Battery cars are forbidden to enter the elevator. What is the best solution for you?

Recently, battery cars have frequently entered the elevator, causing indoor fires and causing casualties and property losses. A video of a battery car spontaneously igniting in an elevator has been widely spread on the Internet, which has aroused great social repercussions. When the battery car enters the elevator, it may not only damage the elevator, but also have great security risks when charging the battery car at home. This is a social problem that needs to be solved urgently.

To solve this problem, the core requirement is to accurately identify the scene of "electric vehicles entering the elevator", and the feasible technical route should have two directions: back-end identification and front-end identification:

1) The back-end identification direction is to put the identified actions behind the management of the monitoring room and computer room, and it is necessary to obtain the video stream in the monitoring background based on the existing video monitoring system, and then realize the target scene identification by using the AI algorithm;

2) Front-end direction recognition means that the recognized action is placed in front of the elevator, and a video recognition device or a sensing monitoring device needs to be installed in the elevator, so that the electric vehicle can be recognized immediately as soon as it enters the elevator, and the information can be processed in collaboration on the spot and pushed to the rear management platform in real time.

The schemes in the two directions have their own advantages, mainly depending on the actual control requirements and business scenarios. Among them, in front-end recognition, I personally recommend a more mature and reliable video recognition technology, because there is only the light curtain +AI algorithm scheme of sensor monitoring equipment at present, but there is still a gap between this scheme and the video recognition scheme of elevator scene in accuracy, reliability and installation convenience. Therefore, for the following scenarios, in the choice of front-end recognition direction, video recognition equipment is recommended by default.

Demand Scenario 1: Your elevator is newly installed, and video surveillance has not been installed.

In this case, choose the front-end identification direction, and suggest that you directly choose the camera that supports electric vehicle identification, and take this part of the demand as part of the monitoring system. Acousto-optic early warning, linkage ladder control and alarm push should be available, and infrastructure such as centralized parking and charging can also be done together.

Demand Scenario 2: The elevator has a monitoring system, and you don't mind installing an extra camera to accomplish this. The system needs to be placed in the local area network.

In this case, in order to choose the front-end identification direction, it is necessary to install wired video identification equipment in the elevator and link the elevator control as needed. Then the back end is equipped with a small server to deploy the management background to receive and process the alarm, and the short message is pushed by a short message cat. This can not only realize front-end identification and early warning, but also realize back-end alarm.

Demand Scenario 3: The elevator already has a monitoring system, so I don't mind installing an extra camera, but I don't want to be so troublesome with wired networking.

In this case, to select the front-end identification direction, it is necessary to install wireless video identification equipment in the elevator, and link the elevator control as needed, and the alarm information will be pushed to the cloud, which can be managed remotely based on applet and web.

Demand scenario 4: I don't want to install anything on the elevator, and I don't want to do on-site early warning and ladder control.

In this case, you have to choose the back-end identification direction. You need to set up a server at the back end, access the existing video stream, analyze the scene with AI algorithm, push the results to the management background, and then push the alarm by SMS.

The picture comes from the network (deleted)

The above scenarios can basically cover most of the situations we usually encounter, which are technically verified and feasible. Of course, they can only be regarded as solutions. The implementation of specific projects still needs to be selected and refined according to their respective landing needs.

The above content is my thinking process in the process of researching solutions for customers, in order to reduce the research time of friends in need. If there are any omissions or mistakes, you can leave me a message, or we can discuss the actual solution requirements together and make progress together!

Ps: If you want a simplified solution, please write down your email address in the message and send it to you when you see it.