A process judgment method for visual recognition based on place of origin

Abstract

The present invention discloses a process judgment method based on visual recognition of a place of origin, which belongs to the field of visual recognition technology of intelligent devices. It solves the problems of scene action data classification and processing and low data processing efficiency. The key points of its technical solution are that the data collected by the camera is processed by a machine vision AI recognition module, and then the data processed by the machine vision AI recognition module is sent to the action classification judgment module for action classification, and the data processed by the action classification judgment module is processed separately by a process judgment module and a timed task process judgment module, and then respectively handed over to the process warehousing and report generation modules for processing into report data, thereby achieving the effect of improving the data processing efficiency of action and process data.

Description

Technical Field

[0001] The present invention relates to the field of smart device visual recognition, and in particular, to a process judgment method for visual recognition based on location of ownership. Background Art

[0002] Place of origin is a scope concept, including but not limited to hospital beds, workbenches, cameras, operating rooms, storage rooms, etc.

[0003] Machine vision is used to manage hospital infection control services. Nursing services are generally a key component of hospital infection control. Nursing services vary widely, but they all require monitoring of multiple actions within a single service. We refer to these multiple actions as a process. This process is monitored for accuracy, duration, omissions, violations, and correct sequence of actions.

[0004] The actions reported by AI machine vision recognition may be the same action, actions in different locations, or actions by different people. The system needs to identify and judge these actions and group them into a common process. Only then can subsequent business calculations be carried out.

[0005] Among them, there is a scenario: each time a person works in a place of belonging, this place of belonging can be a hospital bed, workbench, camera, operating room, storage room, etc., or even a suitable scene in a non-hospital scenario; a work process is only completed in this one scene, and a work process includes some actions.

[0006] To solve the above scenario problems, a process judgment method for visual recognition based on location is currently needed. Summary of the Invention

[0007] The purpose of the present invention is to address the shortcomings of the existing technology and, at least to a certain extent, solve the technical problems in the related technology, and provide a process judgment method based on visual recognition of the place of origin, which is used to determine the process to which the action belongs, as well as the nature and name of the process, thereby giving the machine vision results related business capabilities.

[0008] In order to solve the above technical problems, the technical solution of the present invention is: a process judgment method based on visual recognition of the place of origin, including a camera, a machine vision AI recognition module, an action classification judgment module, a process judgment module, a scheduled task process judgment module, and a process warehousing and report generation module. The data collected by the camera is processed by the machine vision AI recognition module, and then the data processed by the machine vision AI recognition module is sent to the action classification judgment module for action classification. The data processed by the action classification judgment module is processed separately by the process judgment module and the scheduled task process judgment module, and then respectively handed over to the process warehousing and report generation module for processing into report data.

[0009] Preferably, the method for processing an action by the action classification judgment module includes the following steps: judging the attribution of the action based on the action data reported by the machine vision AI recognition module, and defining an attribution value;

[0010] Determine whether the set conditions are met.

[0011] First condition: call the action with the attribute in the memory according to the attribute value; Second condition: the action is the same as the previous one;

[0012] If the first and second conditions are met, the action with higher accuracy is saved and a null value is returned. If the first condition is not met, a null value is returned. If the first condition is met but the second condition is not met, the latest action is kept in the memory and the previous action is returned.

[0013] Preferably, the process judgment module further includes a method for pushing the action judgment algorithm and a process collection method, the steps of which include:

[0014] Determine whether the third condition is met: in the process set, there is a process to be changed;

[0015] If the third condition is met, the action is added to the corresponding process; if the third condition is not met, a new process is generated, the new process is placed in the process set, and the action is added to the corresponding process.

[0016] Preferably, the scheduled task process judgment module includes the following steps:

[0017] Start the task at a fixed time, find all the attributes in the memory, find the action in the memory according to the attribute, and judge whether the fourth condition set is met: the last time of the action is compared with the current time, and whether it exceeds the predetermined period.

[0018] If the fourth condition is met, the last action is placed in the process, the process judgment and reporting process are performed, and the process data in the process set is cleared.

[0019] If the fourth condition is not met, the process ends.

[0020] Preferably, the timed task judgment module automatically enters program execution by a timed interrupt.

[0021] Compared with the background technology, the technical effects of the present invention are mainly reflected in the following aspects:

[0022] 1. Non-contact measurement will not cause any damage to the observer or the observed, thereby improving the reliability of the system.

[0023] 2. It has a wider spectral response range, such as using infrared measurements that are invisible to the human eye, which expands the visual range of the human eye.

[0024] 3. Long-term stable work. It is difficult for humans to observe the same object for a long time, while machine vision can perform measurement, analysis and recognition tasks for a long time. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a schematic diagram of the overall architecture of the process judgment method based on visual recognition of the place of origin involved in the present invention;

[0026] Figure 2 This is a schematic diagram of the algorithm for determining the same action involved in the present invention, and obtaining the start and end time of each action;

[0027] Figure 3 This is an algorithm that concentrates all actions of the same process involved in the present invention, and is a schematic diagram of the algorithm indexed by location and action type;

[0028] Figure 4 This is a schematic diagram of the algorithm for determining the timing task process involved in the present invention, which obtains the last action and determines the process nature and name. DETAILED DESCRIPTION

[0029] The specific embodiments of the present invention are further described below in conjunction with the accompanying drawings to make the technical solutions of the present invention easier to understand and grasp.

[0030] Example:

[0031] Hand hygiene (handwashing, sanitary hand disinfection, surgical hand disinfection, etc.) during hospital infection, isolation of multidrug-resistant bacteria, and disinfection of surgical instruments and endoscopes are all very suitable for monitoring and management using machine vision because of its high accuracy and non-interventional nature.

[0032] The main purpose of the present invention is to provide a process judgment method for visual recognition based on location, which is used to determine the process to which an action belongs, as well as the nature and name of the process, thereby giving machine vision results related business capabilities.

[0033] The objectives of the present invention are achieved through the following technical solutions.

[0034] The method includes a camera, a machine vision AI recognition module, a process judgment module, and a process storage and report generation module. For example, Figure 1 .

[0035] The program method flow included in the process judgment module includes: the same action judgment method, the method of concentrating all actions in the same process, and the method of timing task judgment process, which correspond to Figure 2 、 Figure 3 、 Figure 4 It consists of three parts.

[0036] The same action judgment algorithm is used to determine whether multiple actions uploaded by the machine vision AI recognition program belong to the same action, and record the start time and end time of the same action. Figure 2 .

[0037] The same action judgment algorithm includes the following steps:

[0038] First, the method determines the ownership of the action data reported by the machine vision AI recognition program, uses the ownership as a keyword, and searches for the existence of the action with the ownership. If it does not exist, it returns null.

[0039] Furthermore, if it exists, it is determined whether the current action and the previous action of the attribute are the same action.

[0040] Further, if yes, override the previous action and return null.

[0041] Further, if not, the current action overwrites the previous action and returns to the previous action.

[0042] The algorithm for concentrating all actions of the same process is used to determine whether an action belongs to a certain process based on its attribution, and to concentrate all actions of a process into a set for further processing. Figure 3 .

[0043] All actions of the same process are centralized in the algorithm, which includes the following steps.

[0044] First, based on the action's affiliation, the process set is checked to see if a process based on that affiliation exists. If so, the action is added to the process. If not, a new process is generated, the action is added to the process, and the process is then added to the set.

[0045] The algorithm for determining the process of a scheduled task is to use a scheduled task to obtain the last action of the process, obtain the process from the process set, determine the nature and name of the process, and report it to the business system. For example, Figure 4 .

[0046] The algorithm for determining the timing task includes the following steps.

[0047] First, find out where everything belongs.

[0048] Furthermore, all attributes are traversed, and actions are found according to the attributes, and it is determined whether the last time of the action and the current time exceed a certain period.

[0049] Further, if not, then end.

[0050] Furthermore, if it exceeds, the last action is placed in the corresponding process in the process set, and the nature and name of the process are judged according to the corresponding process in the process set, and finally the process is reported and cleared from the process set.

[0051] Compared with other methods, it has the following advantages:

[0052] 1. Non-contact measurement will not cause any damage to the observer or the observed, thereby improving the reliability of the system.

[0053] 2. It has a wider spectral response range, such as using infrared measurements that are invisible to the human eye, which expands the visual range of the human eye.

[0054] 3. Long-term stable work. It is difficult for humans to observe the same object for a long time, while machine vision can perform measurement, analysis and recognition tasks for a long time.

[0055] Of course, the above are only typical examples of the present invention. In addition, the present invention may also have many other specific implementation methods. Any technical solutions formed by equivalent replacement or equivalent transformation fall within the scope of protection required by the present invention.

Claims

1. A process judgment method based on visual recognition of location, comprising a camera, a machine vision AI recognition module, an action classification judgment module, a process judgment module, a timed task process judgment module, and a process storage and report generation module, characterized by: The data collected by the camera is processed by the machine vision AI recognition module, and then the data processed by the machine vision AI recognition module is sent to the action classification judgment module for action classification. The data processed by the action classification judgment module is processed separately by the process judgment module and the timed task process judgment module, and then respectively handed over to the process storage and report generation modules to be processed into report data; The method for processing an action by the action classification judgment module includes the following steps: judging the attribution of the action based on the action data reported by the machine vision AI recognition module and defining an attribution value; Determine whether the set conditions are met. First condition: call the action with the attribute in the memory according to the attribute value; Second condition: the action is the same as the previous one; If the first and second conditions are met, the action with higher accuracy is saved and a null value is returned. If the first condition is not met, a null value is returned. If the first condition is met but the second condition is not met, the latest action is kept in the memory and the previous action is returned.

2. The process judgment method based on visual recognition of place of origin according to claim 1, characterized in that: The process judgment module also includes a method for pushing the action judgment algorithm and a process collection method, the steps of which include: Determine whether the third condition is met: in the process set, there is a process to be changed; If the third condition is met, the action is added to the corresponding process; if the third condition is not met, a new process is generated, the new process is placed in the process set, and the action is added to the corresponding process.

3. The process judgment method based on visual recognition of place of origin according to claim 1, characterized in that: The timing task process judgment module includes the following steps: Start the task at a fixed time, find all the attributes in the memory, find the action in the memory according to the attribute, and judge whether the fourth condition set is met: the last time of the action is compared with the current time, and whether it exceeds the predetermined period. If the fourth condition is met, the last action is placed in the process, the process judgment and reporting process are performed, and the process data in the process set is cleared. If the fourth condition is not met, the process ends.

4. The process judgment method based on visual recognition of place of origin according to claim 3 is characterized by: The timing task process judgment module automatically enters program execution by timing interruption.

Images