Abnormal recognition result processing method applied to smart city box body detection
A technology for abnormal recognition and result processing, applied in neural learning methods, character and pattern recognition, instruments, etc., can solve the problem of not being able to find abnormal door opening pictures or videos
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0033] Such as figure 1 As shown, the abnormal recognition result processing method applied to smart city cabinet detection in this embodiment includes the following steps:
[0034] S1. Acquire video data in real time, process the video data into frames, and generate several pictures.
[0035] In this embodiment, video data in .264 format is acquired and decoded by FFMpeg. Use the segmentation component in OpenCV to divide the decoded video data into frames.
[0036] S2. Preprocessing the picture, inputting the preprocessed picture into the neural network model for judgment, and obtaining the judgment result, the position information of the box door and the position information of the box body. In this embodiment, during preprocessing, the image is converted into a grayscale image, the grayscale image is decomposed into eight-bit planes, and one of the one-bit plane images is extracted and input into the neural network model for judgment.
[0037] S3. When the judgment resu...
Embodiment 2
[0050] The difference between this embodiment and Embodiment 1 is that in S1 of this embodiment, the face recognition module is also used to determine whether there is a person in the monitoring area; when there is a person in the monitoring area, face recognition is performed to determine whether the person is on record, If it has not been filed, mark the corresponding video data as high-priority video data, and if it has been filed, mark the corresponding video data as low-priority video data. During frame processing, the high-priority video data corresponding to the camera in the preset area is prioritized for frame processing.
[0051] When there is no person in the monitoring area, the possibility of abnormal opening of the box door is small. Every preset time, the pictures are extracted and input into the neural network model for judgment, instead of judging the pictures in real time, which can reduce the neural network. Network models deal with stress. By collecting the ...
Embodiment 3
[0053] The difference between this embodiment and Embodiment 1 is that in this embodiment, S1 also includes:
[0054] S101. Slicing step: Slicing the video data to generate a description file and several media segments; the description file is used to record the shooting date, total duration, number and duration of each media segment of the video data. In this embodiment, video data is sliced in units of milliseconds. In this embodiment, the description file is an m3u8 file, and the media segment is a ts file. For example, the total duration of video data is 10 seconds, sliced into 10 ts files, the duration of a single ts file is 1 second, and the number of ts files is from 001 to 010. The shooting date is, for example, 2021-7-15-12:01:00:001.
[0055] S102. Recognition step: perform binarization processing on the media segment, and determine whether the binarized media segment contains a preset identifier, if it contains a preset identifier, go to the correction step, a...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More