Home > Industry Insights > Analysis of Machinery and Equipment Industry: Market Prospect and Current Development Trend of License Plate Recognition System in 2022
Analysis of Machinery and Equipment Industry: Market Prospect and Current Development Trend of License Plate Recognition System in 2022
In parking lot management, license plate recognition technology is also the main means of identifying vehicle identity. In the "Technical Requirements for the Collection and Transmission System of Vehicle Image and Number Plate Information in Parking Garages (Yards)" constructed by Shenzhen Public Security Bureau, license plate recognition technology has become the main means of vehicle identification. The license plate recognition technology is combined with the electronic non-stop toll collection system (ETC) to identify vehicles. When passing vehicles pass through the crossing, it can realize automatic identification of vehicle identity and automatic toll collection. In the parking lot management, in order to improve the efficiency of vehicle traffic at the entrance and exit, license plate recognition is aimed at vehicles that do not need to charge parking fees (such as monthly trucks, internal free vehicles), and builds an unattended fast lane, which provides a card-free and non-stop access experience. Change the management mode of entering and leaving the parking lot.
In the late 1990s, the license plate recognition system was commercialized due to the needs of traffic management, but limited by the ability of the algorithm itself, the imaging level of the camera was low and the effect was not good, so only a small number of applications were obtained. The early license plate recognition technology has two main characteristics: one is based on traditional algorithms, and the other is based on soft algorithms on the PC side, and the application is limited by both.
After 2006, with the rise and maturity of machine learning, especially deep learning and front-end embedded algorithm deployment technology, as well as the substantial increase in chip computing power, the chip cost has been greatly reduced, and the mainstream form of products has changed from soft recognition to soft recognition. With the integrated camera for license plate recognition, the technology and application of license plate recognition have made great progress, especially in the past five years, license plate recognition has entered the era of comprehensive commercial use. From 2006 to 2014, license plate recognition was mainly used in dynamic traffic, including electronic police, highway checkpoints and highways. After 2014, it also ushered in a big explosion in static traffic, including parking lots, on-street parking and other scenarios. .
In 2018, the scene of the license plate recognition technology has been continuously subdivided, and complex scenes such as smart construction sites, smart gas stations, and unattended weighbridges have appeared. These scenes not only require higher performance of the license plate recognition machine, but also have a unique personality. Functional requirements, it is foreseeable that the license plate recognition integrated machine dedicated to the scene will become the development trend of license plate recognition technology in the future.
Almost every license plate recognition system claims to have a high recognition rate, but in order to avoid mutual shirking of the responsibility for poor operation due to differences in product knowledge between the two parties, users may wish to request a field test when purchasing a license plate recognition system, and the test The time is preferably more than two weeks, and it is better to judge whether the identification result is "exaggerated". Because of the changeable environment, I should be able to grasp about 80% of the situation that the field may affect the recognition rate in two weeks. If it is only measured for a day or even a few hours, it is impossible to understand.
In addition, since license plate recognition is a "system", the quality of the hardware and software architecture will of course affect the "presented results". As for what kind of software and hardware is suitable for what kind of environment, this must vary from environment to environment, because different application environments have different requirements for the recognition rate, and this must be accumulated through experience.
Although there are numerous license plate recognition systems on the market, using the right products and architecture can save a lot of wasted money and time, but more importantly, engineers and system integrators need cooperation and understanding from multiple parties, rather than blindly focusing on A certain brand is better and cheaper, so you won't lose anything by shopping around.
In addition, whether the license plate recognition system can play the greatest role, in addition to software technology, has a lot to do with cameras and on-site construction capabilities. The user can ask the manufacturer to conduct an on-site survey and propose a construction plan, first evaluate the location to be erected, the angle of the camera, whether an auxiliary light source needs to be erected, etc., and then submit a quotation. These actions can not only evaluate the ability of the operator in advance , the user can also achieve product learning and education training, and in the future management, will be more aware of the product's use restrictions and related countermeasures.
For example, a license plate recognition system has an accuracy of more than 90% during the day, but it drops to 80% in the evening, and then drops to 70% at night. This unstable system has an average accuracy of 70% compared to the all-weather license plate recognition system. More difficult to integrate. Because users will think that since the recognition rate during the day is 90%, it is reasonable to have an accuracy rate of 90% all-weather, and such specifications do not include strange environmental disturbances (rainstorm attacks, hail, dense fog sections, etc.) , and erection environmental restrictions (height restrictions, wind and sway restrictions, not easy to suffer from human damage, etc.). For the license plate recognition system industry, accurate and reliable stability is very important.
The current license plate recognition technology mainly has four characteristics: first, the front-end embedded integration; second, the algorithm is basically based on the deep learning architecture; third, the front-end equipment integrates many functions that originally required supporting equipment in addition to the algorithm Fourth, most of them use HiSilicon chips.
Although the license plate recognition technology has made rapid progress in recent years, it has not yet reached the level of universalization. To make the license plate recognition integrated machine play the biggest role, the products of various application scenarios still use special license plate recognition cameras. The effect is better. For example, it is obviously not suitable for the highway to be placed in the parking lot, and the parking lot is also unsatisfactory in scenes such as gas stations and construction sites.
Nowadays, the application of license plate recognition technology has begun to move from the traffic field to the non-traffic field, such as 4S shops, auto repair shops, auto beauty shops, gas stations, weighbridges, charging piles, construction sites and other fields. The recognition characteristics of these complex scenes, The application functions that need to be integrated are very different from traffic scenarios, so many existing license plate recognition products suitable for dynamic traffic or static traffic have many pain points in these more subdivided and complex scenarios. It is against this background that license plate recognition technology focusing on non-traffic fields has gained more and more attention. At present, for the sub-scenarios such as smart car clothing, smart construction sites, smart gas stations, charging piles, weighbridges and other sub-scenarios, license plate recognition all-in-one machines dedicated to various complex scenarios have appeared. Explosion-proof license plate recognition all-in-one machine for gas stations, charging pile anti-occupancy license plate recognition all-in-one machine, etc.
It is foreseeable that in the future, more and more sub-scenarios will have to use scene-specific intelligent license plate recognition all-in-one machines, so that license plate recognition technology can bring more development directions to the industrial revolution, help enterprises achieve big data management, and complete the industry. Reform and progress.
In the late 1990s, the license plate recognition system was commercialized due to the needs of traffic management, but limited by the ability of the algorithm itself, the imaging level of the camera was low and the effect was not good, so only a small number of applications were obtained. The early license plate recognition technology has two main characteristics: one is based on traditional algorithms, and the other is based on soft algorithms on the PC side, and the application is limited by both.
After 2006, with the rise and maturity of machine learning, especially deep learning and front-end embedded algorithm deployment technology, as well as the substantial increase in chip computing power, the chip cost has been greatly reduced, and the mainstream form of products has changed from soft recognition to soft recognition. With the integrated camera for license plate recognition, the technology and application of license plate recognition have made great progress, especially in the past five years, license plate recognition has entered the era of comprehensive commercial use. From 2006 to 2014, license plate recognition was mainly used in dynamic traffic, including electronic police, highway checkpoints and highways. After 2014, it also ushered in a big explosion in static traffic, including parking lots, on-street parking and other scenarios. .
In 2018, the scene of the license plate recognition technology has been continuously subdivided, and complex scenes such as smart construction sites, smart gas stations, and unattended weighbridges have appeared. These scenes not only require higher performance of the license plate recognition machine, but also have a unique personality. Functional requirements, it is foreseeable that the license plate recognition integrated machine dedicated to the scene will become the development trend of license plate recognition technology in the future.
Almost every license plate recognition system claims to have a high recognition rate, but in order to avoid mutual shirking of the responsibility for poor operation due to differences in product knowledge between the two parties, users may wish to request a field test when purchasing a license plate recognition system, and the test The time is preferably more than two weeks, and it is better to judge whether the identification result is "exaggerated". Because of the changeable environment, I should be able to grasp about 80% of the situation that the field may affect the recognition rate in two weeks. If it is only measured for a day or even a few hours, it is impossible to understand.
In addition, since license plate recognition is a "system", the quality of the hardware and software architecture will of course affect the "presented results". As for what kind of software and hardware is suitable for what kind of environment, this must vary from environment to environment, because different application environments have different requirements for the recognition rate, and this must be accumulated through experience.
Although there are numerous license plate recognition systems on the market, using the right products and architecture can save a lot of wasted money and time, but more importantly, engineers and system integrators need cooperation and understanding from multiple parties, rather than blindly focusing on A certain brand is better and cheaper, so you won't lose anything by shopping around.
In addition, whether the license plate recognition system can play the greatest role, in addition to software technology, has a lot to do with cameras and on-site construction capabilities. The user can ask the manufacturer to conduct an on-site survey and propose a construction plan, first evaluate the location to be erected, the angle of the camera, whether an auxiliary light source needs to be erected, etc., and then submit a quotation. These actions can not only evaluate the ability of the operator in advance , the user can also achieve product learning and education training, and in the future management, will be more aware of the product's use restrictions and related countermeasures.
For example, a license plate recognition system has an accuracy of more than 90% during the day, but it drops to 80% in the evening, and then drops to 70% at night. This unstable system has an average accuracy of 70% compared to the all-weather license plate recognition system. More difficult to integrate. Because users will think that since the recognition rate during the day is 90%, it is reasonable to have an accuracy rate of 90% all-weather, and such specifications do not include strange environmental disturbances (rainstorm attacks, hail, dense fog sections, etc.) , and erection environmental restrictions (height restrictions, wind and sway restrictions, not easy to suffer from human damage, etc.). For the license plate recognition system industry, accurate and reliable stability is very important.
The current license plate recognition technology mainly has four characteristics: first, the front-end embedded integration; second, the algorithm is basically based on the deep learning architecture; third, the front-end equipment integrates many functions that originally required supporting equipment in addition to the algorithm Fourth, most of them use HiSilicon chips.
Although the license plate recognition technology has made rapid progress in recent years, it has not yet reached the level of universalization. To make the license plate recognition integrated machine play the biggest role, the products of various application scenarios still use special license plate recognition cameras. The effect is better. For example, it is obviously not suitable for the highway to be placed in the parking lot, and the parking lot is also unsatisfactory in scenes such as gas stations and construction sites.
Nowadays, the application of license plate recognition technology has begun to move from the traffic field to the non-traffic field, such as 4S shops, auto repair shops, auto beauty shops, gas stations, weighbridges, charging piles, construction sites and other fields. The recognition characteristics of these complex scenes, The application functions that need to be integrated are very different from traffic scenarios, so many existing license plate recognition products suitable for dynamic traffic or static traffic have many pain points in these more subdivided and complex scenarios. It is against this background that license plate recognition technology focusing on non-traffic fields has gained more and more attention. At present, for the sub-scenarios such as smart car clothing, smart construction sites, smart gas stations, charging piles, weighbridges and other sub-scenarios, license plate recognition all-in-one machines dedicated to various complex scenarios have appeared. Explosion-proof license plate recognition all-in-one machine for gas stations, charging pile anti-occupancy license plate recognition all-in-one machine, etc.
It is foreseeable that in the future, more and more sub-scenarios will have to use scene-specific intelligent license plate recognition all-in-one machines, so that license plate recognition technology can bring more development directions to the industrial revolution, help enterprises achieve big data management, and complete the industry. Reform and progress.
A comprehensive and in-depth analysis of current industry data on markets, competitors and suppliers.
In-depth research on key industry players and analysis of production and sales operations.
Objective forecasts of future trends in the industry.
WHAT QYRESEARCH OFFER?
Competition
Key players, new entrants,acquisitions, mergers and expansions,development trends and challenges.
Industry Analysis
Rawmaterial, application, product type, demand,supply,downstream, supply chain etc.
Market Size
Capacity, production, sales, revenue, price, cost etc.
Customized Information
We can offer customized survey and information to meet ourclient's need.