A fall detection method for elderly patients for fall prevention management
By analyzing the handrails, stools and water-stained areas in the ward images of geriatric patients, and combining the direction of movement and deviation to calculate the possibility of falling, the problem of difficulty in accurately predicting falls in the existing technology is solved, and the prediction of falls is improved Accuracy and timeliness.
Patent Information
- Application Number
- CN202411585011.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-07
- Publication Date
- 2025-01-03
- Estimated Expiration
- 2044-11-07
AI Technical Summary
The prior art is difficult to accurately and timely predict the possibility of elderly patients falling in hospitals, resulting in the inability to effectively prevent secondary injuries.
By obtaining the ward images of elderly patients, identifying the handrail area, stool area and water-stained area, calculating the angle between the fall environmental impact factor and the direction of movement, and calculating the possibility of falling based on the deviation, using neural network and Otsu threshold segmentation algorithm for fall detection.
It improves the accuracy of predicting falls for elderly patients, can promptly identify patients who are about to fall, and reduces the risk of secondary injury.
Smart Images
Figure CN119091130B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to a fall detection method for elderly patients for fall prevention management. Background Art
[0002] A fall is a condition in which an individual falls to the ground or on a plane lower than the initial position due to a sudden, involuntary, and unintentional change in body position. According to statistics, one-third of the elderly fall at least once a year. As the physical functions of elderly patients gradually decline, the consequences of falls may be more serious for them. Therefore, fall prevention for elderly patients has always been the focus of the geriatric department of the hospital to ensure the safety of elderly patients and prevent secondary injuries. For elderly patients hospitalized in the hospital, as they get out of bed and move around, there is a possibility of falling at any time. In order to avoid this problem, patients' families and hospitals often use monitoring methods. However, this method may not be able to predict falls in elderly patients in a timely and accurate manner. Summary of the invention
[0003] In order to solve the above problems, the present invention provides a fall detection method for elderly patients for fall prevention management, the method comprising:
[0004] Obtain a set of geriatric ward images, each geriatric ward image including a number of geriatric patients;
[0005] The last geriatric ward image in the geriatric ward image set is recorded as the target geriatric ward image, and the water stain areas in the handrail area, the stool area and the floor area in the target geriatric ward image are obtained; the fall environment influencing factor of each geriatric patient is obtained according to the handrail area and the distance between the water stain area and the geriatric patient; the movement direction angle of each geriatric patient in each geriatric ward image is obtained according to the position of each geriatric patient in different geriatric ward images; the fall environment influencing degree of each geriatric patient is obtained according to the fall environment influencing factor and the stool area and the movement direction angle of the geriatric patient;
[0006] According to the angle of motion direction and the deviation of each elderly patient in different geriatric ward images, the possible factors affecting the fall of each elderly patient are obtained; according to the possible factors affecting the fall and the degree of influence of the fall environment, the possibility of fall of each elderly patient is obtained;
[0007] Fall detection is performed on each elderly patient in the target elderly ward image according to the possibility of falling.
[0008] Preferably, the method of obtaining the water stain area in the armrest area, the stool area and the floor area in the target geriatric ward image includes:
[0009] Input the target geriatric ward image into the trained neural network to obtain the floor area, handrail area and stool area in the target geriatric ward image;
[0010] The floor area in the target geriatric ward image is segmented by using the Otsu threshold segmentation algorithm to obtain a segmentation result map of the target geriatric ward image; the connected domain in the segmentation result map corresponds to the area in the target geriatric ward image as the water stain area in the target geriatric ward image.
[0011] Preferably, the step of obtaining the fall environment influencing factor of each elderly patient according to the handrail area and the distance between the water stain area and the elderly patient comprises the following specific methods:
[0012] According to the handrail area, obtain the handrail impact factor;
[0013] For any elderly patient in the target elderly ward image, the number of all pixels in the water stain area in the target elderly ward image is recorded as the area of the water stain area; the Euclidean distance between the centroid of the water stain area and the centroid of the area to which the elderly patient belongs is recorded as the water stain distance influence factor; the ratio of the water stain area to the water stain distance influence factor is recorded as the water stain influence factor; the product of the inverse of the handrail influence factor and the water stain influence factor is taken as the fall environment influence factor of the elderly patient.
[0014] Preferably, the obtaining of the handrail influence factor includes the following specific methods:
[0015] If the handrail area exists in the target geriatric ward image, 1 is used as the handrail influence factor; if the handrail area does not exist in the target geriatric ward image, 0.1 is used as the handrail influence factor.
[0016] Preferably, the method of obtaining the movement direction angle of each elderly patient in each image of the elderly ward according to the position of each elderly patient in different images of the elderly ward includes the following specific methods:
[0017] The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The connecting straight line between the centroids of the areas to which the elderly patients belong is recorded as the first straight line; the angle between the first straight line and the horizontal line is recorded as the second straight line. Image of the geriatric ward The movement direction angle of an elderly patient.
[0018] Preferably, the step of obtaining the fall environment impact degree of each elderly patient according to the fall environment impact factor, the stool area, and the movement direction angle of the elderly patient includes the following specific methods:
[0019] For any elderly patient in the target elderly ward image, the average of the movement angles of the elderly patient in all the elderly ward images in the elderly ward image set is used as the movement direction angle of the elderly patient;
[0020] The Euclidean distance between the centroid of the stool area in the target geriatric ward image and the centroid of the area to which the elderly patients belong is recorded as the obstacle distance; the product of the sine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the first product; the product of the cosine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the second product; the sum of the first product and the second product is recorded as the first sum value; the ratio of the number of all pixel points in the stool area in the target geriatric ward image to the sum value is recorded as the obstacle influence factor; the product of the elderly patient's fall environment influence factor and the obstacle influence factor is taken as the fall environment influence degree of the elderly patient.
[0021] Preferably, the method of obtaining the possible influencing factors of the fall of each elderly patient according to the angle of the movement direction and the deviation of each elderly patient in different elderly ward images includes the following specific methods:
[0022] According to the deviation of each elderly patient in the images of different elderly wards, the movement speed and movement deviation degree of each elderly patient are obtained;
[0023] For any elderly patient in the target geriatric ward image, the product of the variance of the elderly patient's motion angle in all geriatric ward images in the geriatric ward image set, the elderly patient's motion speed and the elderly patient's motion deviation degree is taken as a possible influencing factor for the elderly patient's fall.
[0024] Preferably, the method of obtaining the movement speed and movement deviation degree of each elderly patient according to the deviation of each elderly patient in different elderly ward images includes the following specific methods:
[0025] The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as Image of the geriatric ward The first distance for an elderly patient;
[0026] Among all the geriatric ward images, The variance of the first distance of the elderly patients is used as the The degree of motor deviation in elderly patients;
[0027] The first image of the geriatric ward in the geriatric ward image collection is The centroid of the area to which the elderly patient belongs is the same as the centroid of the area to which the elderly patient belongs in the target elderly patient ward image. The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as The total displacement distance of the elderly patients; The ratio of the total offset distance of the geriatric patients to the total number of all geriatric ward images in the geriatric ward image set is taken as the The movement speed of an elderly patient.
[0028] Preferably, the method of obtaining the possibility of falling of each elderly patient according to the possible factors affecting falling and the degree of influence of the falling environment includes the following specific methods:
[0029] For any elderly patient in the target elderly ward image, the normalized value of the product of the elderly patient's possible fall influencing factor and the degree of fall environment influence on the elderly patient is used as the elderly patient's fall possibility.
[0030] Preferably, the method of performing fall detection on each elderly patient in the target elderly patient ward image according to the fall possibility includes the following specific methods:
[0031] Preset a threshold parameter For any elderly patient in the target elderly ward image, if the elderly patient's fall probability is greater than or equal to the threshold parameter , the elderly patient is recorded as a patient who is about to undergo iteration.
[0032] The beneficial effects of the technical solution of the present invention are as follows: the present invention obtains the degree of environmental influence of falls for each elderly patient based on the fall environment influencing factors and the stool area, as well as the angle of the movement direction of the elderly patient; obtains the possible influencing factors of falls for each elderly patient based on the angle of the movement direction and the deviation of each elderly patient in different elderly ward images; obtains the possibility of falls for each elderly patient based on the possible influencing factors of falls and the degree of environmental influence of falls; performs fall detection on each elderly patient in the target elderly ward image based on the possibility of falls; thereby, by analyzing the influence of the elderly patient's own factors and the influence of the surrounding environmental factors, the possibility of falls for the elderly patient is obtained, thereby improving the accuracy of predicting falls in elderly patients. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0034] Figure 1 A flowchart of the steps of a fall detection method for elderly patients for fall prevention management according to the present invention;
[0035] Figure 2 This is a characteristic relationship flow chart of a fall detection method for elderly patients used for fall prevention management of the present invention. DETAILED DESCRIPTION
[0036] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a fall detection method for elderly patients for fall prevention management proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.
[0037] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
[0038] The specific scheme of the fall detection method for elderly patients for fall prevention management provided by the present invention is described in detail below with reference to the accompanying drawings.
[0039] See also Figure 1 , which shows a flow chart of the steps of a fall detection method for elderly patients for fall prevention management provided by an embodiment of the present invention, the method comprising the following steps:
[0040] Step S001: Acquire a set of geriatric ward images.
[0041] It should be noted that when elderly patients are hospitalized in the hospital, they may fall as they get out of bed and walk, which may cause secondary injuries to the elderly patients. By analyzing the image information collected by the camera, the possibility of falls in the elderly patients in the image can be obtained, which helps nursing staff complete the care of the elderly.
[0042] Specifically, we first need to collect a set of images of geriatric wards. The specific process is as follows:
[0043] A high-definition camera is placed on the corner of the roof near the ward door. The images of the geriatric ward are collected using the high-definition camera every 1 second as a sampling moment, and the collection lasts for 1 minute in total. The set of geriatric ward images at each sampling moment is taken as the geriatric ward image set.
[0044] Preferably, for any geriatric ward image in the geriatric ward image set, the geriatric ward image is input into a trained neural network to obtain a plurality of geriatric patient areas in the geriatric ward image; wherein the neural network used in this embodiment is YOLOv3, and the method for obtaining the data set for training the neural network is:
[0045] A large number of geriatric ward images are collected, and each geriatric patient area is artificially divided into two areas in each geriatric ward image using a bounding box, and the bounding box is used as a label for each geriatric ward image; a large number of geriatric ward images and their corresponding labels are collected to form a data set; the neural network is trained using the data set, and the loss function used in the training process is a mean square error loss function; the specific training process is a well-known content of the neural network, and the specific training process will not be repeated in this embodiment.
[0046] So far, a set of geriatric ward images has been obtained through the above method.
[0047] Step S002: record the last geriatric ward image in the geriatric ward image set as the target geriatric ward image, and obtain the water stain areas in the armrest area, stool area and floor area in the target geriatric ward image; obtain the fall environment influencing factor of each elderly patient according to the armrest area and the distance between the water stain area and the elderly patient; obtain the movement direction angle of each elderly patient in each geriatric ward image according to the position of each elderly patient in different geriatric ward images; obtain the fall environment influencing degree of each elderly patient according to the fall environment influencing factor and the stool area and the movement direction angle of the elderly patient.
[0048] It should be noted that when judging the possibility of elderly patients falling in hospital wards, the environmental factors of hospital wards have a relatively important impact on it; when there are water stains on the floor of the hospital ward and there are no handrails in the middle part of the ward wall, the possibility of elderly patients falling will be greater; on the contrary, there are no water stains on the floor of the hospital ward and there are handrails in the middle part of the ward wall, the possibility of elderly patients falling will be greatly reduced.
[0049] 1. Obtain the environmental factors affecting falls for each elderly patient.
[0050] It should be noted that the smaller the distance between the water stain area in the target elderly patient's ward image and the area to which the elderly patient belongs, the closer the elderly patient is to the water stain area, and the larger the area of the water stain area, the greater the possibility of the elderly patient falling; therefore, when judging the possibility of elderly patients falling in hospital wards, the hospital ward environmental factors have a relatively important influence on it.
[0051] Preferably, in some implementations of the embodiments of the present invention, since water stains on the floor area in the geriatric ward image usually absorb light to form a darker area, the water stain area and other dry areas on the floor area are distinguished based on this feature; the last geriatric ward image in the geriatric ward image set is recorded as the target geriatric ward image; the target geriatric ward image is input into the trained neural network to obtain the floor area, armrest area and stool area in the target geriatric ward image; wherein the neural network used in this embodiment is YOLOv3, and the method for obtaining the data set for training the neural network is:
[0052] A large number of geriatric ward images are collected, and a floor area, an armrest area and a stool area are artificially divided into two areas in each geriatric ward image using a bounding box, and the bounding box is used as a label for each geriatric ward image; a large number of geriatric ward images and their corresponding labels are collected to form a data set; the neural network is trained using the data set, and the loss function used in the training process is a mean square error loss function; the specific training process is a well-known content of the neural network, and the specific training process will not be repeated in this embodiment.
[0053] The floor area in the target geriatric ward image is segmented by using the Otsu threshold segmentation algorithm to obtain a segmentation result map of the target geriatric ward image; the connected domain in the segmentation result map corresponds to the area in the target geriatric ward image as the water stain area in the target geriatric ward image.
[0054] Among them, the Otsu threshold segmentation algorithm is a prior art and will not be described in detail in this embodiment.
[0055] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining the fall environment influencing factor of each elderly patient according to the handrail area and the distance between the water stain area and the elderly patient is:
[0056] According to the handrail area, obtain the handrail impact factor.
[0057] For any elderly patient in the target elderly ward image, the number of all pixels in the water stain area in the target elderly ward image is recorded as the area of the water stain area; the Euclidean distance between the centroid of the water stain area and the centroid of the area to which the elderly patient belongs is recorded as the water stain distance influence factor; the ratio of the water stain area to the water stain distance influence factor is recorded as the water stain influence factor; the product of the inverse of the handrail influence factor and the water stain influence factor is taken as the fall environment influence factor of the elderly patient.
[0058] The specific formula is:
[0059]
[0060] In the formula, It represents the environmental factors affecting falls in elderly patients; represents the number of all pixels in the water-stained area of the target geriatric ward image; The Euclidean distance between the centroid of the water-stained area and the centroid of the area to which the elderly patients belong; Represents the handrail influence factor.
[0061] It should be noted that when there is a water stain area in the target elderly patient's ward image, and the distance between the area and the elderly patient's area is smaller, it means that the elderly patient is closer to the water stain area at this time, and the larger the area of the water stain area is, the greater the possibility of the elderly patient falling.
[0062] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining the handrail influence factor is:
[0063] If the handrail area exists in the target geriatric ward image, 1 is used as the handrail influence factor; if the handrail area does not exist in the target geriatric ward image, 0.1 is used as the handrail influence factor.
[0064] It should be noted that if there are handrails in the image of the geriatric ward, the possibility of the elderly patient falling will be appropriately reduced; on the contrary, if there are no handrails, the possibility of the elderly patient falling will increase.
[0065] At this point, the environmental influencing factors for falls of each elderly patient are obtained.
[0066] 2. Obtain the degree of environmental impact on falls for each elderly patient.
[0067] It should be noted that the degree of environmental impact on falls of elderly patients can be reflected by water stains and handrails in the environment, but it is not accurate enough to only use this as a representative to quantify the degree of environmental impact on falls of elderly patients. It is also necessary to combine whether there are obstacles in the process of elderly patients' movement and the possible impact of the obstacles. These factors will also greatly affect the degree of environmental impact on falls of elderly patients.
[0068] Preferably, in some implementations of the embodiments of the present invention, a specific method for obtaining the motion direction angle of each elderly patient in each elderly patient room image according to the position of each elderly patient in different elderly patient room images is:
[0069] The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The connecting straight line between the centroids of the areas to which the elderly patients belong is recorded as the first straight line; the angle between the first straight line and the horizontal line is recorded as the second straight line. Image of the geriatric ward The movement direction angle of an elderly patient.
[0070] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining the degree of fall environment influence of each elderly patient according to the fall environment influence factor and the stool area, and the angle of the movement direction of the elderly patient is:
[0071] For any elderly patient in the target elderly ward image, the average of the movement angles of the elderly patient in all the elderly ward images in the elderly ward image set is used as the movement direction angle of the elderly patient;
[0072] The Euclidean distance between the centroid of the stool area in the target geriatric ward image and the centroid of the area to which the elderly patients belong is recorded as the obstacle distance; the product of the sine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the first product; the product of the cosine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the second product; the sum of the first product and the second product is recorded as the first sum value; the ratio of the number of all pixel points in the stool area in the target geriatric ward image to the sum value is recorded as the obstacle influence factor; the product of the elderly patient's fall environment influence factor and the obstacle influence factor is taken as the fall environment influence degree of the elderly patient.
[0073] The specific formula is:
[0074]
[0075] In the formula, Indicates the degree of environmental influence on falls in elderly patients; It represents the environmental factors affecting falls in elderly patients; represents the number of all pixels in the stool area in the target geriatric ward image; represents the Euclidean distance between the centroid of the stool area in the target geriatric ward image and the centroid of the area to which the elderly patients belong; Represents the sine value of the angle of the movement direction of the elderly patient; Represents the cosine value of the angle of the elderly patient's movement direction.
[0076] It should be noted that the larger the area occupied by the obstacle in the direction of movement of the elderly patient and the closer the distance between the obstacle and the elderly patient, the greater the possibility that the elderly patient will fall during walking.
[0077] At this point, the degree of environmental impact of falls for each elderly patient is obtained through the above method.
[0078] Step S003: According to the angle of the movement direction and the deviation of each elderly patient in different geriatric ward images, the possible factors affecting the fall of each elderly patient are obtained; according to the possible factors affecting the fall and the degree of influence of the fall environment, the possibility of fall of each elderly patient is obtained.
[0079] It should be noted that for elderly patients, in addition to the influencing factors caused by environmental factors, the elderly patients themselves will also have a certain degree of influence on their falls; when the elderly patients have multiple consecutive changes in walking direction, it means that the elderly patients may have become shaky at this time. Further combined with the gait characteristics of the elderly patients, the impact of the elderly patients' own factors on their possibility of falling is analyzed; then combined with the degree of influence of the falling environment, the possible influencing factors of the elderly patients' falls are analyzed to obtain the possibility of falls in the elderly patients.
[0080] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining the possible influencing factors of the fall of each elderly patient according to the angle of the motion direction and the deviation of each elderly patient in different elderly ward images is:
[0081] According to the deviation of each elderly patient in the images of different elderly wards, the movement speed and movement deviation degree of each elderly patient are obtained;
[0082] For any elderly patient in the target geriatric ward image, the product of the variance of the elderly patient's motion angle in all geriatric ward images in the geriatric ward image set, the elderly patient's motion speed and the elderly patient's motion deviation degree is taken as a possible influencing factor for the elderly patient's fall.
[0083] The specific formula is:
[0084]
[0085] In the formula, Indicates possible factors affecting falls in elderly patients; represents the variance of the motion angle of the elderly patients in all the geriatric ward images; Indicates the movement speed of elderly patients; Indicates the degree of movement deviation in elderly patients.
[0086] It should be noted that for elderly patients, a faster moving speed may cause their center of gravity to become unstable and lead to falls; and when the elderly patients' movement angle keeps changing, it means that the elderly patients are shaky at this time. Further combined with the patient's leg performance, when the length of each step of the elderly patients is very different, the possibility of falls will increase, and the influencing factor of their own factors on the possibility of falls in elderly patients should increase.
[0087] Preferably, in some implementations of the embodiments of the present invention, the method for obtaining the movement speed and movement deviation degree of each elderly patient includes:
[0088] The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as Image of the geriatric ward The first distance for an elderly patient;
[0089] Among all the geriatric ward images, The variance of the first distance of the elderly patients is used as the The degree of motor deviation in elderly patients;
[0090] The first image of the geriatric ward in the geriatric ward image collection is The centroid of the area to which the elderly patient belongs is the same as the centroid of the area to which the elderly patient belongs in the target elderly patient ward image. The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as The total displacement distance of the elderly patients; The ratio of the total offset distance of the geriatric patients to the total number of all geriatric ward images in the geriatric ward image set is taken as the The movement speed of an elderly patient.
[0091] At this point, the possible influencing factors for falls in each elderly patient are obtained.
[0092] Preferably, in some implementations of the embodiments of the present invention, the specific method for obtaining the possibility of falling of each elderly patient according to the possible influencing factors of falling and the degree of influence of the falling environment is:
[0093] For any elderly patient in the target elderly ward image, the normalized value of the product of the elderly patient's possible fall influencing factor and the degree of fall environment influence on the elderly patient is used as the elderly patient's fall possibility.
[0094] The specific formula is:
[0095]
[0096] In the formula, Indicates the likelihood of falls in elderly patients; Indicates possible factors affecting falls in elderly patients; Indicates the degree of environmental influence on falls in elderly patients; represents the linear normalization function.
[0097] So far, the fall probability of each elderly patient is obtained through the above method.
[0098] Step S004: performing fall detection on each elderly patient in the target elderly patient ward image according to the fall probability.
[0099] It should be noted that due to the possibility of a shortage of nursing staff in hospitals and the possibility of multiple elderly patients falling at the same time captured by video surveillance, in order to further optimize prevention management, it is necessary to analyze the possibility of falls among elderly patients and determine the fall trend of each elderly patient so that nursing staff can deal with different situations.
[0100] Preferably, in some implementations of the embodiments of the present invention, the specific method for performing fall detection on each elderly patient in the target elderly patient ward image according to the fall possibility is:
[0101] Preset a threshold parameter , wherein this embodiment is based on This example is described as an example, and this embodiment is not specifically limited. It depends on the specific implementation situation.
[0102] For any elderly patient in the target elderly ward image, if the elderly patient's fall probability is greater than or equal to the threshold parameter , the elderly patient is recorded as a patient who is about to have an iteration and needs immediate care from medical staff.
[0103] See also Figure 2 , which shows a characteristic relationship flow chart of a fall detection method for elderly patients used in fall prevention management.
[0104] At this point, this embodiment is completed.
[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A fall detection method for elderly patients for fall prevention management, characterized in that: The method comprises the following steps: Obtain a set of geriatric ward images, each geriatric ward image including a number of geriatric patients; The last geriatric ward image in the geriatric ward image set is recorded as the target geriatric ward image, and the water stain areas in the handrail area, the stool area and the floor area in the target geriatric ward image are obtained; the fall environment influencing factor of each geriatric patient is obtained according to the handrail area and the distance between the water stain area and the geriatric patient; the movement direction angle of each geriatric patient in each geriatric ward image is obtained according to the position of each geriatric patient in different geriatric ward images; the fall environment influencing degree of each geriatric patient is obtained according to the fall environment influencing factor and the stool area and the movement direction angle of the geriatric patient; The specific method of obtaining the fall environment influencing factor of each elderly patient according to the handrail area and the distance between the water stain area and the elderly patient is as follows: According to the handrail area, obtain the handrail impact factor; For any elderly patient in the target elderly ward image, the number of all pixels in the water stain area in the target elderly ward image is recorded as the water stain area area; the Euclidean distance between the centroid of the water stain area and the centroid of the area to which the elderly patient belongs is recorded as the water stain distance influence factor; the ratio of the water stain area to the water stain distance influence factor is recorded as the water stain influence factor; the product of the reciprocal of the handrail influence factor and the water stain influence factor is taken as the fall environment influence factor of the elderly patient; The method of obtaining the movement direction angle of each elderly patient in each elderly patient room image according to the position of each elderly patient in different elderly patient room images includes the following specific methods: The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The connecting straight line between the centroids of the areas to which the elderly patients belong is recorded as the first straight line; the angle between the first straight line and the horizontal line is recorded as the second straight line. Image of the geriatric ward The angle of movement direction of the elderly patients; The specific method of obtaining the fall environment impact degree of each elderly patient according to the fall environment impact factor, the stool area, and the movement direction angle of the elderly patient is as follows: For any elderly patient in the target elderly ward image, the average of the movement angles of the elderly patient in all the elderly ward images in the elderly ward image set is used as the movement direction angle of the elderly patient; The Euclidean distance between the centroid of the stool area in the target geriatric ward image and the centroid of the area to which the elderly patients belong is recorded as the obstacle distance; the product of the sine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the first product; the product of the cosine value of the elderly patient's movement direction angle and the obstacle distance is recorded as the second product; the sum of the first product and the second product is recorded as the first sum value; the ratio of the number of all pixel points in the stool area in the target geriatric ward image to the sum value is recorded as the obstacle influence factor; the product of the elderly patient's fall environment influence factor and the obstacle influence factor is used as the fall environment influence degree of the elderly patient; According to the angle of motion direction and the deviation of each elderly patient in different geriatric ward images, the possible factors affecting the fall of each elderly patient are obtained; according to the possible factors affecting the fall and the degree of influence of the fall environment, the possibility of fall of each elderly patient is obtained; The method of obtaining the possible influencing factors of the fall of each elderly patient according to the angle of the movement direction and the deviation of each elderly patient in different elderly ward images includes the following specific methods: According to the deviation of each elderly patient in the images of different elderly wards, the movement speed and movement deviation degree of each elderly patient are obtained; For any elderly patient in the target elderly ward image, the product of the variance of the elderly patient's motion angle in all the elderly ward images in the elderly ward image set, the elderly patient's motion speed and the elderly patient's motion deviation degree is taken as a possible influencing factor for the elderly patient's fall; Fall detection is performed on each elderly patient in the target elderly ward image according to the possibility of falling.
2. A fall detection method for elderly patients for fall prevention management according to claim 1, characterized in that: The specific method of obtaining the water stain area in the armrest area, the stool area and the floor area in the target geriatric ward image includes: Input the target geriatric ward image into the trained neural network to obtain the floor area, handrail area and stool area in the target geriatric ward image; The floor area in the target geriatric ward image is segmented by using the Otsu threshold segmentation algorithm to obtain a segmentation result map of the target geriatric ward image; the connected domain in the segmentation result map corresponds to the area in the target geriatric ward image as the water stain area in the target geriatric ward image.
3. A fall detection method for elderly patients for fall prevention management according to claim 1, characterized in that: The specific method of obtaining the handrail influence factor includes: If the handrail area exists in the target geriatric ward image, 1 is used as the handrail influence factor; if the handrail area does not exist in the target geriatric ward image, 0.1 is used as the handrail influence factor.
4. A fall detection method for elderly patients for fall prevention management according to claim 1, characterized in that: The specific method of obtaining the movement speed and movement deviation degree of each elderly patient according to the deviation of each elderly patient in different elderly ward images is as follows: The geriatric ward image collection Image of the geriatric ward The centroid of the area to which the elderly patient belongs is Image of the geriatric ward The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as Image of the geriatric ward The first distance for an elderly patient; Among all the geriatric ward images, The variance of the first distance of the elderly patients is used as the The degree of motor deviation in elderly patients; The first image of the geriatric ward in the geriatric ward image collection is The centroid of the area to which the elderly patient belongs is the same as the centroid of the area to which the elderly patient belongs in the target elderly patient ward image. The Euclidean distance between the centroids of the regions to which the elderly patients belong is recorded as The total displacement distance of the elderly patients; The ratio of the total offset distance of the geriatric patients to the total number of all geriatric ward images in the geriatric ward image set is taken as the The movement speed of an elderly patient.
5. A fall detection method for elderly patients for fall prevention management according to claim 1, characterized in that: The method of obtaining the possibility of falling for each elderly patient based on the possible factors affecting falling and the degree of influence of the falling environment includes: For any elderly patient in the target elderly ward image, the normalized value of the product of the elderly patient's possible fall influencing factor and the degree of fall environment influence on the elderly patient is used as the elderly patient's fall possibility.
6. A fall detection method for elderly patients for fall prevention management according to claim 1, characterized in that: The specific method of performing fall detection on each elderly patient in the target elderly patient ward image according to the fall possibility includes: Preset a threshold parameter For any elderly patient in the target elderly ward image, if the elderly patient's fall probability is greater than or equal to the threshold parameter , the elderly patient is recorded as a patient who is about to undergo iteration.