When it comes to autonomous driving, robot vision, and high-definition cameras, you must think of the camera unit. Previously, this man also talked about some depth cameras and binocular cameras for FPGA applications in high-definition cameras and machine vision. The role of FPGAs in them is mainly Processing the acquired image, the processing of the image requires hardware with good parallel performance, so its processing speed has no advantage compared with ARM CPU? This man is leading today to explore the application of FPGA in edge computing.
When it comes to computing speed, the first thing we can think of is cloud computing. Cloud computing has countless benefits, such as fast calculations, large amounts of computational data, and so on. But the wise man must have a loss. The manufacturer without cloud computing can guarantee that he can do anything. The cloud computing also has its own shortcomings. The biggest shortcoming is that the real-time calculation is not high enough. For example, there are always some delays when you send some data to the cloud for processing. The hardest thing is the time to wait for results. Edge computing requires high computational performance and high real-time performance. FPGA can handle many things in parallel at the same time. It can meet the requirements of data processing speed and real-time requirements. It is the best of both worlds. This year's four-camera ADAS model of the Aldec-based TySOM-2-7Z100 prototype board was presented at the Embedded Vision Summit in Santala, California, as shown in Figure 1. The performance of TySOM is very good, mainly because the core processing component inside is the XiC of the Xilinx Zynq Z-7100.
Fig. 1 TySOM-2-7Z100 prototype board
Figure 2 shows the location of Zynq in the TySOM board. Why does the FPGA in Zynq get such a good application in edge computing? The Zynq-7000 Programmable SoC puts a software-programmable ARM processor and a hardware-programmable FPGA in a single chip that enables digital analysis while also enabling hardware acceleration with integrated CPU, DSP, ASSP and mixed-signal processing. Module. The image processing uses the FPGA module inside Zynq. So what role can the ARM core play in the TySOM card?
The good performance of Aldec's TySOM-2-7Z100 prototype board relies on the dual-core ARM Cortex-A9 processor and an FPGA logic in Zynq. The whole process of image processing begins with the image acquisition by the camera, using an edge detection algorithm (the edge here refers to the perception of the physical edge, such as the boundary of the object or the lane, etc.). This is a computationally intensive task because there are millions of pixels that need to be calculated. The acquired image can only process 3 images per second if it is processed in the ARM CPU. However, in the FPGA, 27.5 images can be processed per second. It can be seen that FPGA plays an important role in Zynq. In other words, with the processing speed of FPGA images, the speed of nearly 10 times has increased.
Fig. 2 Front view of the TySOM-2-7Z100 board
It is not enough to have a high-performance core processing chip. It also requires a lot of peripheral interfaces to interact with other devices. The TySOM is designed to accommodate up to 362 I/O peripheral interfaces, 16 GTX transceivers, and two FMC-HPCs to support expansion daughter card ports. The basic standard interfaces required for the ARM CPU to process data such as DDR3 RAM, USB and HDMI; the ARM core can also support Linux operating systems and other types of real-time operating systems. Not only that, ARM's CPU has 1GB of DDR3 RAM to control it, and can support 32GB of SSD storage space. Network interaction can be achieved via the Gigabit Ethernet PHY via the RJ45 interface, and is also equipped with four USB 2.0 interfaces. Most FPGA interfaces interact with other devices through two FMC-HPC sockets interfaces. In this way, both the ARM core and the FPGA module can interact with the outside world.
Auto-driving is in full swing. With the slow recognition of autonomous driving by national policies, I believe that it is a good thing for technology. Both hardware and algorithms will have their own use. In smart cities, smart life. In the big environment, the development of FPGA will also move forward steadily with the trend, better to embrace change and create opportunities for change.
Paperlike Screen Protector For IPad
Paperlike Screen Protector,iPad Paperlike Screen Protector,Paperlike Screen Protector iPad
Shenzhen Jianjiantong Technology Co., Ltd. , https://www.mct-sz.com