A process judgment method based on camera single-person multi-site visual recognition

Abstract

The application discloses a process judgment method based on single-person multi-site visual identification of a camera, belongs to the technical field of visual identification, and solves the problem of visual identification process data processing for different sites in the hospital infection process.The technical solution points are that multiple cameras are arranged in multiple sites, only one person in the site corresponding to each camera carries out a business and collects the business action, and the collected data is reported to an AI machine visual identification module and processed through the same action judgment module, so that temporary process actions are concentratedly processed, the process judgment processing of a timing task is carried out, data is saved, and the process is ended, the method is used to determine the process to which the action belongs and the nature and name of the process, and thus the business ability related to the machine vision result is given.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of visual recognition, in particular, to a process judgment method based on visual recognition of a single person in multiple sites by a camera. BACKGROUND

[0002] Hand hygiene (hand washing, hand disinfection, surgical hand disinfection, etc.), multiple drug-resistant bacteria isolation, and disinfection operation of surgical instruments and endoscopes require behavior monitoring and management. Since the above management requires isolation operation, it is very suitable for monitoring and management using machine vision.

[0003] However, using machine vision to manage hospital infection business, the business based on hospital infection management is generally nursing business, and the nursing business of a hospital is various, but it is necessary to monitor multiple actions of the same business. We call the multiple actions of the same business a process. The accuracy of the action, the duration of the action, whether the action is omitted, whether there is a rule violation action, whether the sequence of the action is correct, and the like.

[0004] The actions reported by the AI machine vision recognition may be the same action, different actions in different places, or actions of different people. The system needs to identify and judge these actions and classify them into the same process. Then, the subsequent business calculation can be carried out.

[0005] Among them, there is a scene: including one or more different sites, a site has only one camera, and only one person works under the scene or the camera at the same time. A complete working process of the person will be completed in multiple sites, and one or more actions may be performed in each site.

[0006] To solve the above-mentioned scene problem, a process judgment method based on visual recognition of a single person in multiple sites by a camera is needed. SUMMARY

[0007] The present application aims to solve the technical problems in the related art to at least some extent, and provides a process judgment method based on visual recognition of a single person in multiple sites by a camera, which is used to determine the process to which the action belongs, and the nature and name of the process, thereby giving the machine vision result related business capability.

[0008] In order to solve the above technical problems, the technical scheme of the present application is: a process judgment method based on camera single-person multi-site visual recognition, multiple cameras are arranged in multiple sites, only one person in the site corresponding to each camera carries out business and collects business actions, and reports to an AI machine vision recognition module, and then is processed through the same action judgment module, temporary process actions are concentratedly processed, a timing task temporary process is judged and processed, data is saved and ends.

[0009] Preferably, the same action judgment module processing step comprises: sequentially judging a first condition that the existence of a temporary process is queried with a camera ID as a key, a second condition that the current action and the queried temporary process belong to the same process, and a third condition that the queried temporary process property and name exist;

[0010] When the first condition is met, the second condition is not met, and the third condition is not met, the queried temporary process is deleted, a new temporary process is generated, the current action is added to the new temporary process, and the end,

[0011] When the first condition is not met, a new temporary process is generated, the current action is added to the new temporary process, and the end,

[0012] When the first condition is not met and the second condition is met, the current action is added to the queried temporary process, the ID of the queried temporary process is set to the current action, and the end,

[0013] When the first condition is met, the second condition is not met, and the third condition is met, a new temporary process is generated, the current action is added to the new temporary process, and the end.

[0014] Preferably, the temporary process data reporting step of the timing task temporary process judgment processing comprises:

[0015] All temporary processes and actions are taken out from the memory and traversed, and all actions of the current temporary process are taken out from the memory,

[0016] Whether the fourth condition that the time comparison of the last action of the current traversed temporary process and the first action of the current temporary process is whether more than a predetermined period of time and the fifth condition that the related actions of the two temporary processes can be connected is judged,

[0017] When the first condition is met, the end is directly ended;

[0018] When the first condition is not met and the second condition is not met, the end is directly ended;

[0019] When the first condition is not met and the second condition is met, the temporary processes are merged, the ID of the current temporary process is used as the process ID of all actions, the merged process data is saved, and the end.

[0020] Preferably, the timing task is triggered by a timer or a timing interrupt control.

[0021] Preferably, the set of actions or processes is stored in a volatile memory or in a non-volatile memory.

[0022] Compared with the background art, the technical effects of the present application mainly lie in the following aspects:

[0023] 1. Complete the process of visual recognition of a single camera in multiple locations to determine the nature and name of the process to which the action belongs, thereby giving the machine vision results related business capabilities;

[0024] 2. Improve data processing efficiency and reduce hardware data processing card lag. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 is a general flowchart schematic diagram of the process judgment method for visual recognition of a single camera in multiple locations according to the present application;

[0026] Figure 2 is a flowchart schematic diagram of the temporary process action set algorithm, which concentrates multiple actions into a temporary process;

[0027] Figure 3 is a flowchart schematic diagram of the temporary process to process algorithm, which combines multiple temporary processes into a final process and judges the nature and name of the process, and finally saves the flowchart. DETAILED DESCRIPTION

[0028] The specific embodiments of the present application are further described in detail below with reference to the accompanying drawings, so that the technical scheme of the present application is easier to understand and master.

[0029] As for the program, algorithm, and method, it should be interpreted as the flow steps of data processing in the art. The above-mentioned program is carried on the hardware module, for example: the same action judgment program is carried on the same action judgment module, and in one or more hardware modules, the corresponding data processing flow steps can be completed and implemented.

[0030] As for the action, it refers to the business action, and as for the process, it refers to the business action. The business is diversified, for example: each specified action of hand hygiene (hand washing, hand disinfection, and surgical hand disinfection) in the hospital process, which constitutes a qualified process from these continuous compliance actions. Similar processes can be classified and processed in a data set. Temporary process: the process of a business currently captured by the camera.

[0031] Traversal: Traversal refers to the process of following a search path through a tree (or graph) and visiting each node in the tree (or graph) in turn.

[0032] Embodiment:

[0033] The method of the present solution comprises a same action judgment program, a temporary process action set algorithm, a timing task temporary process judgment program, a temporary process to process algorithm, and a data saving program. An example is as follows Figure 1 .

[0034] The temporary process action set algorithm is used to determine whether the action belongs to a certain process according to the attribution, and to collect all actions of a process into a set for further processing. An example is as follows Figure 2 .

[0035] The all actions of the same process algorithm comprises the following steps:

[0036] First, the camera ID is used as a keyword to query in the temporary process set to determine whether a temporary process with the camera ID as a keyword exists.

[0037] Further, if it does not exist, a new temporary process is generated, and the current action is added to the new temporary process.

[0038] Further, if it exists, it is determined whether the temporary process in which the current action is located and the queried temporary process belong to the same temporary process.

[0039] Further, if they are the same temporary process, the ID of the queried temporary process is set to the current action, and the current action is added to the queried temporary process.

[0040] Further, if they are not the same temporary process, it is determined whether the queried temporary process property and name exist.

[0041] Further, if they exist, a new temporary process is generated, and the current action is added to the new temporary process.

[0042] Further, if they do not exist, the queried temporary process is deleted.

[0043] Further, a new temporary process is generated, and the current action is added to the new temporary process.

[0044] The temporary process to process algorithm is used to combine multiple temporary processes into a final process, determine the property and name of the process, and finally save it to the memory.

[0045] An example is as follows Figure 3 .

[0046] A method for converting temporary processes into processes, comprising the following steps:

[0047] First, all actions of the current temporary process are fetched from the memory.

[0048] Further, all temporary processes and all related actions are fetched from the memory, and all temporary processes are traversed.

[0049] Further, a time comparison is made between the last action of the current traversed temporary process and the first action of the current temporary process, to determine whether a certain period of time has been exceeded.

[0050] Further, if the period of time has been exceeded, the current loop is ended.

[0051] Further, if the period of time has not been exceeded, a comparison is made to determine whether the related actions of the two temporary processes can be connected.

[0052] Further, if the actions cannot be connected, the process is ended.

[0053] Further, if the actions can be connected, the temporary processes are merged, and the ID of the current temporary process is used as the process ID of all actions.

[0054] Further, the merged process data is saved, and the process is ended.

[0055] The actions or process sets can be saved in volatile memory (RAM) or non-volatile memory (NVM), such as PROM.

[0056] Of course, the above is only a typical example of the present application, and the present application can have other various specific embodiments, and any technical solution formed by equivalent replacement or equivalent transformation falls within the scope of the present application.

Claims

1. A process judgment method for single-person, multi-location visual recognition based on cameras, characterized by: Multiple cameras are set up in multiple locations. Only one person performs business and collects business actions in the location corresponding to each camera, and reports them to the AI ​​machine vision recognition module. Then, the same action judgment module processes the temporary process actions centrally, including: The system sequentially checks the following conditions: First, the existence of a temporary process is queried using the camera ID as the key; second, the current action and the queried temporary process belong to the same process; and third, the nature and name of the queried temporary process exist. If the first condition is met, but the second and third conditions are not met, delete the queried temporary process, generate a new temporary process, add the current action to the new temporary process, and then terminate. If the first condition is not met, a new temporary process is generated, the current action is added to the new temporary process, and the process ends. If both the first and second conditions are met, the current action is added to the queried temporary process, the ID of the queried temporary process is set to the current action, and the process ends. If the first condition is met, the second condition is not met, and the third condition is met, then a new temporary process is generated, the current action is added to the new temporary process, and the process ends. The scheduled task performs a temporary process to check and process the data, saves the data, and then terminates.

2. The process judgment method for single-person, multi-location visual recognition based on a camera as described in claim 1, characterized in that: The steps for judging and processing the reported temporary process data in the scheduled task temporary process include: Retrieve all temporary procedures and actions from memory, and iterate through them. Retrieve all actions of the current temporary procedure from memory. To determine if the fourth condition is met: compare the time of the last action of the current temporary process with the time of the first action of the current temporary process to see if it exceeds the predetermined time limit; the fifth condition: whether the related actions of the two temporary processes can be connected. When the fourth condition is met, the process ends directly. If neither the fourth nor the fifth condition is met, the process ends directly. If the fourth condition is not met, but the fifth condition is met, merge the temporary processes, use the ID of the current temporary process as the ID of all action processes, save the merged process data, and end the process.

3. The process judgment method for single-person, multi-location visual recognition based on a camera as described in claim 1, characterized in that: Scheduled tasks are triggered by timers or timer interrupts.

4. The process judgment method for single-person, multi-location visual recognition based on a camera as described in claim 1, characterized in that: Among them A set of actions or processes is stored in volatile memory or in non-volatile memory.

Images