Search

CN-115346245-B - Assessment method and device for hand contact risk, electronic equipment and storage medium

CN115346245BCN 115346245 BCN115346245 BCN 115346245BCN-115346245-B

Abstract

The application discloses an assessment method, a device, electronic equipment and a storage medium of hand contact risk, wherein the method comprises the steps of collecting video data of an indoor scene by utilizing a depth camera, wherein the video data comprise two-dimensional color images and depth distance data, respectively calculating 25 points of human skeleton point coordinates and at most 21 points of hand skeleton key point coordinates in a first video stream and a second video stream, and assessing the ROI of an object with surface contact risk and Mask coordinates belonging to the surface of the object in the ROI, so as to identify a contact risk value of each person in any time period in the indoor scene and a contact exposure dose value of any object in the scene. Therefore, the problems that the hand contact risk of personnel cannot be quantitatively evaluated in an indoor scene, the contact exposure dose of the object surface in a specific scene and the like in the related technology are solved.

Inventors

  • Xi Ziang
  • LIU LI
  • YUAN YUFENG
  • WANG YING

Assignees

  • 清华大学
  • 武汉大学中南医院

Dates

Publication Date
20260505
Application Date
20220824

Claims (8)

  1. 1. A method for assessing risk of hand contact, comprising the steps of: acquiring video data of an indoor scene by using a depth camera, wherein the video data comprises a two-dimensional color image and depth distance data; Calculating 25 points of human skeleton point coordinates and at most 21 points of hand skeleton key point coordinates in a first video stream and a second video stream respectively according to the two-dimensional color image, and evaluating ROI of an object with surface contact risk and Mask coordinates belonging to the surface of the object in the ROI Identifying a contact risk value to which each person in any time period in the indoor scene belongs and a contact exposure dose value of any object in the scene according to the 25-point human skeleton point coordinates, the at most 21-point hand skeleton key point coordinates and the ROI of the object of which the evaluation surface is in contact risk and Mask coordinates belonging to the surface of the object in the ROI; The calculating, in the first video stream and the second video stream, 25 points of coordinates of human skeleton points, at most 21 points of coordinates of key points of hand skeleton points, and Mask coordinates belonging to the surface of the object in the ROI and evaluating the risk of surface contact, respectively, the two-dimensional color image includes: Giving each detected person a unique identity mark, binding the human skeleton point coordinates with the identity mark, tracking the positions of all the persons by a single person in a camera view scene, and setting up a contact risk value for each identity mark; for people with front and visual angles meeting preset conditions, taking the extracted coordinates of at most 21 points of hand skeleton key points as marking points for evaluating hand contact risks; For people with hands blocked and visual angles not meeting the preset conditions, estimating the positions of hand contact marking points by using left/right elbows and left/right wrists in skeleton points of the human body; The calculating the coordinates of the 25 points of human skeleton points, the coordinates of the key points of the hand skeleton at most 21 points, the ROI of the object for evaluating the risk of surface contact and Mask coordinates belonging to the surface of the object in the ROI in the first video stream and the second video stream respectively comprises the following steps: Giving each detected object a unique object ID, binding Mask coordinates with the object ID, tracking all objects in the view scene of the camera, and setting up a contact exposure dose value for each object ID; and marking distance information for each pixel point by utilizing the depth distance data.
  2. 2. The method according to claim 1, wherein the identifying the contact risk value to which each person in any time period in the indoor scene belongs and the contact exposure dose value of any object in the scene based on the 25-point human skeletal point coordinates and the at most 21-point hand skeletal key point coordinates and the ROI of the object at which the evaluation surface is at contact risk and Mask coordinates belonging to the object surface within the ROI comprises: judging whether the coordinates of the hand marking points in each frame of image are overlapped with Mask coordinates of each object in the ROI or not; If the hand mark points are overlapped, calculating the difference between the depth distance of the hand mark point coordinates and the depth distance in the original depth information, and determining the contact state according to the difference.
  3. 3. The method of claim 2, wherein said determining a contact state from said gap comprises: and when the difference is smaller than or equal to 0.1m, judging the contact state to be a contact state, otherwise, judging the contact state to be a non-contact state.
  4. 4. An assessment device for hand contact risk, comprising: The acquisition module is used for acquiring video data of an indoor scene by using a depth camera, wherein the video data comprises a two-dimensional color image and depth distance data; a calculation module for calculating 25 points of human skeleton point coordinates and at most 21 points of hand skeleton key point coordinates in a first video stream and a second video stream respectively, and assessing the ROI of an object at risk of surface contact and Mask coordinates belonging to the surface of the object in the ROI The identifying module is used for identifying a contact risk value of each person in any time period in the indoor scene and a contact exposure dose value of any object in the scene according to the 25-point human skeleton point coordinates, the at most 21-point hand skeleton key point coordinates and the ROI of the object of the evaluation surface contact risk and Mask coordinates belonging to the surface of the object in the ROI; wherein the computing module comprises: The first tracking unit is used for giving each detected person a unique identity mark, binding the coordinates of the skeleton points of the human body with the identity mark, tracking the positions of all the persons by a single person in a camera view scene, and setting up a contact risk value for each identity mark; The extraction unit is used for taking the extracted coordinates of at most 21 points of hand skeleton key points as marking points for evaluating hand contact risks for the personnel in front and with visual angles meeting preset conditions; the estimating unit is used for estimating the hand contact marking point position by using the left/right elbows and the left/right wrists in the skeleton points of the human body for the person with the hands blocked and the visual angle not meeting the preset conditions; the computing module further includes: The second tracking unit is used for giving a unique object ID to each detected object, binding Mask coordinates with the object ID, tracking all objects in the view scene of the camera, and setting up a contact exposure dose value for each object ID; and the marking unit is used for marking the distance information for each pixel point by utilizing the depth distance data.
  5. 5. The apparatus of claim 4, wherein the identification module comprises: The judging unit is used for judging whether the coordinates of the hand marking points in each frame of image are overlapped with Mask coordinates of each object in the ROI or not; And the determining unit is used for calculating the difference between the depth distance of the hand marking point coordinates and the depth distance in the original depth information when the hand marking point coordinates are coincident, and determining the contact state according to the difference.
  6. 6. The apparatus according to claim 5, wherein the determining unit is specifically configured to, And when the difference is smaller than or equal to 0.1m, judging the contact state to be a contact state, otherwise, judging the contact state to be a non-contact state.
  7. 7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the method of assessing risk of hand contact according to any one of claims 1-3.
  8. 8. A computer-readable storage medium, on which a computer program is stored, characterized in that the program is executed by a processor for implementing a method of assessing risk of hand contact according to any one of claims 1-3.

Description

Assessment method and device for hand contact risk, electronic equipment and storage medium Technical Field The present application relates to the field of risk assessment technologies, and in particular, to a method and apparatus for assessing risk of hand contact, an electronic device, and a storage medium. Background At present, people pay attention to hand hygiene safety gradually, and the risk of hand hygiene infection is larger for hospitals, and hand hygiene is also one of important assessment indexes of hospital infection management work. At the same time, techniques for detecting hand hygiene have also emerged. Related art color RGB two-dimensional image evaluation depth information of a target can be acquired by using an RGB (Red-Green-Blue) camera to judge whether the hand of a person to be identified is in contact with an object, thereby detecting the hand hygiene condition. However, the related art only evaluates depth information through a two-dimensional image, which is difficult to effectively improve the accuracy of RGB image recognition, and can not quantify the risk of hand contact of an evaluator and the exposure dose of contact of the surface of an object in a specific scene in an indoor scene, which is needed to be solved. Disclosure of Invention The application provides an assessment method and device for hand contact risk, electronic equipment and storage medium, and aims to solve the problems that the hand contact risk of personnel cannot be quantitatively assessed in an indoor scene, the contact exposure dose of the surface of an object in a specific scene is quantitatively assessed in the related technology. An embodiment of the first aspect of the application provides a method for evaluating a hand contact risk, which comprises the steps of acquiring video data of an indoor scene by using a depth camera, wherein the video data comprises a two-dimensional color image and depth distance data, respectively calculating 25 points of human skeleton point coordinates and at most 21 points of hand skeleton key point coordinates in a first video stream and a second video stream and Mask coordinates belonging to the surface of the object in the ROI, and identifying a contact risk value of each person belonging to any time period in the indoor scene and a contact exposure dose value of any object in the scene according to the 25 points of human skeleton point coordinates and the at most 21 points of hand skeleton key point coordinates and the Mask coordinates belonging to the surface of the object in the ROI. Optionally, in one embodiment of the present application, the calculating the coordinates of the 25 points of human skeleton points and the coordinates of the at most 21 points of human skeleton points in the first video stream and the second video stream respectively, and the ROI of the object with the assessed surface contact risk and the Mask coordinates belonging to the surface of the object in the ROI include assigning each detected person a unique identity, binding the coordinates of the human skeleton points with the identity, tracking the positions of all the persons by a single person in the view scene of the camera, setting up a contact risk value for each identity, using the coordinates of the at most 21 points of human skeleton points extracted as the marking points for assessing the hand contact risk for persons with the front and view angles meeting the preset conditions, and using the left/right elbows and the left/right wrists in the human skeleton points for persons with the hands blocked and the view angles not meeting the preset conditions to assess the positions of the hand contact marking points. Optionally, in one embodiment of the present application, the calculating the two-dimensional color image in the first video stream and the second video stream respectively calculates 25 points of human skeleton point coordinates and at most 21 points of hand skeleton key point coordinates, and ROI of the object for evaluating the risk of surface contact and Mask coordinates belonging to the surface of the object in the ROI, including assigning each detected object a unique object ID (Identity Document, identity), binding Mask coordinates with the object ID, tracking all objects in the camera view scene, and setting up a contact exposure dose value for each object ID, and marking distance information for each pixel point by using the depth distance data. Optionally, in one embodiment of the present application, the identifying the contact risk value of each person belonging to any time period in the indoor scene and the contact exposure dose value of any object in the scene according to the 25-point human skeleton point coordinates, the at most 21-point hand skeleton key point coordinates, the ROI of the object with the evaluation surface contact risk, and Mask coordinates belonging to the object surface in the ROI includes determining whether the hand marking point coordi