A disinfection door spray precision positioning and identification method based on machine vision

Abstract

The application discloses a disinfection door spray precise positioning and identification method based on machine vision, image acquisition of passing crowds is carried out through a high-speed camera to obtain high-quality images; a wavelet threshold denoising method is used to carry out denoising treatment on initial images, and then a weighted average method is used to carry out gray-scale processing; then, the images are subjected to binarization processing; a sobel operator is used to carry out image edge detection and segmentation, and a double-edge tracking algorithm is used to concatenate target edges; then, it is judged whether the images meet spray requirements; if yes, an optimization algorithm is used to calculate the spray time, spray pressure and spray angle of each spray head of N ultrasonic spray nozzles installed on the disinfection door; then, spray disinfection operation is carried out; finally, image processing is used to detect the spray concentration, the nozzle input is adjusted, and the working efficiency and precision of the disinfection door are ensured, so that strong support is provided for inhibiting virus transmission.

Description

TECHNICAL FIELD

[0001] The present application mainly relates to the disinfection technical field, more particularly, to a disinfection door spray precise positioning and identification method based on machine vision. BACKGROUND

[0002] The current commonly used virus disinfection method is to manually hold a sprayer to spray disinfectant. Such a virus disinfection method is not only time-consuming and laborious, but also poses a certain infection risk to the personnel involved in virus disinfection. The disinfection door currently used on the market mainly uses infrared sensing, that is, when someone is detected to pass through, the disinfection door sprays a certain amount of spray to disinfect. However, such a disinfection door cannot spray disinfectant according to the body shape of the person, and there are problems of incomplete disinfection and waste of disinfectant, and excessive disinfectant also poses a certain harm to human health. SUMMARY

[0003] In view of the problems of low efficiency, high drug consumption, and infection risk in the current commonly used virus disinfection method, the purpose of the present application is to provide a disinfection door spray precise positioning and identification method based on machine vision.

[0004] In order to solve the above technical problems, the present application provides a disinfection door spray precise positioning and identification method based on machine vision, which specifically comprises the following steps:

[0005] Step one: image acquisition of the human body to be passed through the disinfection door to obtain an initial image;

[0006] Step two: wavelet threshold denoising method is used to denoise the initial image obtained by collection, and then the denoised image is subjected to gray scale processing by weighted average method;

[0007] Step three: the gray scale processed image is subjected to local adaptive binaryzation;

[0008] Step four: the binaryzation processed image is subjected to image edge detection and segmentation by Sobel operator to extract the human body edge contour;

[0009] Step five: the human body edge image is obtained by edge detection of the human body edge contour, the target edge of the human body edge image is concatenated by double edge tracking algorithm, and the center of the target point is calculated;

[0010] Step six: it is judged whether the clarity of the concatenated human body edge image meets the spray requirement, if not, it is returned to step one to start image collection again;

[0011] After judging that the human body continuous edge image meets the liquid spraying requirement, the spray time, spray pressure and spray angle of the spray head are calculated according to the human body edge image; and the disinfectant is accurately controlled to be applied.

[0012] As a preferred embodiment of the present application, the wavelet threshold denoising method in step two has the following process: 1) decomposition process: selecting a wavelet to perform n-layer wavelet decomposition on the image signal; 2) threshold processing process: performing threshold processing on the coefficients of each layer of the decomposition to obtain wavelet coefficients; and 3) reconstruction process: performing wavelet reconstruction according to the obtained wavelet coefficients to obtain the denoised image signal.

[0013] As a preferred embodiment of the present application, the denoised image is further evaluated by a denoising quality measurement evaluation method after the denoising processing in step two, and the denoising quality measurement evaluation method uses a peak signal-to-noise ratio, which is usually defined as:

[0014]

[0015] MAX I : maximum value of the color of the image point; and MSE: mean square error.

[0016] The mean square error (MSE) is defined as follows:

[0017]

[0018] H: height of the image; W: width of the image; X(i,j): pixel point of the initial image; and Y(i,j): pixel point of the denoised image.

[0019] As a preferred embodiment of the present application, the gray scale processing formula in step two is:

[0020] Gray(i,j) = 0.299 * R(i,j) + 0.587 * G(i,j) + 0.114 * B(i,j)

[0021] Gray(i,j): gray scale value; R(i,j): red component of the denoised image; G(i,j): green component of the denoised image; and B(i,j): blue component of the denoised image.

[0022] As a preferred embodiment of the present application, step three specifically comprises: performing threshold calculation for each pixel point in the gray scale image, and the threshold of the image is determined by the local gray scale characteristics and the pixel gray scale value of all pixel points in the image; and setting the pixel points less than the threshold in the image to one pixel value, and setting the pixel points greater than the threshold in the image to another pixel value.

[0023] As the preferred of the present application, Sobel operator is used for edge detection in the step four, A represents the binary image pixel point, Gx and Gy represent the image gray value detected by horizontal and vertical edge respectively.

[0024]

[0025] Therefore, the gray scale of this point: |G| = |G x + |G y |

[0026] If the gradient G is greater than a certain threshold, the point (x, y) is considered as an edge point.

[0027] As the preferred of the present application, the calculation of the edge center in the step five, which mainly includes finding the upper and lower boundaries of the target. The calculation form is as follows:

[0028]

[0029]

[0030]

[0031]

[0032] Further, the center coordinates can be obtained:

[0033]

[0034] y upper : the highest point of the image in y coordinate; y lower : the lowest point of the image in y coordinate; x left : the leftmost point of the image in x coordinate; x right : the rightmost point of the image in x coordinate; W: pixel width; H: pixel height; I(x, y): brightness function of the point.

[0035] As the preferred of the present application, the quasi-position identification method can also collect the spray photo sprayed by the spray head, and the edge of the mist droplet is detected and recognized through image processing to identify the spray concentration, so as to perform negative feedback adjustment on the working parameters of the spray head, and ensure the working efficiency and working accuracy of the disinfection door.

[0036] The present application also provides a disinfection door based on the disinfection door spray precision positioning identification method based on machine vision, which comprises a disinfection door body, an image acquisition device arranged at the top position of the disinfection door body, a spray precision positioning identification device and a liquid spraying device for spraying disinfectant.

[0037] The image acquisition device includes a high-speed camera, which is used to acquire the image of a human body about to pass through the disinfection door and acquire a spray picture sprayed by the liquid spraying device and transmit to the spray precise positioning and identification device; the spray precise positioning and identification device is provided with a disinfection door spray precise positioning and identification method based on machine vision, which can analyze and process the crowd image and the spray picture transmitted by the image acquisition device to adjust the liquid spraying device; the liquid spraying device includes a plurality of ultrasonic spray nozzles mounted on the disinfection door body, and the ultrasonic spray nozzles are used to spray disinfectant.

[0038] Compared with the prior art, the present application has the following advantages:

[0039] 1. The present application can precisely position the disinfectant spray through the edge recognition method, ensure the comprehensive coverage of disinfectant, and improve the virus killing effect.

[0040] 2. The present application can calculate the spray time, spray pressure and spray angle of the spray head according to the human body edge image through the precise positioning and identification method; precise control of the application of disinfectant can effectively reduce the use of disinfectant, save a large amount of medical resources, and avoid the physiological damage to the human body caused by excessive disinfectant dose of traditional disinfection door, thereby protecting the health of the human body.

[0041] 3. The present application can identify the crowd passing through the disinfection door through machine vision technology, and kill viruses without manually holding a sprayer to spray disinfectant, thereby effectively controlling the spread of viruses. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is the overall flowchart of the disinfection door spray precise positioning and identification method based on machine vision of the present application.

[0043] Figure 2 is a disinfection door diagram in one specific embodiment of the disinfection door spray precise positioning and identification method based on machine vision of the present application. DETAILED DESCRIPTION

[0044] The present application will be further described below in conjunction with the drawings.

[0045] It should be noted that in the present application, unless otherwise explicitly specified and limited, the terms "connection", etc. should be understood in a broad sense, for example, "connection" can be fixed connection, or detachable connection, or integral; can be directly connected, or indirectly connected through an intermediate medium; can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise explicitly limited. For ordinary skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0046] In the present embodiment, as shown inFigure 1 and Figure 2 A machine vision-based disinfection door spray precision positioning and identification method is shown as follows:

[0047] Step one: first, install an image acquisition device on the top of the disinfection door, which acquires images of people about to pass through the disinfection door through a high-speed camera, obtaining high-definition and high-quality initial images;

[0048] Step two: then a series of processing is performed on the acquired images, and a wavelet threshold denoising method is used to denoise the acquired initial images: first, select a wavelet to perform n-layer wavelet decomposition on the image signal; then, perform threshold processing on the decomposed coefficients of each layer to obtain estimated wavelet coefficients; finally, perform wavelet reconstruction according to the wavelet coefficients to obtain the denoised image signal. After denoising, the denoising quality measurement evaluation method is used for evaluation, and the peak signal-to-noise ratio is used as the denoising quality measurement evaluation method, which is usually defined as:

[0049]

[0050] MAX I : maximum value of image point color; MSE: mean square error.

[0051] And the mean square error (MSE) is defined as follows:

[0052]

[0053] H: height of the image; W: width of the image; X(i,j): initial image pixel value; Y(i,j): denoised image pixel value.

[0054] Then, the denoised image is processed by the weighted average method to make it recognizable by the computer; and the gray processing formula is:

[0055] Gray(i,j)=0.299*R(i,j)+0.587*G(i,j)+0.114*B(i,j)

[0056] Gray(i,j): gray value; R(i,j): red component of the denoised image; G(i,j): green component of the denoised image; B(i,j): blue component of the denoised image.

[0057] Step three: the image is binarized locally and adaptively, and the threshold value of each pixel in the gray image is calculated point by point. The threshold value is determined by the local gray characteristics of the points in the neighborhood of the pixel and the pixel gray value. The pixel points less than the threshold value in the image are set to one pixel value, and the pixel points greater than the threshold value are set to another pixel value. At this time, the pixel gray value in the image is only 0 or 255;

[0058] Step four: the edge part of the image concentrates most of the information of the image, and is also the recognition target. Therefore, the Sobel operator is used to detect and segment the edges of the image. A represents the pixel points of the binarized image, Gx and Gy represent the gray values of the image detected by the horizontal and vertical edges respectively.

[0059] Then:

[0060]

[0061] Therefore, the gray value of this point is: |G| = |G x |+|G y |; if the gradient G is greater than a certain threshold value, it is considered that the point (x, y) is an edge point.

[0062] The human body image edge is identified and the human body edge contour is extracted.

[0063] Step five: the separated human body edge image is obtained by edge detection, the target edge is connected by double edge tracking algorithm, and the center of the target point is calculated to obtain the human body edge contour graph. The main content is to find the upper and lower boundaries of the target, and the calculation form is:

[0064]

[0065]

[0066]

[0067]

[0068] Further, the center coordinates can be obtained:

[0069]

[0070] y upper : the highest point of the image in the y coordinate; y lower : the lowest point of the image in the y coordinate; x left : the leftmost point of the image in the x coordinate; x right : the rightmost point of the image in the x coordinate; W: pixel width; H: pixel height; I(x, y): brightness function of the point.

[0071] Step six: judge whether the human body edge image obtained by concatenation meets the liquid spraying requirement, if the human body contour obtained by recognition is too fuzzy, return to step one, start image collection and image processing again, and accurately extract human body contour lines;

[0072] Step seven: after obtaining the continuous human body edge contour, according to the contour characteristics, the spraying time, spraying pressure and spraying angle of each nozzle of the N ultrasonic spray nozzles installed on the disinfection door are calculated by using an optimization algorithm, and the nozzles are adjusted, and the spraying is prepared;

[0073] Step eight: the N ultrasonic spray nozzles installed on the disinfection door can accurately control the application of disinfectant according to the calculation results, complete the disinfection target while reducing the amount of disinfectant, protect human health and save resources.

[0074] Step nine: the high-speed camera can take pictures of the spray of the ultrasonic spray nozzle, identify the spray concentration through image processing, and then adjust the nozzle working parameters of the ultrasonic nozzle, so as to ensure the working efficiency and accuracy of the disinfection door.

[0075] As shown in Figure 2 In order to more clearly understand the method of the application, a disinfection door adopting the disinfection door spray accurate positioning and recognition method based on machine vision is also provided, which comprises a disinfection door body, an image acquisition device arranged at the top of the disinfection door body, a spray accurate positioning and recognition device, and a liquid spraying device for spraying disinfectant.

[0076] The image acquisition device comprises a high-speed camera, which is used to acquire the image of the human body about to pass through the disinfection door and the spray picture sprayed by the liquid spraying device and transmit it to the spray accurate positioning and recognition device; the spray accurate positioning and recognition device can analyze and process the crowd image and the spray picture transmitted by the image acquisition device to adjust the liquid spraying device; the liquid spraying device comprises a plurality of ultrasonic spray nozzles installed on the disinfection door body, and the ultrasonic spray nozzles are used to spray disinfectant.

[0077] Its working process is as follows: when the human body is at a position 2 meters away from the disinfection door, a plurality of images are collected by using a high-speed camera, the collected images are screened to ensure that a front, unobstructed, high-brightness high-definition human body image is obtained, and the image resolution is 3024*4032. Then the initial image collected is denoised by using a wavelet threshold denoising method; then the denoised image is grayed by using a weighted average method; then the grayed image is locally self-adaptive binarized; then the binarized image is subjected to image edge detection and segmentation by using a sobel operator; finally, a discontinuous human body edge image is obtained by edge detection, the target edge is concatenated by a double-edge tracking algorithm, and the center of the target point is calculated to obtain a human body edge contour graph.

[0078] According to the contour characteristics of the obtained human body edge contour graph, an optimization algorithm is used to calculate the spraying time, spraying pressure and spraying angle of each nozzle of the 10 ultrasonic spray nozzles installed on the disinfection door, and the nozzles are adjusted for spraying. The ultrasonic atomizing nozzles installed on the disinfection door are mainly distributed as two on the top and two on the bottom with a spacing of 20 cm, and three on the left and three on the right with a spacing of 40 cm. The atomizing nozzles are connected to the disinfection door in a rotatable manner, and the atomizing nozzles can be rotated by 60° around the axis as the center, which is conducive to precise control of the spraying angle. The total spraying amount of each nozzle is set based on the recognized human body edge contour to set the spraying time and spraying pressure of each nozzle, and the total spraying amount of the nozzle is controlled in the range of 3-6 ml / person to ensure moderate use and protect human health while ensuring virus disinfection effect. At the same time, the high-speed camera also collects pictures when the nozzle sprays, analyzes the concentration and range of the spray, and thus adjusts the nozzle working parameters of the ultrasonic nozzle in real time to ensure the working efficiency and accuracy of the disinfection door.

[0079] In summary, the disinfection door spraying precise positioning and recognition method based on machine vision provided by the present application reduces the use of disinfectant while ensuring the virus disinfection effect and protecting human health, and can adjust the nozzle working parameters of the ultrasonic nozzle in real time to ensure the working efficiency and accuracy of the disinfection door. It provides strong support for inhibiting virus transmission

[0080] Finally, it should be noted that: the above only describes the preferred embodiments of the present application and is not intended to limit the present application, although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent replacements for some technical features. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A machine vision-based disinfection door spray precision positioning and identification method, characterized in that, The method comprises the following steps: Step 1: image acquisition of a human body passing through a disinfection door to obtain an initial image; Step 2: wavelet threshold denoising method is used to denoise the initial image, and then weighted average method is used to gray the denoised image; Step 3: local adaptive binaryzation of the grayed image; Step 4: Sobel operator is used to detect and segment the edge of the binaryzed image to extract the edge contour of the human body; Step 5: discontinuous human body edge image is obtained by edge detection of the edge contour of the human body, the target edge of the human body edge image is connected by double edge tracking algorithm, and the center of the target point is calculated; Step 6: whether the connected human body edge image meets the liquid spraying requirement is judged, if not, the image acquisition is restarted; After the continuous human body edge image is obtained, the spraying time, spraying pressure and spraying angle of the nozzle are calculated according to the human body edge image, and the disinfectant is accurately controlled.

2. The method according to claim 1, wherein, The wavelet threshold denoising method in step 2 comprises the following steps: 1) decomposition process: selecting a wavelet to decompose an image signal by n layers of wavelet; 2) threshold processing process: processing the coefficients of each layer by threshold to obtain wavelet coefficients; and 3) reconstruction process: reconstructing the wavelet according to the obtained wavelet coefficients to obtain the denoised image signal. The denoised image is further evaluated by a denoising quality measurement evaluation method, and the peak signal-to-noise ratio is used as the denoising quality measurement evaluation method, which is usually defined as:

3. The method of claim 1, wherein the method comprises: And the mean square error (MSE) is defined as follows: MAX I : maximum value of the color of the image point; MSE: mean square error, H, W: height and width of the image; X(i,j): initial image pixel point; Y(i,j): denoised image pixel point. The gray processing formula in step 2 is:

4. The method of claim 1, wherein the method is a method of precise positioning and recognition of a disinfection door spray based on machine vision. Gray(i,j) = 0.299 * R(i,j) + 0.587 * G(i,j) + 0.114 * B(i,j) Gray(i,j): gray value; R(i,j): red component of the denoised image; G(i,j): green component of the denoised image; B(i,j): blue component of the denoised image. The step 3 is specifically: threshold calculation is performed for each pixel in the gray image, the threshold of the image is determined by the local gray characteristics and pixel gray value of all pixel points in the image; the pixel points less than the threshold in the image are set as a pixel value, and the pixel points greater than the threshold are set as another pixel value.

5. The method of claim 1, wherein, The Sobel operator is used for edge detection in step 4, A represents the binaryzed image pixel point, Gx and Gy represent the image gray values detected by horizontal and vertical edges respectively; 6. The method of claim 1, wherein the method is a machine vision-based precise positioning and recognition method for disinfection door spray. If the gradient G is greater than a certain threshold, the point (x, y) is considered as an edge point. The calculation of the edge center in step 5 mainly includes finding the upper and lower boundaries of the target, and the calculation form is: Thus the point gray scale: |G| = |G x |+|G y | And the center coordinates are obtained:

7. The method of claim 1, wherein the method is a machine vision-based precise positioning and recognition method for disinfection door spray. H: pixel height; ​ I(x,y): brightness function of the point. y upper : image at y coordinate highest; y lower : image at y coordinate lowest; x left : image at x coordinate leftmost; x right : image at x coordinate rightmost; W: pixel width; ​ ​ 8. The method for precise positioning and identification of disinfection gate spray based on machine vision according to claim 1, characterized in that, The precision positioning identification method can also collect spray photos sprayed by the spray head, detect the edges of the mist droplets through image processing to identify the spray concentration, thereby performing negative feedback adjustment on the working parameters of the spray head, and ensure the working efficiency and precision of the disinfection door.

9. A disinfection door based on the machine vision-based spray precision positioning identification method according to any one of claims 1-8, comprising a disinfection door body, an image acquisition device arranged at the top of the disinfection door body, a spray precision positioning identification device, and a liquid spraying device for spraying disinfectant; The image acquisition device comprises a high-speed camera, which is used to acquire images of human bodies about to pass through the disinfection door and spray pictures sprayed by the liquid spraying device and transmit them to the spray precision positioning identification device; the spray precision positioning identification device can analyze and process the crowd images and spray pictures transmitted by the image acquisition device to adjust the liquid spraying device; the liquid spraying device comprises a plurality of ultrasonic spray nozzles mounted on the disinfection door body, which are used to spray disinfectant.

Images