Machine Learning and Intelligent Vision for the Industrial EdgeOctober 22, 2020 by Jeff Steinheider, NXP Semiconductors
NXP’s i.MX 8M Plus applications processor enables machine learning and intelligent vision for consumer applications and the industrial edge. Learn about the features of this processor and how it can be used in embedded vision systems.
The increase in automation of manufacturing and industrial processes creates more demand for intelligent vision-based systems. These systems capture visual input and then use a variety of processing techniques to make decisions based on that input. These vision systems are often integrated in larger control systems, and so they must be capable of real-time communication with other devices.
NXP’s i.MX 8M Plus applications processor enables machine learning and intelligent vision for the industrial edge and a wide range of other applications. In this article, learn more about the features of the i.MX 8M Plus applications processor and how it can be used in embedded vision systems.
The Benefits of Edge Computing
When utilizing cloud computing, the user relies on computational resources outside the bounds of their local network. With edge computing, much of the processing is brought back within the local network’s bounds, and as a result, sensitive data can also be kept within the local network.
Edge devices can perform a wide variety of tasks. In the cloud computing paradigm, they were often utilized for filtering, pre-processing, and storing or buffering the data. New advancements, such as integrated neural processing units (NPUs), open possibilities of what can be accomplished in edge devices. Edge devices that can make decisions based on a variety of data sources, such as camera input, and enable users to keep essential data within the local network. This reduces the amount of data that must be uploaded to the cloud and increases the system’s overall reliability and security. It can also enable faster real-time decision making, as transmitting data to the cloud and waiting for control responses adds latency, preventing a cloud computing architecture from being able to address some applications.
Figure 1. NXP’s i.MX 8M Plus Applications Processor Evaluation Kit
With an embedded NPU, an edge processing system can discard unwanted or noisy results from an input device, such as a camera, and only upload relevant data sets to the cloud. This saves bandwidth from the local site and reduces the amount of storage and processing required in the cloud, which, in turn, reduces the long-term operating cost.
The i.MX 8M Plus Applications Processor
NXP helps to enable vision-based applications at the edge with the new i.MX 8M plus applications processor by integrating two MIPI CSI camera interfaces and dual camera image signal processors (ISPs) with a supported resolution of up to 12 megapixels, along with a 2.3 TOPS neural processing unit (NPU) to accelerate machine learning.
The integrated ISPs reduce the system’s overall complexity, cost, and power requirements while offering an optimal imaging solution, particularly at two megapixels and higher resolutions. The multimedia capabilities of the i.MX 8M Plus SoC also includes video encode and decode, 2D and 3D graphic acceleration, and audio and voice functionalities.
The i.MX 8M Plus processor brings in communication interfaces such as two gigabit Ethernet controllers. One of them is time-sensitive networking (TSN) capable, while the second supports audio-video bridging (AVB). The device also includes dual CAN-FD interfaces, two USB 3.0/2.0 ports, one PCIe gen 3.0 interface, and three SDIO 3.0 controllers.
The i.MX 8M Plus SoC makes use of heterogeneous computing, by providing developers with several different programmable processing cores. The Arm® Cortex®-A53 cores are enabled with Yocto Linux®, providing enablement for application developers. The integrated M7 core provides real-time performance, or it can be used to enable low power operating modes.
Figure 2. This evaluation board showcases the wide variety of available ports to connect the i.MX 8M Plus to existing equipment.
Many manufacturing and controlling applications require high reliability. To ensure high reliability, NXP included DRAM inline ECC support and ECC on on-chip RAM to detect and correct memory errors.
Putting all of the features mentioned above together, the i.MX 8M Plus applications processor is well suited to enable any vision-based embedded system at the edge. These systems can range from quality control mechanisms in a factory that detect faulty items on a production line, to intelligent smart home controllers, for example, an automated climate and security controller.
Figure 3. The i.MX 8M Plus Processor Works in a Wide Variety of Applications.
Extreme Operating Conditions, Longevity, and Reliability
Many industrial use cases require a processor that can run in the field for extended periods of time. The i.MX 8M Plus applications processor is industrial qualified, capable of running for up to 10 years in an extended temperature range from -40°C up to 105°C.
In addition, the device is a part of the NXP 15-year longevity program, guaranteeing supply for 15 years after product launch. This makes the i.MX 8M Plus processor work for applications that require much longer product life cycles, either due to certification requirements or extended software development timelines.
The i.MX 8M Plus applications processor is well-suited to be utilized in a wide variety of applications, ranging from industrial-grade controllers to energy-efficient consumer devices. Some of the target applications include industrial HMI, automation controllers, machine vision systems, medical equipment, home security and surveillance, and fleet management. Find a list of all the target applications of the i.MX 8M Plus on the i.MX 8M Plus processor product page.
The following three application notes highlight a few use-cases and studies:
- Using Code-Signing Tool with Hardware Security Module
- Strengthening Public Key Cryptography using CAAM Secure Key
- i.MX 8M Plus NPU Warmup Time
There are also a few boards available that allow the evaluation of the i.MX 8M Plus SoC:
Industry Articles are a form of content that allows industry partners to share useful news, messages, and technology with All About Circuits readers in a way editorial content is not well suited to. All Industry Articles are subject to strict editorial guidelines with the intention of offering readers useful news, technical expertise, or stories. The viewpoints and opinions expressed in Industry Articles are those of the partner and not necessarily those of All About Circuits or its writers.