Lots of existing and emerging markets for edge AI processors fall into the multi-camera AI use case. Multi-camera AI is similar to AI Box. The difference is that multiple cameras are directly connected to the system through MIPI CSI-2 in multi-camera AI, while encoded bitstreams from remote cameras are streamed into the system through Ethernet in AI Box. The use cases of multi-camera AI include but are not limited to the following applications:

• Surveillance camera is one of the most popular multi-camera AI use cases. The applications of surveillance cameras include security, traffic monitoring, and so forth. For security, cameras monitor and record a specific area such as in a home, workplace, or public places to maintain safety of the area. For traffic monitoring, cameras are mounted at interactions, school zones, and frequently congested roads, and monitor traffic flow and accidents for the purpose of traffic management. In general, a surveillance camera system is connected to a network and the activities recorded are encoded and uploaded to the cloud for saving and for remote viewing.
• Mobile DVR is widely used to record the video footage from the cameras mounted inside and outside vehicles, which is useful for in-cabin monitoring, theft protection, and evidence in case of accidents. While mobile DVR is very similar to the surveillance camera system, mobile DVR saves the recorded video footage in the local storage devices such as a secure digital (SD) card or a solid-state drive (SSD).

Figure 3-3 shows the data flow for a multi-camera AI use case example on the AM69A. In this example, eight 2MP cameras are combined using two MIPI CSI-2 aggregators and captured at 30 fps through two MIPI CSI-2 RX ports, which is the main difference from AI Box. The images through ISP (Image Signal Processor), demosaic and lens distortion correction are scaled to smaller resolution by MSC. The outputs obtained through DL pre-processing, DL network on MMA, and DL post-processing are encoded by the hardware accelerated H.264 or H.265 encoder, and streamed out or saved in storage. Table 3-3 shows the resource utilization and estimated power consumption for the multi-camera AI use case with eight channels of a 2MP camera.