CN113903052A - Indoor human body collision alarm method and device based on image processing and mechanical analysis - Google Patents
Indoor human body collision alarm method and device based on image processing and mechanical analysis Download PDFInfo
- Publication number
- CN113903052A CN113903052A CN202111052459.2A CN202111052459A CN113903052A CN 113903052 A CN113903052 A CN 113903052A CN 202111052459 A CN202111052459 A CN 202111052459A CN 113903052 A CN113903052 A CN 113903052A
- Authority
- CN
- China
- Prior art keywords
- human body
- value
- collision
- image processing
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000004458 analytical method Methods 0.000 title claims abstract description 30
- 238000012545 processing Methods 0.000 title claims abstract description 28
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000004088 simulation Methods 0.000 claims abstract description 35
- 230000006378 damage Effects 0.000 claims abstract description 10
- 230000007613 environmental effect Effects 0.000 claims abstract description 7
- 210000000746 body region Anatomy 0.000 claims abstract description 4
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 4
- 238000003708 edge detection Methods 0.000 claims abstract description 4
- 210000000988 bone and bone Anatomy 0.000 claims description 45
- 238000001514 detection method Methods 0.000 claims description 27
- 238000006073 displacement reaction Methods 0.000 claims description 24
- 238000005259 measurement Methods 0.000 claims description 17
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 238000009529 body temperature measurement Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 239000000463 material Substances 0.000 claims description 3
- 208000010392 Bone Fractures Diseases 0.000 description 11
- 206010017076 Fracture Diseases 0.000 description 9
- 201000010099 disease Diseases 0.000 description 9
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 9
- 210000003414 extremity Anatomy 0.000 description 8
- 208000027418 Wounds and injury Diseases 0.000 description 6
- 208000014674 injury Diseases 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- 230000006399 behavior Effects 0.000 description 5
- 210000003205 muscle Anatomy 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 208000024779 Comminuted Fractures Diseases 0.000 description 2
- 238000005452 bending Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 230000036544 posture Effects 0.000 description 2
- 208000034656 Contusions Diseases 0.000 description 1
- 238000005299 abrasion Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000008280 blood Substances 0.000 description 1
- 210000004369 blood Anatomy 0.000 description 1
- 208000034526 bruise Diseases 0.000 description 1
- 230000003139 buffering effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 210000003275 diaphysis Anatomy 0.000 description 1
- 229940079593 drug Drugs 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 238000001125 extrusion Methods 0.000 description 1
- 210000002683 foot Anatomy 0.000 description 1
- 210000004744 fore-foot Anatomy 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000035876 healing Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 210000001930 leg bone Anatomy 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000474 nursing effect Effects 0.000 description 1
- 230000011164 ossification Effects 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 108010048734 sclerotin Proteins 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 210000004872 soft tissue Anatomy 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G08—SIGNALLING
- G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
- G08C17/00—Arrangements for transmitting signals characterised by the use of a wireless electrical link
- G08C17/02—Arrangements for transmitting signals characterised by the use of a wireless electrical link using a radio link
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses an indoor human body collision alarm method and device based on image processing and mechanical analysis, which comprises the following steps: s1, collecting environmental image information, and identifying human bodies and objects in the images; s2, Canny edge recognition is carried out on the object image, and openposition algorithm analysis is adopted for analyzing the human body limbs; s3, searching a contactable area of the object and the human body based on the height and the width of the contrast object and the human body pixel points, and judging a contacted trunk part according to the limbs of the human body; s4, based on the division in the step S3, carrying out edge detection on the human body region of the object at the same height, and based on the pixel position, obtaining the distance between different trunk positions in the contact region and the object; s5, acquiring the moving speed of the human body in real time, and judging the force applied when the human body collides with an object; and S6, introducing the force and the collision point received when the human body collides with the object into a simulation formula to obtain the information of the damage condition.
Description
Technical Field
The invention belongs to the technical field of anti-collision alarm, and particularly relates to an indoor human body collision alarm method and device based on image processing and mechanical analysis.
Background
Along with work load is big, elder generation's number increases, parent's the more and more circumstances of going out work simultaneously, old man at home, child are unattended, and old man's sclerotin is relatively poor in addition, child is less, bearing capacity is low, love the activity, be injured after colliding with indoor object personnel easily, and injured back medical personnel can't accurately learn the wounded condition, influence rescue efficiency. How to accurately know the injury condition of the personnel at home becomes a more important thing. In addition, the existing home monitoring equipment only reads infrastructure through a portable sensor, for example, CN110477882A, the finger ring type monitoring equipment and the monitoring system device read heartbeat blood oxygen and the like, the sensor is not in medical grade, misjudgment is easily caused, the monitoring system device needs to be worn artificially, and people are difficult to wear articles in an autonomous manner in real life. For example, CN 107454830A: although the health monitoring method and the health monitoring clothes judge whether the personnel need to be rescued according to the information, the condition of the wounded personnel before and after the accident is not inferred, so that the guardian who is on the spot before the accident easily cares about the potential danger, and the rescue efficiency is reduced because the rescuers after the accident are not aware of the necessary equipment.
Disclosure of Invention
In order to solve the night personnel when indoor activity because of the sight dim and can't learn injured and the old man child can't assess the injured condition after the atress collides under unmanned nursing at home with other people behind the object collision. The invention provides an indoor human body collision warning device based on image processing and mechanical analysis, which is combined with an infrared temperature sensor, a laser ranging sensor, a binocular camera, a steering engine and the like. The binocular camera is used for reading indoor pictures and recognizing the postures and skeletons of people, the binocular camera and the laser ranging sensor are used for measuring, calculating and correcting distances, and the contact process condition of people and objects is calculated under the three-dimensional space coordinate. The collision point and the force thereof are obtained based on the situation, the mechanical simulation is carried in to calculate the possible injury degree of the speculated personnel, the alarm is given to the injury degree, the condition is sent to the guardian and the hospital after the accident happens, the rescue personnel can conveniently carry appropriate equipment to rescue, and the efficiency is improved.
The invention is realized by at least one of the following technical schemes.
An indoor human body collision alarm method based on image processing and mechanical analysis comprises the following steps:
s1, collecting and detecting environmental image information by a collecting and detecting device, and identifying human bodies and objects in the images;
s2, the acquisition detection device carries out edge recognition on the object image and analyzes the object image to the limbs of the human body;
s3, based on edge recognition, dividing the pixel of the object image and the pixel of the human body image into length and height, comparing the height and width of the pixel points of the object and the human body to find a contactable area of the object and the human body, and judging a contacted trunk part according to the limbs of the human body;
s4, based on the division in the step S3, carrying out edge detection on the human body region of the object at the same height, and based on the pixel position, obtaining the distance between different trunk positions in the contact region and the object;
s5, acquiring the moving speed of the human body in real time, and judging the force applied when the human body collides with an object;
and S6, introducing the force and the collision point received when the human body collides with the object into a simulation formula to obtain the information of the damage condition.
Preferably, the force applied when the human body collides with the object is determined by the following formula:
V＝|D1-D2|/Δt
Ft＝m V1-m V2
where D1 denotes the distance between the human body and the object at time t1, D2 denotes the distance between the human body and the object at time t2, Δ t denotes the interval time between time t1 and time t2, V denotes the speed at which the human body moves within the interval time Δ t, m denotes the mass of the human body, V denotes the mass of the human body, and1、V2respectively, the human body velocity at time t1 and time t 2.
Preferably, the simulation in step S6 is implemented by mechanical simulation to obtain the relationship between the forces applied to different parts of the human body and the simulated displacement values of the human body, the positions of the human body identified by the openposition algorithm are substituted into the corresponding bones for simulation, the obtained values are bone displacement values, and the fracture degree is determined according to the bone displacement values to obtain the corresponding disease conditions, wherein the simulation process is as follows:
based on the CT image of the bone part of the existing human body, a three-dimensional model of the bone body is formed by reverse three-dimensional fitting, the density and the material of the model are set, based on ANSYS software, force application points are set at different parts, joints are set as fixed points, simulation is carried out, bone displacements under different forces and different parts are obtained, aiming at the same force application point, the displacement is taken as the Y axis, the force is taken as the X axis, and different forces correspond to different displacements, and a plurality of groups of data are recorded; and (5) performing data fitting through MATLAB to derive a corresponding formula.
The device of the indoor human body collision alarm method based on image processing and mechanical analysis comprises a base and a rotating part which is rotationally connected with the base; the base is connected with the rotating part through the rotating part; the rotating part is controlled by the controller, the rotating part is provided with a collection detection device used for collecting environmental information, and the collection detection device transmits collected information to the controller to obtain collision information of a human body and an object and collision conditions of the human body through the controller.
Preferably, the controller comprises a high-performance processor, the high-performance processor is used for controlling the steering engine and receiving and processing information collected by the collecting and detecting device, and the human body collision condition is judged through the human bone mechanics simulation model.
Preferably, the acquisition and detection device comprises a binocular camera, an infrared temperature measurement sensor, a laser distance measurement sensor and a rotating shaft; the infrared temperature measurement sensor measures the temperature of the human body in the detection range, and therefore whether the human body is detected is judged.
Preferably, the binocular camera searches for objects and human bodies in a range based on a model trained by TensorFlow, and performs distance measurement based on the binocular camera.
Preferably, the acquisition detection device compares the value read by the laser ranging sensor with the ranging value of the binocular camera, the brightness value of the image is set to be H, when the brightness value is larger than H, the ranging distance of the binocular camera is taken as the standard, the ranging value of the laser ranging sensor is taken as the auxiliary, and when the difference between the reading value of the laser ranging sensor and the value of the binocular camera accounts for no more than 10% of the ratio of the value of the binocular camera, the value read by the laser ranging sensor is taken as the distance value between a person and an object; when the rate exceeds 10%, taking the value of the binocular camera as a final result;
when the brightness is smaller than H, the distance measurement value of the laser distance measurement sensor is taken as the standard, and when the difference between the reading value of the laser distance measurement sensor and the reading value of the binocular camera does not exceed 10 percent of the reading ratio of the laser distance measurement sensor, the setting is correct, and correction is not carried out; when the ratio is exceeded, the reading of the laser ranging sensor is taken as the final result.
Preferably, the rotating member has an attitude sensor or a gyroscope built therein.
Preferably, the acquisition and detection device reads and identifies the distance between the person and the object, and calculates two-dimensional coordinates of the person and the object through a trigonometric function according to the rotation angle relative to the original point when the distance between the person and the object is detected, wherein the direction of the device facing the person and the object is taken as a Y axis, and the direction parallel to the person and the object is taken as an X axis; when the recognized people and objects have similar coordinates at the Y axis and the proportion of the difference value of the two values in the Y axis coordinate value of the people is less than or equal to 2 percent, the two values are considered to have the possibility of collision in the X axis direction, and if the difference value exceeds 2 percent, the recognition is abandoned;
when the rotating part rotates, the attitude sensor or the gyroscope feeds back the acquisition and detection device to detect personnel based on the rotation angle of the rotating part under the three-dimensional coordinate, and when the personnel are detected, the acquisition and detection device rotates around the Z axis relative to the original point by a corresponding angle, namely the rotation angle relative to the original point.
Compared with the prior art, the invention has the beneficial effects that:
the invention has the advantages that the mechanical simulation treatment is added, and the image recognition, the distance measurement and the temperature sensor are added simultaneously, so that the collision between the human body and the indoor article and the injury condition after the collision can be more accurately recognized. In addition, the invention carries out identification based on the data of the camera, and adds a distance measuring and temperature sensor to carry out data fusion, thereby improving the identification degree. Meanwhile, the 4G, SD card and the temperature sensor are added, so that the existing condition can be inferred by using information detected by the device and combining mechanical simulation under the condition that personnel do not wear portable equipment, the injury condition of the personnel is sent to the family members and the hospital through 4G, the existing patient condition can be more intuitively known by the hospital and the family members under the condition that the family members are not in the field, corresponding medicines and devices are prepared in advance, and the rescue efficiency is improved. The invention can be used for analyzing and deducing the injury condition of the household personnel, and is convenient for the caretaker and the hospital to know the condition of an illness. The early warning of wearing the device and timely analysis and the occurence of failure to the state of illness has been avoided, simultaneously through the people's bone mechanics simulation model analysis and the uploading of state of illness information after the personnel collision, medical personnel can in time carry necessary article and rescue, improve rescue efficiency.
Drawings
FIG. 1 is a perspective view of an indoor human body collision warning device based on image processing and mechanical analysis according to the present invention;
FIG. 2 is a side view of an indoor human body collision warning device based on image processing and mechanical analysis according to the present invention
FIG. 3 is a schematic diagram of the overall apparatus of the present invention;
in the figure, 1-binocular camera, 2-infrared temperature measuring sensor, 3-laser distance measuring sensor, 4-rotating shaft, 5-voice playing port, 6-base, 7-4G main antenna, 8-4G auxiliary antenna, 9-SIM card seat, 10-SD card seat and 11-power interface.
Detailed Description
The technical scheme of the invention is clearly and completely described in the following with reference to the accompanying drawings. In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", and the like indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience of description and simplification of description, but do not indicate or imply that the device or apparatus to which the description refers must have a specific orientation, be constructed and operated in a specific orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance.
In the description of the present invention, it is to be noted that, unless otherwise expressly specified or limited, the terms "loading," "attaching," "mounting," "arranging," and the like are to be construed broadly, such that mounting refers to securing a machine or equipment in a certain position according to a certain procedure or method. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
As shown in fig. 1, an indoor human body collision warning device based on image processing and mechanical analysis comprises a base 6 and a rotating part rotatably connected with the base 6, wherein a collecting and detecting device for collecting environmental information is mounted on the rotating part, and the collecting and detecting device comprises a binocular camera 1, an infrared temperature measuring sensor 2 and a laser distance measuring sensor 3; the base 6 is connected with the rotating part through the rotating shaft 4.
A controller and a 4G transmission module are arranged in the base 6; the controller is composed of a high-performance processor (Jetson NX high-performance processor) and a peripheral circuit thereof, the rotating part is controlled by the controller, and the controller controls the steering engine to enable the steering engine to drive the rotating shaft 4 to rotate and drive the rotating part to rotate together with the rotating shaft 4. The binocular camera 1, the infrared temperature measuring sensor 2 and the laser distance measuring sensor 3 rotate due to the rotation of the rotating component.
The 4G transmission module comprises a 4G main antenna 7, a 4G auxiliary antenna 8 and a 4G chip module, and uploads the injured condition to a hospital related system and a guardian mobile phone through a 4G network;
the base 6 is also provided with a voice playing port 5, a SIM card seat 9, an SD card seat 10, a power interface 11 and the like which are connected with the controller; the 4G main antenna 7 and the 4G auxiliary antenna 8 reduce signal attenuation, improve signal to noise ratio, avoid the problem that communication cannot be carried out in old school areas or under the condition of angle wall separation, and ensure that information data can be uploaded to hospitals and guardians' mobile phones in time.
The controller comprises a high-performance processor (Jetson NX high-performance processor), the high-performance processor is used for controlling the steering engine and receiving and processing information collected by the collecting and detecting device, and the human body collision condition is judged through the human bone mechanics simulation model.
An attitude sensor (gyroscope) is further arranged in the rotating part, the attitude sensor is placed right above the rotating shaft 4, the attitude sensor outputs three-dimensional space coordinates to a high-performance processor in the base 6, the coordinate origin is set when the plane where the lens of the binocular camera 1 is located in the rotating part is parallel to and closest to the plane where the voice playing port 5 of the base is located, the direction of the binocular camera 1 facing a person and an object is set as a Y axis, and the direction of the binocular camera 1 facing the person and the object is set as an X axis.
The method for the indoor human body collision warning device based on image processing and mechanical analysis comprises the following steps:
s1, collecting and detecting environmental image information by a collecting and detecting device, and identifying human bodies and objects in the images;
s2, the acquisition detection device transmits the acquired information to the controller, a high-performance processor in the controller performs Canny edge recognition on the object image, and the openposition algorithm is adopted to analyze the body limb of the human body;
s3, based on edge recognition, dividing the pixel of the object image and the pixel of the human body image into length and height, comparing the height and width of the pixel points of the object and the human body to find a contactable area of the object and the human body, and judging a contacted trunk part according to the limbs of the human body;
s4, based on the division in the step S3, carrying out edge detection on the human body region of the object at the same height, and based on the pixel position, obtaining the distance between different trunk positions in the contact region and the object;
s5, acquiring the moving speed of the human body in real time, and judging the force applied when the human body collides with an object;
and S6, introducing the force and the collision point received when the human body collides with the object into a simulation formula to obtain the information of the damage condition.
The high-performance processor analyzes the human skeleton based on an open-source human posture recognition project openposition and judges the position based on human skeleton analysis.
When the mobile phone is initially operated, the controller sends instructions to the binocular cameras 1, the infrared temperature measuring sensor 2, the laser distance measuring sensor 3 and the attitude sensor (gyroscope) to start to initialize coordinates, and the mobile phone is connected with the controller through Bluetooth. On the premise that a user does not manually adjust, the high-performance processor reads a three-dimensional coordinate fed back by the attitude sensor of the current rotating component to serve as an original point, controls the steering engine to rotate back and forth, and drives the rotating component to rotate left and right through the rotating shaft 4. At the moment, the high-performance processor reads information fed back by the attitude sensor, reads the maximum angles a and b of the rotating component rotating towards the left side and the right side during back-and-forth rotation, adds the angles of the left side and the right side and divides the angle by 2, subtracts the maximum angle a to obtain the angle of the actual origin rotating towards the two sides, initializes the attitude sensor again after the rotating component rotates, and resets the origin coordinate. When the user adjusts because of individual demand, after equipment relocates the initial point, can manually set for device equipment turned angle's size on APP to and the angle of each rotation of both sides, if not setting up, then acquiesce the left and right sides and respectively rotate 180.
After the device is initialized to operate, the device enters a working state, the high-performance processor controls the steering engine to rotate back and forth according to a set range under the feedback of the attitude sensor, and the rotating component is driven to scan back and forth through the rotating shaft 4.
The binocular camera 1 searches for objects and human bodies in the range based on a model trained by TensorFlow, and carries out distance measurement based on the binocular camera 1, compares the value read by the laser distance measuring sensor 3 with the distance measurement value of the binocular camera 1, and reads the brightness value based on the image, wherein the value is higher and brighter. The full brightness of 50% is the limit, sets for H, when the luminance value is greater than H, uses binocular camera 1 range finding distance as the standard, and the 3 range finding numerical value of laser range finding sensor is supplementary, and when the numerical difference of 3 readings of laser range finding sensor and 1 numerical value of binocular camera accounts for binocular camera 1 numerical value proportion and does not exceed 10%, sets for correctly, does not revise. And when the proportion is exceeded, taking the value of the binocular camera 1 as a final result. When the brightness is smaller than H, the distance measurement value of the laser distance measurement sensor 3 is taken as the standard, when the difference between the reading of the laser distance measurement sensor 3 and the value of the binocular camera 1 accounts for no more than 10% of the reading proportion of the laser distance measurement sensor 3, the value is set as a correct value, the value is a distance value, and correction is not performed (the correct value refers to that the data is reasonable distance data, and data read by other sensors is not required to be referred as the distance value). When the ratio is exceeded, the laser ranging sensor 3 is read as the final result.
The acquisition and detection device reads and identifies the distance between a person and an object, and according to the rotation angle of the acquisition and detection device relative to an original point when the distance between the person and the object is detected (the original point refers to the position of a gyroscope of the acquisition and detection device of the indoor human body collision alarm method based on image processing and mechanical analysis when the distance between the person and the object is not rotated, the relative rotation angle refers to the position of the gyroscope built in the acquisition and detection device, when the acquisition and detection device rotates, the gyroscope feeds back the rotation angle based on three-dimensional coordinates, the acquisition and detection device detects the person based on the rotation of a rotating part, when the person is detected, the gyroscope rotates around a Z axis relative to the original point by a certain angle, the angle is the rotation angle relative to the original point), two-dimensional coordinates of the person and the object are obtained through trigonometric function calculation, the plane where the voice playing port 5 is located is the front of the indoor human body collision alarm device based on image processing and mechanical analysis, the plane that 4G main antenna 7 and base 6 are connected is the back, and the axle center of axis of rotation 4 when the one side of the collection detection device contact that 1 camera lens of binocular camera and 1 place were on a parallel with the plane of pronunciation broadcast mouth 5 place is the original point, with the front that is on a parallel with an indoor human body collision alarm device based on image processing and mechanics analysis be the X axle, and the left side is the negative X axle, and the right side is positive X axle, with the positive place ahead in the front of an indoor human body collision alarm device based on image processing and mechanics analysis be positive Y axle. And taking the axial direction parallel to the rotating shaft 4 upwards as a positive Z axis, and when the high-performance processor reads the picture of the binocular camera 1 and recognizes that the coordinates of people and objects at the Y axis are close, and the proportion of the difference value of the two values in the Y axis coordinate value of the people is less than or equal to 2%, the two values are considered to have collision possibility in the X axis direction, and at the moment, the device starts to perform the next calculation. If the ratio exceeds 2%, the recognition is discarded. The infrared temperature measuring sensor 2 measures the temperature of the human body within the range, so as to judge whether the human body is a human body.
When the acquisition and detection device reads and identifies the distance between a person and an object and feeds the distance back to a high-performance processor in the controller, when the high-performance processor identifies that the person and the object are likely to collide, the high-performance processor identifies the edge of the object based on the previously identified person and object, performs openpos solution on the human body, reads the pixel coordinate of the edge line of the object and the pixel coordinate of the human body skeleton region in an image, subtracts the two coordinates, obtains the minimum value which is the region likely to collide with the object under the motion of the human body, takes the pixel coordinate of the human body in the minimum value into the openpos region, identifies the human body bone part of the pixel coordinate, calls a corresponding human bone mechanics simulation model, inversely fits the human body bone mechanics simulation model into the three-dimensional bone formation model based on the CT image of the human body part, sets the density and the material of the model, sets a force application point at different parts based on ANSYS software, set for the fixed point with joint department, carry out the simulation to this can obtain the displacement of bone atress under different power, different positions, to same stress point, use the displacement as the Y axle, power is the X axle, and different power correspond different displacements, based on simulation many times like this, record multiunit data, carry out data fitting through MATLAB based on this multiunit data, derive corresponding formula, this formula is:
Z＝A*x1^n+B*x2^n1+C，
because different parts have different bones and different muscles and soft tissues, the coefficients A, B, C are different, x1 is displacement, x2 is a force value, C is a buffer value from the muscle at the corresponding point, n is the power number of the displacement, n1 is the power number of the force value, and Z is the displacement value. The image data in the visual angle of the camera is a two-dimensional plane, so that the identified and judged impact area is an impact point, in the simulation, 2mm of Z-axis interval on the sagittal plane of a bone part is taken as a simulation point, different force values are applied to each simulation point for simulation, displacement simulation numerical values of different force values of the same point are collected, a formula is fitted, each formula corresponds to one impact point, and each impact point is arranged from head to tail according to the numerical sequence.
In the model, the bone displacement refers to the angle of the bone before and after applying force on a certain point, and the bone is bent after being stressed from the whole view. Bone displacement is a numerical value and in fact the displacement is caused by bending of the bone, which to a certain extent will fracture, with the bone displacement at the fracture being the critical point. Selecting different formulas according to the impact points, substituting the force into the formulas, judging the fracture when the obtained numerical value is greater than the critical point, and if the obtained numerical value is not greater than the critical point, dividing the pain degree according to the actually investigated relation between the pain sense and the bending of the bone body. At this time, the acquisition and detection device continuously reads the pixel distance S between the human body and the object in the picture. Based on the interval time t, according to the formula: obtaining human body speeds at different moments based on the momentum theorem Ft mv1-mv2 and a preset weight value, obtaining the force applied to a contact point when the human body is contacted (the distance between two coordinates of the contact point of the human body and the object is 0), bringing the force value, time and bone area into a mechanical simulation model for calculation, obtaining a bone deviation value, a pressure value and the reaction of surrounding bones and muscle tissues, comparing the pressure with the pain feeling value and comparing the femoral deviation with the fracture possibility, and obtaining the possible external expression condition and possible subsequent state information when the human body is injured. The method is defined based on the common medical science each time (for example, for leg bone fracture, the appearance known in the common medical science is that a human body cannot act, or the bone healing is slow along with that a fracture part is not timely treated, for example, the bone is simply collided but not fractured, and bruise can be formed at an impact point on the common medical level. The guardian can use cell-phone APP to control transmission information, carries out the numeralization show to the state of an illness on the APP, and the user can set for when the value reaches when many, and the device is automatic to be reported and to seek help to the hospital. When the human body and the object are within the preset alarm distance, the device gives an alarm through the voice playing port 5 to remind people to pay attention to the front object.
When a person falls, it is graphically: the human body is used as the center, namely the point of falling over the forefoot is used as the origin, and the human body rotates around the point by a certain angle until the human body contacts an object or the contact point and the origin are on the same horizontal line. Because the limbs may change and are not straight and straight in the falling process of the human body, the collecting and detecting device identifies the human body, calculates the human body at each skeleton position based on opencast, identifies the edges of each skeleton to the periphery based on each skeleton, identifies the limbs of the human body, identifies the human body as falling when the limbs have large angle change relative to the front position, obtains the pixel length from the head to the foot of the human body in different pictures based on the pixel point coordinates of the human body in the pictures, takes the pixel length as the radius R, obtains the angular velocity according to w-theta/t based on the interval time t, obtains the linear velocity based on V-wR, considers that the linear velocity is the velocity vertical to the contact point at the moment because the t time is short, obtains the force applied to the corresponding position when the human body falls based on the momentum theorem, and brings the force magnitude, the force value, the time and the bone region into the mechanical simulation model for calculation, and obtaining a bone deviation value, a pressure value, and the reaction of surrounding bones and muscle tissues, and obtaining the external manifestation and the subsequent possible state information which may appear when the person is injured according to the comparison of the pressure and the value of pain feeling and the comparison of the femoral deviation and the fracture possibility. And feeds the information back to the guardian and the medical staff for subsequent processing.
When some diseases occur, the disease is expressed as frequent, regular or irregular physical behaviors of the person. Based on a model after deep learning of human body behaviors (the model structure is a case model stored by a device for recording the behaviors of a patient with frequent disease-induced limb behaviors when a disease occurs, the human body action swing position is identified frame by video based on OpenCV, Openpos and TensorFlow, and then the frame by frame picture position connection is carried out to form a swing amplitude change value of a corresponding part under the disease occurrence, an image is cut in a change area to form a swing model of the corresponding part, the swing of different parts is linked with the swing conditions of other parts under corresponding time to form a human body model again The real-time human behavior video has corresponding matching degree in the model, and when the matching degree reaches 90%, the condition possibly occurring in the personnel is deduced. And generates corresponding information based on the guardian and the medical personnel. Meanwhile, an alarm is sent out through the voice playing port 5 to remind surrounding personnel that a person is in a disease condition nearby.
The human bone mechanics simulation model is simulated based on mimics and ANSYS to obtain the diaphysis offset, extrusion, abrasion and the like of human bones at different positions under different stress conditions, meanwhile, the coefficient values of the buffering influence of skin and muscle on the force are led in, the obtained continuous values are brought into fitting, the force analysis formulas of different bone parts under different environments are reversely deduced, analyzing the bone position and the stress condition according to openposition, calling stress analysis formula models of different bone positions, so as to obtain the required data of bone body displacement, etc., and further deduce which kind of possible disease condition is based on the data and the set disease condition grade, and to what extent, such as by a prior medical rating, such as fracture and comminuted fracture, the medical rating as such, on the human bone mechanics simulation model level, comminuted fracture is fracture in which bone is subjected to great force at multiple points or a single point, resulting in bone fracture.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (10)
1. An indoor human body collision alarm method based on image processing and mechanical analysis is characterized by comprising the following steps:
s1, collecting and detecting environmental image information by a collecting and detecting device, and identifying human bodies and objects in the images;
s2, the acquisition detection device carries out edge recognition on the object image and analyzes the object image to the limbs of the human body;
s3, based on edge recognition, dividing the pixel of the object image and the pixel of the human body image into length and height, comparing the height and width of the pixel points of the object and the human body to find a contactable area of the object and the human body, and judging a contacted trunk part according to the limbs of the human body;
s4, based on the division in the step S3, carrying out edge detection on the human body region of the object at the same height, and based on the pixel position, obtaining the distance between different trunk positions in the contact region and the object;
s5, acquiring the moving speed of the human body in real time, and judging the force applied when the human body collides with an object;
and S6, introducing the force and the collision point received when the human body collides with the object into a simulation formula to obtain the information of the damage condition.
2. The indoor human body collision warning method based on image processing and mechanical analysis according to claim 1, wherein the force applied when the human body collides with the object is determined by the following formula:
V＝|D1-D2|/Δt
Ft＝m V1-m V2
where D1 denotes the distance between the human body and the object at time t1, D2 denotes the distance between the human body and the object at time t2, Δ t denotes the interval time between time t1 and time t2, V denotes the speed at which the human body moves within the interval time Δ t, m denotes the mass of the human body, V denotes the mass of the human body, and1、V2respectively, the human body velocity at time t1 and time t 2.
3. The indoor human body collision warning method based on image processing and mechanical analysis as claimed in claim 1, wherein the simulation of step S6 is to obtain the relationship between the forces applied to different parts of the human body and the human body simulation displacement values through mechanical simulation, the applied forces are substituted into the corresponding bones according to the human body positions identified by openposition algorithm for simulation, the obtained values are bone displacement values, and the fracture degree is determined according to the bone displacement values to obtain the corresponding illness states, the simulation process is as follows:
based on the CT image of the bone part of the existing human body, a three-dimensional model of the bone body is formed by reverse three-dimensional fitting, the density and the material of the model are set, based on ANSYS software, force application points are set at different parts, joints are set as fixed points, simulation is carried out, bone displacements under different forces and different parts are obtained, aiming at the same force application point, the displacement is taken as the Y axis, the force is taken as the X axis, and different forces correspond to different displacements, and a plurality of groups of data are recorded; and (5) performing data fitting through MATLAB to derive a corresponding formula.
4. The device for alarming human body collision in the room based on image processing and mechanical analysis as claimed in claim 1, which is characterized by comprising a base (6), a rotating part rotatably connected with the base (6); the base (6) is connected with the rotating part through the rotating part; the rotating part is controlled by the controller, the rotating part is provided with a collection detection device used for collecting environmental information, and the collection detection device transmits collected information to the controller to obtain collision information of a human body and an object and collision conditions of the human body through the controller.
5. The device of claim 4, wherein the controller comprises a high-performance processor therein, the high-performance processor is used for controlling the steering engine and receiving, processing and collecting information collected by the detection device, and judging the human body collision condition through a human bone mechanics simulation model.
6. The device for alarming indoor human body collision based on image processing and mechanical analysis as claimed in claim 4, wherein the collecting and detecting device comprises a binocular camera (1), an infrared temperature measuring sensor (2), a laser distance measuring sensor (3) and a rotating shaft (4); the infrared temperature measurement sensor (2) measures the temperature of the human body in the detection range, so as to judge whether the human body is detected.
7. The apparatus for the indoor human body collision warning method based on image processing and mechanical analysis according to claim 6, wherein the binocular camera (1) searches for an object, a human body within a range based on a TensorFlow trained model, and performs ranging based on the binocular camera (1).
8. The device for alarming indoor human body collision based on image processing and mechanical analysis as claimed in claim 6, wherein the collecting and detecting device compares the value read by the laser ranging sensor (3) with the ranging value of the binocular camera (1), sets the brightness value of the image as H, when the brightness value is greater than H, the ranging distance of the binocular camera (1) is taken as the standard, the ranging value of the laser ranging sensor (3) is taken as the auxiliary, when the ratio of the reading of the laser ranging sensor (3) and the value of the binocular camera (1) to the value of the binocular camera (1) is not more than 10%, the value read by the laser ranging sensor (3) is taken as the distance value between the human and the object; when the rate of the image data exceeds 10%, taking the value of the binocular camera (1) as a final result;
when the brightness is smaller than H, the distance measurement value of the laser distance measurement sensor (3) is taken as the standard, and when the difference between the reading value of the laser distance measurement sensor (3) and the reading value of the binocular camera (1) is less than 10% of the reading proportion of the laser distance measurement sensor (3), the setting is correct, and correction is not carried out; when the proportion is exceeded, the reading of the laser ranging sensor (3) is taken as the final result.
9. The apparatus for an indoor human body collision warning method based on image processing and mechanical analysis according to claim 6, wherein the rotating member is built-in with an attitude sensor or a gyroscope.
10. The device of the indoor human body collision warning method based on image processing and mechanical analysis as claimed in claim 9, wherein the acquisition detection device reads and identifies the distance between the person and the object, and obtains the two-dimensional coordinates of the person and the object through trigonometric function calculation according to the rotation angle relative to the origin when the distance between the person and the object is detected, and the direction of the device facing the person and the object is taken as the Y axis, and the direction parallel to the person and the object is taken as the X axis; when the recognized people and objects have similar coordinates at the Y axis and the proportion of the difference value of the two values in the Y axis coordinate value of the people is less than or equal to 2 percent, the two values are considered to have the possibility of collision in the X axis direction, and if the difference value exceeds 2 percent, the recognition is abandoned;
when the rotating part rotates, the attitude sensor or the gyroscope feeds back the acquisition and detection device to detect personnel based on the rotation angle of the rotating part under the three-dimensional coordinate, and when the personnel are detected, the acquisition and detection device rotates around the Z axis relative to the original point by a corresponding angle, namely the rotation angle relative to the original point.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111052459.2A CN113903052A (en) | 2021-09-08 | 2021-09-08 | Indoor human body collision alarm method and device based on image processing and mechanical analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111052459.2A CN113903052A (en) | 2021-09-08 | 2021-09-08 | Indoor human body collision alarm method and device based on image processing and mechanical analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113903052A true CN113903052A (en) | 2022-01-07 |
Family
ID=79188856
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111052459.2A Pending CN113903052A (en) | 2021-09-08 | 2021-09-08 | Indoor human body collision alarm method and device based on image processing and mechanical analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113903052A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115886683A (en) * | 2023-01-06 | 2023-04-04 | 珠海视新医用科技有限公司 | Endoscope anti-collision alarm method and device |
| CN116704553A (en) * | 2023-06-13 | 2023-09-05 | 长江大学 | Human body characteristic identification auxiliary system based on computer vision technology |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108109197A (en) * | 2017-12-15 | 2018-06-01 | 北京中睿华信信息技术有限公司 | A kind of image procossing modeling method |
| CN109620184A (en) * | 2019-01-29 | 2019-04-16 | 上海海事大学 | Mobile phone-wearable device integral type human body burst injury real-time monitoring alarming method |
| CN112069933A (en) * | 2020-08-21 | 2020-12-11 | 董秀园 | Skeletal muscle stress estimation method based on posture recognition and human body biomechanics |
| CN112200126A (en) * | 2020-10-26 | 2021-01-08 | 上海盛奕数字科技有限公司 | Method for identifying limb shielding gesture based on artificial intelligence running |
-
2021
- 2021-09-08 CN CN202111052459.2A patent/CN113903052A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108109197A (en) * | 2017-12-15 | 2018-06-01 | 北京中睿华信信息技术有限公司 | A kind of image procossing modeling method |
| CN109620184A (en) * | 2019-01-29 | 2019-04-16 | 上海海事大学 | Mobile phone-wearable device integral type human body burst injury real-time monitoring alarming method |
| CN112069933A (en) * | 2020-08-21 | 2020-12-11 | 董秀园 | Skeletal muscle stress estimation method based on posture recognition and human body biomechanics |
| CN112200126A (en) * | 2020-10-26 | 2021-01-08 | 上海盛奕数字科技有限公司 | Method for identifying limb shielding gesture based on artificial intelligence running |
Non-Patent Citations (1)
| Title |
|---|
| 柴象海等: "面向受害人体型特征的交通事故再现计算方法", 《振动与冲击》, no. 04, 25 April 2010 (2010-04-25) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115886683A (en) * | 2023-01-06 | 2023-04-04 | 珠海视新医用科技有限公司 | Endoscope anti-collision alarm method and device |
| CN116704553A (en) * | 2023-06-13 | 2023-09-05 | 长江大学 | Human body characteristic identification auxiliary system based on computer vision technology |
| CN116704553B (en) * | 2023-06-13 | 2024-01-26 | 长江大学 | Human body characteristic identification auxiliary system based on computer vision technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8214082B2 (en) | Nursing system | |
| US10506952B2 (en) | Motion monitor | |
| US20220331028A1 (en) | System for Capturing Movement Patterns and/or Vital Signs of a Person | |
| CN113903052A (en) | Indoor human body collision alarm method and device based on image processing and mechanical analysis | |
| JP3570163B2 (en) | Method and apparatus and system for recognizing actions and actions | |
| JP6466139B2 (en) | Robot measuring instrument that measures human movement | |
| CN105364915B (en) | Intelligent home service robot based on three-dimensional machine vision | |
| WO2013058985A1 (en) | Method and apparatus for detecting deterioration of health status | |
| JP2016080671A5 (en) | ||
| CN107788990A (en) | A kind of wearable falling detection device and fall detection system | |
| Nizam et al. | A study on human fall detection systems: Daily activity classification and sensing techniques | |
| CN106214163B (en) | Recovered artifical psychological counseling device of low limbs deformity correction postoperative | |
| US20210000677A1 (en) | State estimation program, trained model, rehabilitation support system, learning apparatus, and state estimation method | |
| CN113364843A (en) | Indoor timely rescue system who meets accident alone of old person | |
| CN211155811U (en) | Monitoring bracelet equipment containing gesture recognition | |
| Patsadu et al. | Survey of smart technologies for fall motion detection: Techniques, algorithms and tools | |
| KR102153781B1 (en) | Method for measuring physical exercise ability of rehabilitation patient using motion recognition band | |
| CN113041092A (en) | Remote rehabilitation training system and method based on multi-sensor information fusion | |
| CN109199712A (en) | A kind of evaluation and test of intelligent hand motor function and recovery training wheel chair | |
| CN110517750A (en) | A kind of more human action method for catching of fusion WIFI positioning and inertia sensing | |
| Niu et al. | A survey on IMU-and-vision-based human pose estimation for rehabilitation | |
| Almazaydeh et al. | A panoramic study of fall detection technologies | |
| Crha et al. | Navigation of robotic platform for gait disorders monitoring | |
| US20220167880A1 (en) | Patient position monitoring methods and systems | |
| Nistane et al. | An IoT based rectification of posture using biofeedback |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |