Collision analysis method and system based on vector space cosine similarity and storage medium
A technology of cosine similarity and collision analysis, applied in the field of image processing, can solve the problems of heavy workload, missing the best time for handling cases, low efficiency, etc., to improve accuracy, improve query efficiency and query accuracy, and improve flexibility Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0042] Such as figure 1 As shown, the present invention provides an embodiment of a collision analysis method based on vector space cosine similarity, comprising the following steps:
[0043] S1, receiving image data to be processed; the image data includes image compression package data and video data;
[0044] S2, processing the image data to obtain picture data; the picture data includes several pictures;
[0045] S3, call the GPU to carry out the recognition operation on the image data, and obtain the Base64 encoding value information of each image;
[0046] S4, based on the vector space cosine similarity, perform collision analysis on the Base64 encoded value information, and obtain a collision analysis result.
Embodiment 2
[0048] The present invention provides an embodiment of a collision analysis method based on vector space cosine similarity, comprising the following steps:
[0049] Step 1, create a new collision analysis task, including its basic task information, such as task name, task collision callback address, operation rules of task collision, etc. In the embodiment of the present invention, there are two types of operation rules, one is reference collision, and the other is many-to-many multipoint collision. Receive two files uploaded by the user. In the embodiment of the present invention, the first file is image compression package data, and the second file is video data. In other embodiments, the two files can both be image compression package data, or both can be video data;
[0050] Step 2, judging the formats of the two files, and decompressing the data in the image compression package to obtain the first image data. Store the image address of the first image data in the server...
Embodiment 3
[0058] The present invention also provides a collision analysis system based on vector space cosine similarity, including:
[0059] A data receiving module, configured to receive image data to be processed;
[0060] The data processing module is used to process the image data to obtain picture data; the picture data includes several pictures;
[0061] The identification calculation module is used to call the GPU to perform identification calculations on the picture data, obtain the Base64 encoded value information of each picture and store it in the Redis hash;
[0062] The collision analysis module is used to perform collision analysis on the Base64 encoded value information in the Redis hash based on the vector space cosine similarity, and obtain the collision analysis result;
[0063] Clickhouse database module for storing collision analysis results;
[0064] and a retrieval module, which is used to obtain retrieval parameters and send them to the collision analysis modul...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More