Outdoor advertisement online collapse prediction method based on LSTM neural network

Abstract

The invention discloses an outdoor advertisement online collapse prediction method based on an LSTM neural network. The method comprises the following steps: S1, collecting environmental data and vibration data; S2, performing normalization processing on the acquired data; S3, performing training by utilizing data under normal conditions at the initial stage of installation, comparing the environment data, the mapping outdoor advertisement environment and the nonlinear relation of the vibration data with actual data through an LSTM neural network, comparing the fitting result with the actual data, learning the vibration mode of the outdoor advertisement under the action of wind power, and combining the LSTM neural network used in the step with an LSTM layer and a full connection layer, wherein the LSTM layer is used for extracting sequence data features, and the full connection layer fuses the sequence features and changes the shape of output data; and S4, carrying out long-term onlinedetection on the outdoor advertisement based on the training result of S3, and judging the abnormity of the outdoor advertisement. All-weather online anomaly detection can be realized.

Description

technical field [0001] The invention relates to the application of computer technology and artificial intelligence in the field of fault detection, in particular to an online collapse prediction method for outdoor advertisements based on LSTM neural network. Background technique [0002] Outdoor advertisements are neon lights, outdoor advertisements, posters, etc. set up on the exterior of buildings or in outdoor public places such as streets and squares. Outdoor advertising is aimed at all the public, so it is difficult to choose specific target objects, but outdoor advertising can display the image and brand of the company in a fixed place for a long time, so it is very effective for improving the popularity of the company and the brand. [0003] Since the location of outdoor advertisements has a certain height, if there are problems in the installation or use of outdoor advertisements, such as screw corrosion, broken brackets and other factors, the outdoor advertisements ...

AI Technical Summary

Problems solved by technology

[0004] The existing outdoor advertisement collapse prediction technology usually collects sensor data and uses a certain algorithm to predict the state of outdoor advertisements. This method cannot adapt to outdoor advertisements in different environments, and has poor generalization ability and low accuracy. Prone to false positives or non-reports of abnormal states

Another type of technology uses machine vision methods to detect the status of outdoor advertisements. This type of method can only detect visual abnormalities, and usually cannot accurately predict the collapse of outdoor advertisements. Moreover, this method is subject to relatively large interference noise in the environment and is prone to occurrence. false positive

Images