to strengthen multi -modal visual perception. Combining data analysis, it will create a digital base for application scenarios. The original governance model is gradually transformed and upgraded to build a future city.

- END -