autonomous image recognition algorithm based on

deep learning can flexibly apply regional focus, SIFT

feature matching, and pattern recognition to achieve

an excellent recognition success rate. The robot can

read the data of various meters, automatically

record and judge while driving, and raise the alarm

when needed. The robot automatically evaluates and

recognizes the environment digitally, the error rate

is less than 5%, and the success rate is over 95%.

Due to the highly compact size of the EC70A-SU, the

robot body has a small and exquisite design, which

is convenient for manual handling and is also

suitable for transportation by vehicles. In addition to

technology, customers also benefit from DFI's

excellent business promotion capabilities, technical

support capabilities, and ability to grasp and

respond to customers' needs in real-time fully.

Another Chinese customer, established in 2007, is

committed to supplying robots and artificial

intelligence-related products and overall solutions.

With cutting-edge technologies in related fields as

the core, it focuses on automatic robot control,

navigation and positioning, machine learning,

graphics, and image analysis, and developed various

innovative robot products widely used in power,

chemical, energy, and security railways, and other

fields.

Its self-developed intelligent inspection robot takes

the general robot technology as the core. It

integrates multi-sensor fusion technology, intelligent

image recognition and analysis technology,

autonomous navigation and positioning technology,

non-contact detection, and other technologies to

achieve total autonomy efficiency in all-weather

substations. As a result, intelligent inspection and

monitoring, while effectively reducing the labor

intensity of employees, improve the automation and

intelligence of operation and maintenance work. And

provide innovative applications for substations with

none or few people on duty.

The computing core of its miniature intelligent

inspection robot also uses DFI's EC70A-SU, which

has a sturdy and durable structure. It comprises

industrial-grade components inside and outside,

supports 15~36V DC wide-voltage power input, and

can operate stably at temperature ranges between

-20°C to 60°C. In addition, the onboard memory

has a stronger shock resistance, which meets the

durability of this miniaturized intelligent inspection

robot during movement. The EC70A-SU also

provides six COM ports with galvanic isolation

protection and supports RS-232/422/485 by

adjusting the jumper, which improves customization

flexibility.

Therefore, in addition to essential functions such as

autonomous navigation, automatic charging, sound

and light alarm, short message notification, and

intelligent inspection, it is also convenient to

customize for customers in different industries.

Furthermore, it can also integrate exclusive

functions, such as infrared temperature

measurement, three-phase temperature difference

judgment, meter recognition, leak detection, defect

detection, defect recognition, foreign body

recognition, and other functions to become essential

features of this miniaturized intelligent inspection

robot.

Finally, the EC70A-SU supports a 15-years long-term

supply of CPUs until the first quarter of 2030 and

has a diversified selection of processor models for

long-term deployment. Furthermore, the compact,

lightweight, and powerful EC70A-SU is used as the

computing brain of miniaturized intelligent

inspection robot, bringing higher value to the

unmanned substation.

DFI Application Story

Low Power, High Performance, and Multi-Interface Fanless Embedded System

Create the Trustworthy Substation Inspection Robot

Substation inspection robots mainly serve various

substations, concentrated primarily on large-scale

substations above 500kV. Therefore, they are used

mainly to inspect substation equipment. They are

divided into visible light inspections and infrared

inspections. The visual light inspection examines the

appearance of the equipment, the status of the

equipment from opening to closing, and the meter

reading explicitly. The infrared inspection is mainly

used to measure the temperature, observe whether

the equipment's temperature is in the normal range,

and analyze any abnormal temperature rise.

Early substation inspection robots belonged to

Autonomous Guided Vehicles (AGVs) that needed to

lay magnetic tracks in advance and moved forward

or backward along the magnetic rails. At each point

that the AGVs are required to be stopped for the

detection or turning and speed adjustment, an

additional RFID tag needs to be buried next to the

magnetic track. The RFID tags contain information

about the adjustment, such as detection, speed

adjustment, turning, etc. After the robot scans these

tags, it performs corresponding actions.

With the rapid development of autonomous

positioning and navigation technology, such as laser

radar navigation and real-time simultaneous

positioning and environmental map establishment

(SLAM, Simultaneous Localization, and Mapping)

algorithm technology, the environment does not

need to be deployed in planning (such as rebuilding

factories, laying track, and leveling the ground) for

emerging autonomous mobile robot (AMR).

Autonomous Mobile Robots (AMR) are commonly

used in different industrial fields because they do

not need to be deployed in pre-planned

environments (such as rebuilding factories, laying

tracks, leveling the ground) like Autonomous Guided