Application of Deep-learning-based Transmission Lines to Prevent External Disruption--Detection of Large Construction Vehicles under Transmission Lines

Transmission is an important component of the power system. It constitutes the overall function of the power system together with power generation, substation, power distribution and power consumption. Through the transmission line, the power plants and load centers that are far apart are connected, so that the development and utilization of electric energy exceeds the geographical limit. Transmission lines can be divided into overhead transmission lines and underground transmission lines according to the structure. The former consists of line towers, wires, insulators, etc., which are erected on the ground, and the latter are mainly laid on underground or underwater.

At present, long-distance power transmission mainly adopts overhead transmission lines and undertakes huge power transmission. The safe operation of overhead transmission lines is closely related to the stability of the power grid, and operation and maintenance work is particularly important. The overhead transmission line has the characteristics of special geographical location, complex operating environment, wide operation and maintenance working points, and complicated tasks. It is very necessary to build a technologically advanced online monitoring system to prevent damage to transmission lines and provide an effective means of securing the safe operation of the line.

In recent years, with the rapid development of urban construction, construction operations have continued everywhere, and external engineering construction has brought great hidden dangers to the safe and stable operation of transmission lines. Construction workers lack power knowledge and evade approval for the construction period. The phenomenon is not uncommon. In various types of faults such as external force damage of transmission lines, illegal construction of large vehicles has become a major factor in the external damage of transmission lines. Due to the sudden and dispersive nature of large-scale vehicles and mechanical construction damage accidents, the power operation department is invincible.

Second, the prevention of external situation

In view of the prevention of external transmission lines, the power sector has been doing its best to make the best use of it. In the comprehensive management of transmission lines, the use of scientific management thinking, comprehensive human defense, physical defense, technical defense and other scientific means A clear, well-defined protection system. In terms of scientific and technological means, the power sector continues to promote the development of online monitoring products and systems for transmission lines. With the technological development of high-tech enterprises and in-depth research of business, some professional online monitoring technology companies have successively developed transmission lines with complete functions. The monitoring system is gradually applied to the actual support of the transmission line.

In the online monitoring system of transmission lines, in view of the illegal construction of large vehicles, it is generally the routine front-end equipment to capture pictures or videos to the back-end center for human analysis, which requires personnel to continuously pay attention and judgment for a long time, and the labor cost is huge, and Man-made efficiency and behavior are difficult to constrain, and it is impossible to ensure that large-scale vehicle illegal construction can be discovered in time.

At present, some enterprises can automatically identify large construction vehicles through an online monitoring system, mainly through pattern recognition technology (a process of classifying the object based on its observation value), and the application of such technology is reduced. The labor cost of the enterprise will increase the efficiency of the company. However, pattern recognition is limited by its own technical bottleneck, the false positive rate is high, the accuracy is very low, and when the training data reaches a certain amount, the recognition effect can no longer be further improved.

Third, based on deep learning, transmission line anti-extrusion application

The concept of "deep learning" originates from the research of artificial neural network. It is a new field in machine learning research. The motivation is to establish and simulate the neural network of human brain for analysis and learning. It mimics the mechanism of human brain to interpret data. In a vivid way, it can be understood that deep learning is a process that mimics human beings when dealing with problems, from shallow to deep, from simple to difficult, from intuitive to abstract. The effect of the traditional machine learning algorithm can not be further improved after the amount of data reaches a certain amount. Compared with traditional machine learning, deep learning can learn with independent features and can achieve better results under the driving of big data.

Based on massive security data, Hikvision has been working hard on deep learning for many years. In 2016, Hikvision launched several series of deep learning-based smart applications, such as “squat”, “super brain” and “blade”. "Facebook" "Falcon" and other equipment. Hikvision has an in-depth understanding of the anti-outbreak business of transmission lines, fully researching, understanding and absorbing, to quickly discover illegal vehicles under transmission lines and solve the problem of difficult management of transmission lines in violation of construction lines, to ensure the safe operation of transmission lines and reduce power. For the purpose of the department's operational input costs, the application of deep-learning-based transmission lines to prevent large-scale vehicle external force damage has been introduced.

The application consists of two systems: front-end and back-end. The front-end consists mainly of Hikvision's low-power intelligent network dome. The back-end consists of Hikvision's deep-learning-based blade analysis equipment and visualization service platform. The application captures pictures through multiple preset positions of the transmission line front-end monitoring device, and the back-end receives the picture transmitted from the front end to automatically analyze whether there are large construction vehicles in the scene. Once a large construction vehicle is detected, the alarm is automatically presented to the management personnel on the service platform. The application is described in detail as follows:

Application diagram

Application of Depth Prevention of Transmission Lines Based on Deep Learning——Detection of Large Construction Vehicles under Transmission Lines

1) Front-end monitoring:

The front-end monitoring system mainly consists of Hikvision low-power smart ball machine, which has the following characteristics:

 wide voltage power supply design, the minimum power consumption of the ball machine ≤ 2.6W;

 Support timing capture, plan capture, and capture multiple preset positions can be set;

 Integrated 4G wireless module, compatible with 3G/4G wireless network transmission;

The wide voltage and low power design is mainly designed for the unstable design of power supply solar panels. It integrates 4G wireless transmission module, is compatible with a variety of wireless standards, guarantees network transmission, supports timing, planning and multi-preset capture, and ensures a monitoring. The point meets multiple monitoring directions.

2) Analysis of the back-end blade image:

“Blade” is a device that Hikvision has launched for intelligent structured analysis and processing based on pictures, embedded with deep learning algorithms developed by Hikvision. The product series is mainly located in the center intelligent structured processing, fully meeting the needs of massive image structured analysis processing.

In the anti-external application of transmission lines, the blade equipment performs intelligent structural analysis and processing on the pictures transmitted at the front end, deeply excavates the key information in the vehicle pictures, identifies the specific types of vehicles, and provides data basis for identifying the application analysis of large vehicles.

3) Visual business platform system

Transmission line anti-external visualization business platform, combined with Hikvision deep learning intelligent analysis equipment, positioning transmission line to prevent large-scale vehicle detection applications, meeting transmission line multi-monitoring points, single monitoring points, multiple scenes, large-scale construction vehicle detection, automatic Identify large-scale construction vehicles in the transmission line monitoring scene. Once identified, the alarms will be presented to the platform management personnel. The platform management personnel can make judgments and judgments, analyze whether the violations and possible hazards, timely confirm and stop the violations, and guarantee production. Safety.

Fourth, the conclusion

Under the impetus of scientific development, the security industry is changing with each passing day, and the wave of deep learning is pushing the security artificial intelligence to the cusp. In the future security industry, there must be hundreds of battles and thousands of sails. How to apply artificial intelligence to security lines? All industries are issues that every security company needs to consider, leading or being towed?

As the bloodline of the power system, transmission lines play an important role in preventing and stopping the occurrence of external transmission lines, which is a long-term and arduous task. Applying artificial intelligence to the transmission line to prevent external damage and improving the diversification and intelligence of security measures is fundamentally the development of artificial intelligence to the field of transmission lines. The expansion of application methods has left the accumulation of unskilled, and the expansion of application innovation needs to better integrate the business. We are willing to work tirelessly.