id:BSN_2021
Public account: BSN Study Club
Garbage classification refers to the general term for a series of activities that classify, store, place and transport garbage according to certain regulations or standards, thereby transforming it into public resources. The purpose of garbage classification is to increase the resource value and economic value of garbage, reduce the amount of garbage processed and the use of processing equipment, reduce processing costs, and reduce the consumption of land resources. It has social, economic, ecological and other benefits.
In recent years, with the continuous growth of our country's economy and the continuous improvement of people's economic and living standards, our country has produced a large amount of domestic and production waste every day. Inappropriate garbage disposal methods not only damage the environment and cause a lot of waste of resources, but also bring many thorny problems to local governments and competent authorities. In this regard, relevant departments in my country have issued documents many times, clearly requiring "implementing the responsibilities of urban entities, promoting the development of people's habits, and accelerating the construction of classified facilities..."
On May 19, 2022, the BSN Study Club invited Shen Fengping, a BSN qualified developer and CEO of Ningbo Huanlian Big Data Co., Ltd. Shen Fengping used the theme of "Blockchain Application of 5G + Artificial Intelligence Technology in Garbage Classification for Inclusive Carbon Benefits of Residents", focusing on the problems encountered in garbage classification at the current stage, how to use artificial intelligence and blockchain technology to improve quality, "Ai Xiao "Kitchen" smart waste classification system, implementation application case introduction, etc., brought a wonderful sharing to developers and industry users.
1. Current difficulties in garbage classification
In the current garbage classification work, there are three main difficulties: weak classification awareness, unsatisfactory supervision effects, and difficult government management.
Weak classification awareness
Citizens’ classification knowledge is not sufficiently popularized, and classification habits are difficult to change.
Supervision effect is not ideal
Manual non-round-the-clock supervision. Supervisors are inefficient, costly and ineffective. Some supervisors have become sorters, which is not conducive to the improvement of residents' ability to sort independently.
Government management is difficult
Unable to detect illegal placement in time and obtain evidence for wrong behavior. It is impossible to provide targeted and classified counseling to residents. Smart device manufacturers operate independently, and the government does not have a unified ledger and supervision system.
2.Solution
In response to these pain points, Ningbo Huanlian Big Data Co., Ltd. launched Ai Xiaochu - a digital garbage classification solution driven by "artificial intelligence".
Solution - solution logic
Internet of Things (IOT) technology and Artificial Intelligence (AI) technology are used to build an all-weather verification and management system. Traditional garbage collection points are paired with "Ai Xiaochu" to conduct real-time verification, and feedback the verification results to residents to correct erroneous garbage collection. Carry out supervision, continue to improve the public's awareness and ability of garbage classification, and help the government establish a long-term and effective supervision mechanism for garbage classification operations.
In the past, the government has promoted green accounts and green points for many years in order to encourage residents to protect the environment, but it has never achieved the desired results. If incentives are based on the weight of garbage put in, then the more garbage produced, the more incentives will be given, which is not in line with the concept of environmental protection. If incentives are based on the amount of garbage put out, people who do not put out garbage will enjoy large rewards, which is not in line with the government's original intention of encouraging citizens to actively participate in environmental protection.
Reliance on Ai Xiaochu provides corresponding "carbon conversion" incentives based on the amount, frequency, and accuracy of garbage put in by citizens. Rely on this reward method to guide some people to participate in garbage classification. Then through these seed customers, we can further drive more people in the community to participate. It ultimately enhances the environmental awareness of every citizen.
In the past, enterprise data started from the upstream, and the downstream referenced the upstream data. This model is no longer feasible: once a link is falsified, the error will continue. Each node of Ai Xiaochu collects data through trusted sensors; each node is independent, and the data is collected and transmitted through trusted transmission on the chain. All data can be traced back to the original state; finally, relying on carbon accounting, calculation Show the value of interdependent behavior.
Shen Fengping said, "At first we built our own blockchain platform, and later came into contact with BSN. We gradually migrated many applications to BSN. On the one hand, it significantly reduced the company's data storage costs, and also reduced the operation and maintenance costs. , and it’s safe and reliable.”
business function architecture
Solution - full scene supervision instructions for the community
Classified assessment + directional guidance + low-carbon incentive business process
Solution - Instructions for digital supervision of delivery sites
Solution - front-end module
Solution—data aggregation
Core Technology
Ai Xiaochu - a digital garbage classification solution driven by "artificial intelligence" has four core technologies, including MOPS hierarchical identification technology, Ai Xiaochu smart eye, full-scenario automated training technology, and financial-level data transactions.
MOPS layered recognition technology. Real-time capture of the delivery status of each point, the identity of the deliverer, the environment inside the trash can, the height of the garbage in the box, the delivery weight... The self-developed MOPS superimposed image layered recognition technology realizes garbage classification and assessment at the home.
Ai Xiaochu has a wise eye. The continuously evolving evaluation engine uses real-life community data as training samples. It has good food waste identification capabilities (can identify more than 100,000 types of domestic waste) and continues to evolve.
Full-scenario automated training technology. The form of garbage may change. In order to quickly improve the fault tolerance of the engine, Ningbo Huanlian Cloud has developed panoramic automated training technology so that waste has nowhere to hide.
Financial grade data trading. Based on blockchain technology, multi-node on-chain storage and full data desensitization are on-chain. Currently, nodes in Ningbo, Beijing, Shenyang, Xinyang and other cities have been opened. Financial-level authentication transmission can be sourced at any time without worrying about data leakage and loss.
Through layered photography, edge computing, neural networks, 5G technology and artificial intelligence analysis, each delivery system for residents will record and analyze the accuracy of the delivery. The system backend will provide real-time feedback to residents and supervisors based on the results, while retaining the data for next-step decisions.
Digital Ai Xiaochu
Practice has shown that in most cases, the garbage classification and recognition capabilities of Ai Xiaochu AI engine have caught up with normal people; in a small number of cases, it is still slightly inferior to humans, with a gap of about 10%. In the case of 20,000 samples, the recognition ability of the AI engine is close to that of humans because its working state can remain stable, while as the workload increases, the human state will decline.
Solution – Backend Platform
Solution – Targeted Counseling and Decision Making
Ai Xiaochu's targeted counseling decision-making is divided into three steps. First, the mixed investment situation of residents is analyzed based on big data, a portrait of the residents is made, and a targeted counseling work order is automatically generated. Secondly, the automatically generated calendar delivery record allows customers to see at a glance the accuracy of their daily delivery behavior. Finally, mobile phone text messages and WeChat real-time notifications.
Solution - problem solved
Ai Xiaochu’s solution mainly solves three problems. The first is machine substitution. Machines replace manual supervision, provide 24/7 supervision and guidance, improve the pass rate of residents' garbage disposal, and promote the development of people's habits. At the same time, government spending will be significantly reduced. The second is data empowerment. With the help of 5G, big data analysis, and artificial intelligence technology, automatic analysis of evidence collection will accelerate the implementation of environmental protection policies based on the rule of law. The last step is to promote reform. Artificial intelligence technology is used to accurately calculate and guide the weight and separation rate of various types of domestic waste for residents, laying the foundation for domestic waste charging and helping to achieve peak carbon neutrality.
3. Practical demonstration
Ai Xiaochu has been established in many communities. Before installation, there were many problems with garbage classification in the community. Specifically, the classification quality is poor and the pass rate is low; the supervisors are too old and have low cultural quality, and they only play the role of garbage sorters; they are unable to monitor the specific situation of garbage classification, and they cannot play a role in some residents who maliciously litter. Despite warnings and education, there is no evidence of punishment, and conflicts between instructors and residents are prominent.
After installation, the effect is obvious. Specifically, the site of the community garbage collection station is kept clean and tidy, and malicious mixed garbage disappears; a detection module is installed in the garbage collection port to detect garbage classification around the clock; a smart speaker is set up to remind residents of broken bags for delivery, and timely voice reminders are issued for incorrect garbage collection; reporting Violation information is sent to the law enforcement system to provide regulatory authorities with a complete chain of law enforcement evidence.
Practical display——Anji Wanhua Mansion
4. Operational capabilities
Product R&D unit
Localization project support
