Personnel frequency counting optimization method

Abstract

The invention relates to an optimization method for counting the frequency of people, and the method comprises the following steps: 1, carrying out the recognition of people in a face library through a face recognition technology, and carrying out the frequency counting; carrying out automatic warehousing recognition on people not in the face database, and carrying out frequency statistics; step 2, matching and counting are carried out on the personnel which are subjected to frequency statistics according to hours through cache data; matching and counting are carried out on the personnel which are subjected to frequency statistics according to days in a cache and database matching mode; and step 3, carrying out post video and picture query operation on alarm personnel. According to the method, strangers can be put in storage by controlling the picture quality, so that the number of persons who fail in modeling is greatly reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of face recognition and relates to a method for counting personnel frequencies, in particular to an optimization method for counting personnel frequencies. Background Art

[0002] With the advancement of science and technology, embedded systems have been widely used in various fields and have become an important force in promoting the development of intelligence. With its small size, low power consumption, high integration and strong reliability, embedded systems play an important role in many fields such as industrial automation, smart home, medical equipment, and security monitoring. In terms of personnel frequency statistics, embedded systems can provide real-time and accurate data processing capabilities to meet the needs of rapid acquisition and analysis of personnel flow information.

[0003] With the acceleration of urbanization and the increasing importance of public safety, video surveillance systems have been widely used in various cities and public places. However, most traditional video surveillance systems rely on manual monitoring, which has problems such as low monitoring efficiency and easy false alarms. Therefore, the demand for intelligent analysis of video surveillance is growing. The embedded personnel frequency counting system was developed based on this demand. It can automatically analyze and process surveillance videos, count the flow frequency of personnel in real time, and provide strong data support for urban management and public safety.

[0004] However, the existing personnel frequency statistics methods have the following defects:

[0005] 1. People are directly entered into the database without being identified, resulting in a large number of unfamiliar people being unable to be successfully modeled.

[0006] 2. Only daily statistics are supported, not short-term statistics.

[0007] 3. Storing data directly in flash will cause problems with the life of the flash.

[0008] 4. The entire device only supports one frequency statistics result and cannot meet the statistics of multiple sets of frequency data.

[0009] Therefore, the present invention proposes an optimization method for counting personnel frequency, which can solve the above problems.

[0010] After searching, no public documents of the prior art identical or similar to the present invention were found. Summary of the invention

[0011] The purpose of the present invention is to overcome the shortcomings of the prior art and propose an optimization method for statistical personnel frequency, which can control the quality of pictures to enter strangers into the database, greatly reducing the number of people who fail to model.

[0012] The present invention solves the practical problem by adopting the following technical solutions:

[0013] An optimization method for counting personnel frequency includes the following steps:

[0014] Step 1: Use face recognition technology to identify people in the face database and perform frequency statistics; for people not in the face database, automatically enter the database for identification and perform frequency statistics;

[0015] Step 2: For personnel whose frequency is counted by hour, matching and counting are performed through cache data; for personnel whose frequency is counted by day, matching and counting are performed through the combination of cache and database;

[0016] Step 3: For the person who called the police, perform post-event video and image query operations.

[0017] Moreover, the specific steps of step 1 include:

[0018] (1) The video frames sent by the camera are extracted at a certain ratio, and the video frame content is sent to the algorithm model for face detection. If a face is detected according to industry standards, the face image is captured and matched with the existing people in the face database;

[0019] (2) If the similarity with the existing persons in the face database is greater than or equal to the threshold configured by the user, the number of times the person appears is output to the frequency statistics module for statistics; if the similarity with the existing persons in the face database is less than the threshold configured by the user, the face whose face size, face angle, and face clarity meet the requirements for entry into the database is entered into the stranger database, and the number of times the person appears is sent to the frequency statistics module.

[0020] Moreover, the specific steps of step 2 include:

[0021] (1) Analyze the data entering the frequency statistics module and divide it into personnel with recorded frequency and personnel without recorded frequency;

[0022] (2) For personnel data with recorded frequency, directly update the data to the existing cache, and for personnel data with unrecorded frequency, create a new cache space to record it;

[0023] (3) According to the frequency statistics method, when the alarm threshold is reached, it is sent to the alarm module for alarm;

[0024] Moreover, the specific method of step 2 (3) is:

[0025] For personnel who need to be counted by hour, after updating to the cache, the cache content is judged by the threshold, and the personnel who reach the alarm threshold are reported; for personnel who need to be counted by day, after updating to the cache, the cache memory count of each person and the number of times each person appears in the database are merged and integrated. If the number of times a single person appears meets the threshold condition preset by the user, the personnel who reach the alarm threshold are reported.

[0026] Advantages and beneficial effects of the present invention:

[0027] The present invention provides an optimization method for counting personnel frequency, which controls the image quality to enter strangers into the database, greatly reducing the number of personnel whose modeling fails. The present invention supports frequency statistics by day or hour, which can meet the needs of users in different scenarios. The present invention reduces the number of flash writes and increases the service life of the flash by caching. The present invention supports 16 groups of frequency statistics at the same time, and can output 16 groups of frequency statistics results at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of personnel identification and automatic warehousing process of the present invention;

[0029] Figure 2 It is a personnel frequency statistics flow chart of the present invention. DETAILED DESCRIPTION

[0030] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0031] An optimization method for counting personnel frequency, such as Figure 1 and Figure 2 As shown, the following steps are included:

[0032] Step 1: Use face recognition technology to identify people in the face database and perform frequency statistics; for people not in the face database, automatically enter the database for identification and perform frequency statistics;

[0033] The specific steps of step 1 include:

[0034] (1) The video frames sent by the camera are extracted at a certain ratio, and the video frame content is sent to the algorithm model for face detection. If a face is detected according to industry standards, the face image is captured and matched with the existing people in the face database;

[0035] (2) If the similarity with the existing persons in the face database is greater than or equal to the threshold configured by the user, the number of times the person appears is output to the frequency statistics module for statistics; if the similarity with the existing persons in the face database is less than the threshold configured by the user, the face whose face size, face angle, and face clarity meet the requirements for entry into the database is entered into the stranger database, and the number of times the person appears is sent to the frequency statistics module.

[0036] Step 2: For personnel whose frequency is counted by hour, matching and counting are performed through cache data; for personnel whose frequency is counted by day, matching and counting are performed through the combination of cache and database;

[0037] The specific steps of step 2 include:

[0038] (1) Analyze the data entering the frequency statistics module and divide it into personnel with recorded frequency and personnel without recorded frequency;

[0039] (2) For personnel data with recorded frequency, directly update the data to the existing cache, and for personnel data with unrecorded frequency, create a new cache space to record it;

[0040] (3) According to the frequency statistics method, when the alarm threshold is reached, it is sent to the alarm module for alarm;

[0041] The specific method of step (3) of step 2 is:

[0042] For personnel who need to be counted hourly, after updating to the cache, the cache content is judged by threshold, and the personnel who reach the alarm threshold are reported;

[0043] The data that needs to be counted on a daily basis is updated to the cache, and the cache memory count of each person and the number of times each person appears in the database are merged and integrated. If the number of times a single person appears meets the threshold conditions preset by the user, an alarm will be reported for the person who reaches the alarm threshold.

[0044] Step 3: For the person who called the police, perform post-event video and image query operations.

[0045] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation manner. Any other implementation manners derived by those skilled in the art based on the technical solution of the present invention also fall within the scope of protection of the present invention.

Claims

1. A method for optimizing the frequency of statistical personnel, characterized in that: The following steps are involved: Step 1: Use face recognition technology to identify people in the face database and perform frequency statistics; For people who are not in the face database, automatic database entry and identification are adopted, and frequency statistics are performed; Step 2: For personnel whose frequency is counted by hour, matching and counting are performed through cache data; for personnel whose frequency is counted by day, matching and counting are performed through the combination of cache and database; Step 3: For the person who called the police, perform post-event video and image query operations.

2. The optimization method for counting personnel frequency according to claim 1, characterized in that: The specific steps of step 1 include: (1) The video frames sent by the camera are extracted at a certain ratio, and the video frame content is sent to the algorithm model for face detection. If a face is detected according to industry standards, the face image is captured and matched with the existing people in the face database; (2) If the similarity with the existing persons in the face database is greater than or equal to the threshold configured by the user, the number of times the person appears is output to the frequency statistics module for statistics; if the similarity with the existing persons in the face database is less than the threshold configured by the user, the face whose face size, face angle, and face clarity meet the requirements for entry into the database is entered into the stranger database, and the number of times the person appears is sent to the frequency statistics module.

3. The optimization method for counting personnel frequency according to claim 1, characterized in that: The specific steps of step 2 include: (1) Analyze the data entering the frequency statistics module and divide it into personnel with recorded frequency and personnel without recorded frequency; (2) For personnel data with recorded frequency, directly update the data to the existing cache, and for personnel data with unrecorded frequency, create a new cache space to record it; (3) According to the frequency statistics method, when the alarm threshold is reached, it is sent to the alarm module for alarm.

4. The method for optimizing the frequency of personnel statistics according to claim 3, characterized in that: The specific method of step (3) of step 2 is: For personnel who need to be counted by hour, after updating to the cache, the cache content is judged by the threshold, and the personnel who reach the alarm threshold are reported; for personnel who need to be counted by day, after updating to the cache, the cache memory count of each person and the number of times each person appears in the database are merged and integrated. If the number of times a single person appears meets the threshold condition preset by the user, the personnel who reach the alarm threshold are reported.

Images