CN112927199B - Method for selecting injection point for non-intravenous injection treatment - Google Patents

Method for selecting injection point for non-intravenous injection treatment Download PDF

Info

Publication number
CN112927199B
CN112927199B CN202110200103.2A CN202110200103A CN112927199B CN 112927199 B CN112927199 B CN 112927199B CN 202110200103 A CN202110200103 A CN 202110200103A CN 112927199 B CN112927199 B CN 112927199B
Authority
CN
China
Prior art keywords
image
gray
pixel point
binary
pixel
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.)
Active
Application number
CN202110200103.2A
Other languages
Chinese (zh)
Other versions
CN112927199A (en
Inventor
胡斌
黄定梁
石繁槐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tongji University
Original Assignee
Tongji University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Tongji University filed Critical Tongji University
Priority to CN202110200103.2A priority Critical patent/CN112927199B/en
Publication of CN112927199A publication Critical patent/CN112927199A/en
Application granted granted Critical
Publication of CN112927199B publication Critical patent/CN112927199B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformation in the plane of the image
    • G06T3/40—Scaling the whole image or part thereof
    • G06T5/70—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/136—Segmentation; Edge detection involving thresholding
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20032—Median filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30101—Blood vessel; Artery; Vein; Vascular

Abstract

A method of selecting a non-intravenous therapy injection site, comprising: s1, imaging the part to be injected of the user by using near-infrared imaging equipment to obtain a near-infrared vein image ORIGINAL of the part to be injected; s2, performing preprocessing such as cutting and zooming on the near-infrared vein image ORIGINAL obtained in S1 to obtain a GRAY image GRAY; s3, carrying out GRAY scale normalization processing on the GRAY scale image GRAY obtained in the S2 to obtain a normalized GRAY scale image NORM; s4, carrying out image binarization operation on the normalized grayscale image NORM obtained in S3 to obtain a near-infrared BINARY image BINARY; s5, carrying out filtering operation on the near-infrared BINARY image BINARY obtained in the S4 to obtain a blood VESSEL BINARY image VESSEL; and S6, collecting candidate pixel points of the vascular binary image VESSEL obtained in the S5, respectively calculating the size of a range square of each candidate pixel point, comparing the sizes of the range squares to obtain pixel point positions corresponding to one or more specific positions which are most suitable as injection points, and completing the selection of the non-intravenous injection treatment injection points.

Description

Method for selecting injection point for non-intravenous injection treatment
Technical Field
The utility model relates to the use of computer vision in the medical field.
Background
Injection therapy is a commonly used treatment. According to different positions of the needle head reaching human tissues, the injection therapy is divided into different specific treatment modes such as intradermal injection, subcutaneous injection, intramuscular injection, intravenous injection and the like. Different injection treatment modalities have different clinical applications. Compared with other treatment methods, the injection treatment can ensure that the medicine is absorbed more quickly, thereby achieving the treatment effect more quickly. In addition, some special medicines are not taken orally, such as insulin, and must be injected into human body for treatment.
However, current injection therapy also has some drawbacks and disadvantages. On the one hand, the current injection therapy usually requires the operation of professional medical staff, which is a great burden for the medical staff. On the other hand, although injection therapy can achieve good therapeutic effect, the needle and the medicine need to directly enter human tissues, and serious medical accidents are likely to be caused by careless operation, so that serious consequences are caused.
Medical accidents which easily occur in the injection treatment are mostly related to blood vessels, and the acquisition of accurate blood vessel distribution is crucial to the successful execution of various injection treatments. However, in a complex practical medical environment, accurate vascularity is difficult to obtain. During the ordinary injection treatment, medical staff needs to spend a great deal of energy to carefully identify the blood vessels of a patient, to find the blood vessel distribution of the patient, and then to find a proper and safe injection point under the condition of visible light. In some emergency situations, such as war rescues, etc., it is more difficult to accurately obtain the vascularity and find a suitable injection site. In addition, the medical subject's own condition also presents challenges to this task. Compared with healthy young and old people, the blood vessel distribution conditions of the groups such as children, old people, chronic disease and trauma patients, and patients with blood vessel collapse are more complex, the success rate of injection therapy with the first injection is difficult to guarantee, and multiple injections can cause massive hemorrhage complications to bring new risks to patients.
Near-infrared light refers to electromagnetic radiation waves between the visible and mid-infrared, with wavelengths generally ranging from 780nm to 2500 nm. Near-infrared light of a specific wavelength has excellent properties that are not possessed by some visible light, and thus is widely used. Compared with visible light, near infrared light with the wavelength of about 780nm to 1100nm has stronger penetrating power to human tissues, and the absorption capacity of the deoxyhemoglobin in blood to the near infrared light is obviously higher than that of tissues around blood vessels such as fat, melanin and the like. Therefore, one can more clearly observe blood vessels under the skin under near infrared light.
Disclosure of Invention
Aiming at the problems of injection treatment in the prior art and combining the characteristics of near-infrared light and human tissue absorption, the utility model specifically aims at non-intravenous injection treatment (including intradermal injection, subcutaneous injection, intramuscular injection and other injection treatment modes, but not including intravenous injection treatment) based on near-infrared vein images, and uses an image processing method to find the position of a proper injection point in the near-infrared vein images, thereby realizing autonomous treatment or assisting medical staff in treatment, simultaneously reducing the possibility of medical accidents, improving the success rate of injection treatment, and enabling the injection treatment to be simpler and safer.
The utility model aims to overcome the defects in the prior art, and provides an injection point selection method based on near-infrared vein images and image processing aiming at non-intravenous injection treatment, so that the injection point can be kept away from dangerous areas such as blood vessels and the like as far as possible, doctors are assisted in performing non-intravenous injection treatment on patients, and the treatment safety is improved.
The purpose of the utility model can be realized by the following technical scheme:
a method for selecting a non-intravenous injection treatment injection point based on near-infrared vein images and image processing, comprising the following steps:
s1, imaging the part to be injected of the user by using near-infrared imaging equipment to obtain a near-infrared vein image ORIGINAL (gray level image) of the part to be injected;
s2, performing preprocessing such as cutting and zooming on the near-infrared vein image ORIGINAL obtained in S1 to obtain a GRAY image GRAY;
s3, carrying out GRAY scale normalization processing on the GRAY scale image GRAY obtained in the S2 to obtain a normalized GRAY scale image NORM;
s4, carrying out image binarization operation on the normalized grayscale image NORM obtained in S3 to obtain a near-infrared BINARY image BINARY;
s5, carrying out filtering operation on the near-infrared BINARY image BINARY obtained in the S4 to obtain a blood VESSEL BINARY image VESSEL;
and S6, collecting candidate pixel points of the vascular binary image VESSEL obtained in the S5, respectively calculating the size of a range square of each candidate pixel point, comparing the size of the range square of each candidate pixel point to obtain pixel point positions corresponding to one or more specific positions which are most suitable for injection points, and finishing the selection of non-intravenous injection treatment injection points.
Further, in step S1, the near-infrared imaging device is used to image the to-be-injected region of the user, and the imaging result of the device is derived to obtain a near-infrared vein image ORIGINAL (gray scale image) of the to-be-injected region.
The light source used for the near-infrared vein image is near-infrared light in a certain range, and specifically, the wavelength range thereof is approximately 780nm to 1100 nm. The position of the near-infrared imaging device is adjusted according to the part to be injected of the user, so that the imaging range of the near-infrared imaging device covers the part to be injected and has a proper distance with the surface of the human body. After the near-infrared imaging is completed, the imaging result is derived from the device, and a near-infrared vein image ORIGINAL (gray level image) is obtained.
Further, in step S2, the near-infrared vein image ORIGINAL obtained in step S1 is preprocessed by clipping, scaling, and the like according to actual needs and specific situations, so as to obtain a GRAY level image GRAY.
Considering that the near-infrared vein image ORIGINAL derived from the near-infrared imaging device may contain some information unrelated to the part to be injected, the near-infrared vein image ORIGINAL is clipped according to actual needs so that only the image corresponding to the part to be injected is retained. And according to the actual requirement on the image size, zooming the image after cutting, and adjusting the image size to the desired size to obtain the GRAY image GRAY.
Further, in step S3, the GRAY scale normalization process is performed on the GRAY scale image GRAY obtained in step S2, and the GRAY scale value of each pixel is adjusted to obtain a normalized GRAY scale image NORM.
The GRAY level normalization is performed on the GRAY level image GRAY, so that the GRAY level of the GRAY level image GRAY is probably relatively concentrated and the contrast is low, the GRAY level normalization operation is performed to enable the GRAY level value of the GRAY level image GRAY to be distributed in the [0,255 ] interval more uniformly and dispersedly, the image contrast is enhanced, and the subsequent image processing is facilitated to obtain a better effect.
For the GRAY image GRAY, it is assumed that the GRAY value of the pixel in which the GRAY value is the largest is GRAYmaxThe gray value of the pixel with the smallest gray value is graymin. Then, for any pixel in the GRAY image GRAY, assuming that the GRAY value of the pixel before the GRAY normalization is g, after the GRAY normalization, the GRAY value g' is:
Figure BDA0002948354840000031
where round means rounding. That is, the normalized grayscale image NORM obtained through the grayscale normalization operation has an overall grayscale value ranging from [ gray ]min，graymax]Is adjusted to [0,255]。
Further, in step S4, a binarization threshold is selected for the normalized grayscale image NORM obtained in S3, and an image binarization operation is performed to obtain a near-infrared BINARY image BINARY.
Image binarization requires that a threshold value be selected first. There are many ways to select the threshold value, there is no fixed method, and according to the image processing knowledge and practical experience, the threshold value selection methods that can be used include but are not limited to: artificially given fixed thresholds, Dajin algorithm, etc. After a threshold value is selected for the NORM of the normalized gray level image, the binarization operation is carried out, and the specific steps are as follows: traversing each pixel point of the normalized grayscale image NORM, comparing the grayscale value with a selected threshold, and if the grayscale value is less than or equal to the threshold, setting the grayscale value of the pixel point to be 0 (black); otherwise, setting the gray value of the pixel point to be 255 (white), namely the gray value calculation expression of each pixel point after the image binarization is as follows:
Figure BDA0002948354840000041
where th is the selected threshold and is [0,255 ]]G is the gray value of a certain pixel point in the normalized gray image NORM, graybinaryThe gray value of the pixel point after binarization.
The gray value of each pixel point in the normalized gray image NORM can be adjusted to be one of 0 or 255 through the formula, and the near-infrared BINARY image BINARY obtained through adjustment has only two gray values, and is visually represented as only two colors of black and white.
Further, in step S5, a suitable image filtering method is adopted to perform filtering operation on the near-infrared BINARY image BINARY obtained in step S4, so as to reduce noise influence, and obtain a vascular BINARY image VESSEL.
According to the difference of the near-infrared vein image ORIGINAL acquisition quality, the distribution condition and the characteristics of the blood vessel of the part to be injected and the like, the near-infrared BINARY image BINARY may have noise with different degrees, and the noise influence is weakened through filtering processing. There are a variety of filtering methods that can be used to process the image based on the knowledge of the image processing, and filtering methods that can be used herein include, but are not limited to, median filtering. Median filtering is better suited to handle larger, discretely distributed, denser noise, where it is more appropriate.
The specific steps of median filtering are as follows: firstly, selecting a filtering template which is generally square, wherein the pixel length of the square template can be selected to be a proper value according to the situation; and traversing each pixel point in the original image to perform filtering operation, namely for any pixel point, firstly coinciding the center of the filtering template with the pixel point, then counting the gray values of all the pixel points in the filtering template, and selecting the median of the gray values as the gray value of the pixel point after filtering.
After the near-infrared BINARY image BINARY is subjected to filtering operation, the obtained result is still a BINARY image, which is defined as VESSEL here. The VESSEL can reflect the pixel position of the blood VESSEL more clearly, and since the color of the blood VESSEL portion is generally significantly darker (significantly darker) in the near-infrared vein image, through the above processing, it can be considered that the pixel point of black (gray value of 0) in the obtained blood VESSEL binary image VESSEL most likely corresponds to the blood VESSEL portion of the site to be injected.
Further, in the step S6, searching for a pixel point location corresponding to one or more specific locations that are most suitable for an injection point according to the blood VESSEL binary image VESSEL obtained in S5 includes: collecting candidate pixel points according to the vascular binary image VESSEL; respectively calculating the size of the square of the range of each candidate pixel point; and comparing the size of the square range of each candidate pixel point to obtain the pixel point position which is most suitable for corresponding to one or more specific positions of the injection point, and finishing the selection of the non-intravenous injection treatment injection point.
Further, for the blood VESSEL binary image VESSEL obtained in S5, image coordinates of pixels are first defined. Assuming that the vascular binary image VESSEL has m rows of pixels and n columns of pixels (both m and n are positive integers), the x-axis direction of the specified image coordinate is from top to bottom, and the x-axis coordinate of any pixel point is the number of rows counted from top to bottom; the y-axis direction of the defined image coordinates is from left to right, and the y-axis coordinate of any pixel point is the number of columns counted from left to right. Thus, each pixel in the blood VESSEL binary image VESSEL has a unique image coordinate.
Considering that the size of the blood VESSEL binary image VESSEL is possibly large, if each pixel point is traversed to search the most appropriate injection point position, the calculation time is possibly long; in addition, within a certain range, the differences of the pixel points close in position as injection points are not large, so that before the most suitable injection point is searched, candidate pixel points are collected firstly, and the calculation amount is reduced on the premise of ensuring the accuracy of searching the injection point.
The specific steps of collecting the candidate pixel points according to the vascular binary image VESSEL are as follows: firstly, selecting a line sampling interval k and a column sampling interval l (both k and l are positive integers) according to actual conditions and requirements, and then calculating the number of sampling lines and the number of sampling columns a and b, wherein the specific calculation formula is as follows:
Figure BDA0002948354840000051
Figure BDA0002948354840000052
wherein
Figure BDA0002948354840000053
Meaning rounding down. Finally, selecting a pixel point with an image coordinate of (ki, lj) as a candidate pixel point, wherein i is 1, 2. j is 1, 2.
Further, the size of the square of the range of each candidate pixel point is calculated. The size of the square in the range of any pixel can quantitatively describe the quality degree of the corresponding position of the pixel as an injection point. For each pixel point, the pixel point is taken as the center, the pixel point is taken as the unit, the pixel point is expanded to the periphery to form a square, in the limited vascular binary image VESSEL, a maximum square exists, all the pixel points in the coverage range of the square are white (the gray value is 255), and the square is called as a 'range square' of the pixel point. The sizes of the range squares corresponding to different pixel points may be different, and the pixel side lengths of the range squares (i.e., the number of pixels corresponding to the side length of the range squares) may be used to quantitatively describe the sizes of the range squares. And for each candidate pixel point, finding a corresponding range square, recording the side length of the pixel of the range square, and finishing the calculation of the size of the range square of each candidate pixel point.
Further, the size of the square range of each candidate pixel point is compared to obtain the pixel point position which is most suitable for corresponding to one or more specific positions of the injection point, and the selection of the non-intravenous injection treatment injection point is completed.
The requirement of the non-intravenous injection treatment task on the injection point is to be as far away from the blood VESSEL region as possible, and in the blood VESSEL binary image VESSEL, the blood VESSEL is composed of black pixel points (the gray value is 0), so according to the definition of the 'range square', the larger the 'range square' of the pixel points in the blood VESSEL binary image VESSEL is, the farther the pixel points are from the blood VESSEL region, and the more suitable the injection point is.
And selecting one or more of the largest (namely the largest side length of the pixel) range squares corresponding to all the candidate pixel points, and finding out the candidate pixel points corresponding to the one or more range squares, wherein the candidate pixel points are the pixel point positions corresponding to the specific positions which are most suitable to be used as injection points.
To this end, selection of a non-intravenous therapy injection site based on near infrared vein images and image processing is accomplished.
Compared with the prior art, the utility model has the following advantages.
1. Compared with the current common visible light manual observation, the near infrared imaging method can more easily obtain clearer blood vessel distribution conditions of the part to be injected;
2. the 'range square' is defined and used for quantitatively describing the quality degree of any pixel point as an injection point, so that the injection point can be selected more accurately and reliably, the success rate of injection treatment is improved, and the possibility of medical accidents is reduced.
3. The method is completely automatically completed by algorithms such as image processing and the like, manual intervention is not needed, and injection point selection can be rapidly, efficiently and repeatedly performed. The technical scheme of the utility model does not form a complete treatment scheme per se, but can be used as a technical means for assisting and supporting medical staff to treat. By means of the technology, the burden of medical staff participating in treatment is favorably reduced, meanwhile, errors caused by artificial factors such as overwork and the like are avoided, and the treatment success rate of the medical staff is improved.
Drawings
FIG. 1 is a schematic flow chart of the present invention.
Fig. 2 is a schematic diagram of a near-infrared vein image ORIGINAL used in the present invention.
Fig. 3 is a GRAY diagram of a GRAY image used in the present invention.
FIG. 4 is a schematic diagram of a normalized grayscale image NORM used in the present invention.
Fig. 5 is a schematic diagram of a near-infrared BINARY image BINARY used in the present invention.
Fig. 6 is a diagram of a median filtering template used in the present invention.
Fig. 7 is a schematic diagram of a vascular binary image VESSEL used in the present invention.
FIG. 8 is a schematic diagram of image coordinates for use with the present invention.
Fig. 9 is a schematic diagram of a candidate pixel point collection method used in the present invention.
FIG. 10 is a diagram illustrating the square effect of the range of each pixel defined by the present invention.
Fig. 11 is a pseudo code for step S6 of selecting an injection point according to the present invention.
FIG. 12 is a graphical representation of the results of a non-IV treatment injection site selection in accordance with the present invention.
Detailed Description
The utility model is described in detail below with reference to the figures and specific embodiments. The present example is carried out on the premise of the technical scheme of the present invention, and a detailed implementation manner and a specific operation process are given, but the protection scope of the present invention is not limited to the following examples.
As shown in fig. 1, the present embodiment provides a method for selecting a non-iv injection treatment injection point based on near-infrared iv images and image processing, comprising the following steps:
step S1, imaging the part to be injected of the user by using near-infrared imaging equipment to obtain a near-infrared vein image ORIGINAL (gray level image) of the part to be injected;
step S2, preprocessing the near-infrared vein image ORIGINAL obtained in the step S1 such as cutting and zooming to obtain a GRAY level image GRAY;
step S3, carrying out GRAY scale normalization processing on the GRAY scale image GRAY obtained in the step S2 to obtain a normalized GRAY scale image NORM;
step S4, carrying out image binarization operation on the normalized gray level image NORM obtained in the step S3 to obtain a near-infrared BINARY image BINARY;
step S5, carrying out filtering operation on the near-infrared BINARY image BINARY obtained in the step S4 to obtain a blood VESSEL BINARY image VESSEL;
and S6, collecting candidate pixel points of the vascular binary image VESSEL obtained in the S5, respectively calculating the size of a square range of each candidate pixel point, comparing the sizes of the squares to obtain pixel point positions corresponding to one or more specific positions which are most suitable for injection points, and completing selection of non-intravenous injection treatment injection points.
In step S1, a near-infrared vein image is obtained by: the near-infrared imaging method comprises the steps of taking near-infrared light with the wavelength range of 780nm to 1100nm as a light source, adjusting the position of near-infrared imaging equipment according to a to-be-injected part of a user to enable the imaging range of the near-infrared imaging equipment to cover the to-be-injected part and have a proper distance with the surface of a human body, and then imaging the to-be-injected part of the user by using the near-infrared imaging equipment. After the near-infrared imaging is completed, the imaging result is exported from the equipment, and a near-infrared vein image ORIGINAL (gray image) of the part to be injected is obtained. The near-infrared vein image ORIGINAL schematic diagram of the part to be injected is shown in figure 2, which is obtained by imaging the part to be injected of a human body by using a scotopic low-light-level VIVO500S projection type near-infrared blood vessel imager.
In step S2, the near-infrared vein image ORIGINAL obtained in step S1 is preprocessed to obtain a GRAY-scale image GRAY, and the process specifically includes:
(1) clipping near-infrared vein image ORIGINAL
And cutting the near-infrared vein image ORIGINAL according to actual needs, removing information irrelevant to the part to be injected, and only keeping the image corresponding to the part to be injected. In the present embodiment, the band of the near-infrared vein image ORIGINAL is 640 × 480, that is, 640 pixels per row and 192 pixels per column. According to actual needs, only images corresponding to injection sites need to be reserved, namely rectangles surrounded by pixels on the 128 th row and the 255 th row (including the 128 th row and the 255 th row) from top to bottom and pixels on the 192 th column and the 413 th column (including the 192 th column and the 413 th column) from left to right. The size of the gray image obtained after cropping is 222 × 128, that is, 222 pixels per row and 192 pixels per column.
(2) Scaling the cropped image
And according to the actual requirement on the image size, zooming the image after the image is cut, and adjusting the image size to the expected size. In this embodiment, the image size obtained after the cropping is relatively appropriate, and the GRAY image GRAY is directly obtained without further scaling operation (scaling ratio of 1).
A schematic diagram of a GRAY-scale image GRAY obtained by preprocessing the near-infrared vein image ORIGINAL is shown in fig. 3.
In step S3, the GRAY scale image GRAY obtained in S2 is subjected to GRAY scale normalization processing to obtain a normalized GRAY scale image NORM in which the GRAY scale values of the image are distributed more uniformly and dispersedly at [0,255 []In the interval, the image contrast is enhanced, and the subsequent image processing can obtain better effect. The specific method for carrying out GRAY level normalization on the GRAY level image GRAY is that: let GRAY value of pixel with maximum GRAY value in GRAY image GRAY be GRAYmaxThe gray value of the pixel with the smallest gray value is graymin. Then, for any pixel in the GRAY image GRAY, assuming that the GRAY value of the pixel before the GRAY normalization is g, after the GRAY normalization, the GRAY value g' is:
Figure BDA0002948354840000081
where round means rounding. Traversing each pixel in the GRAY level image GRAY, calculating a new corresponding pixel GRAY level value by using the above formula, and replacing the original GRAY level value of the corresponding pixel to obtain a normalized GRAY level image NORM after GRAY level normalization. That is, the normalized grayscale image NORM obtained through the grayscale normalization operation has an overall grayscale value ranging from [ gray ]min，graymax]Is adjusted to [0,255]。
A schematic diagram of the normalized grayscale image NORM obtained by performing grayscale normalization on the grayscale image GRAY is shown in fig. 4.
In step S4, an image binarization operation is performed on the normalized grayscale image NORM obtained in S3 to obtain a near-infrared BINARY image BINARY, which specifically includes:
(1) binary threshold selection
Before the image binarization operation is performed, an appropriate threshold value is selected first. There are many ways to select the threshold value, there is no fixed method, and according to the image processing knowledge and practical experience, the threshold value selection methods that can be used include but are not limited to: artificially given fixed thresholds, Dajin algorithm, etc. In the embodiment, the Dajin algorithm is used for selecting the image binarization threshold, and the algorithm can be realized by directly calling functions of open source libraries such as OPENCV.
(2) Image binarization operation according to threshold value
After a threshold value is selected for the NORM of the normalized gray level image, the binarization operation is carried out, and the specific steps are as follows: traversing each pixel point of the normalized grayscale image NORM, comparing the grayscale value with a selected threshold, and if the grayscale value is less than or equal to the threshold, setting the grayscale value of the pixel point to be 0 (black); otherwise, setting the gray value of the pixel point to be 255 (white), namely, the gray value calculation expression of each pixel point after the image binarization is as follows:
Figure BDA0002948354840000091
where th is the selected threshold and is [0,255 ]]G is the gray value of a certain pixel point in the normalized gray image NORM, graybinaryThe gray value of the pixel point after binarization.
The gray value of each pixel point in the normalized gray image NORM can be adjusted to be one of 0 or 255 through the formula, and the near-infrared BINARY image BINARY obtained through adjustment has only two gray values, and is visually represented as only two colors of black and white.
The effect of the near-infrared BINARY image BINARY obtained by performing image binarization on the normalized grayscale image NORM is shown in fig. 5.
In step S5, a suitable image filtering method is adopted to perform filtering operation on the near-infrared BINARY image BINARY obtained in step S4, so as to reduce the influence of noise, and obtain a blood VESSEL BINARY image VESSEL. According to the difference of the near-infrared vein image ORIGINAL acquisition quality, the distribution condition and the characteristics of the blood vessel of the part to be injected and the like, the near-infrared BINARY image BINARY may have noise with different degrees, and the noise influence is weakened through filtering processing. There are a variety of filtering methods that can be used to process the image based on image processing knowledge, and filtering methods that can be used herein include, but are not limited to, median filtering. Median filtering is better suited to handle larger, discretely distributed, denser noise, where it is more appropriate. The present embodiment uses median filtering to reduce the influence of noise on a binary image.
The specific steps of median filtering are as follows: firstly, selecting a filtering template which is generally square, wherein the pixel length of the square template can be selected to be a proper value according to the situation; and traversing each pixel point in the original image to perform filtering operation, namely, for any pixel point, firstly coinciding the center of the filtering template with the pixel point, then counting the gray values of all the pixel points in the filtering template, and selecting the median of the gray values as the gray value of the pixel point after filtering. Therefore, the result obtained after median filtering the binary image is still a binary image. In this embodiment, the median filtering template is selected as a square with a side length of 5 pixels, as shown in fig. 6.
After the near-infrared BINARY image BINARY is subjected to filtering operation, the obtained result is still a BINARY image and is marked as VESSEL. The VESSEL can reflect the pixel position of the blood VESSEL more clearly, and since the color of the blood VESSEL portion is generally significantly darker (significantly darker) in the near-infrared vein image, through the above processing, it can be considered that the pixel point of black (gray value of 0) in the obtained blood VESSEL binary image VESSEL most likely corresponds to the blood VESSEL portion of the site to be injected.
The median filtering operation can reduce the noise influence in a limited way, and a filtered vascular binary image VESSEL effect diagram is shown in FIG. 7.
In step S6, candidate pixel points are acquired from the blood VESSEL binary image VESSEL obtained in step S5, the size of the square of the range of each candidate pixel point is calculated and compared, and a pixel point position corresponding to one or more specific positions most suitable as an injection point is obtained, so that the selection of a non-intravenous injection treatment injection point is completed, and the process specifically includes:
(1) collecting candidate pixel points according to vascular binary image VESSEL
For the blood VESSEL binary image VESSEL obtained in S5, image coordinates of pixels are first defined. Assuming that the vascular binary image VESSEL has m rows of pixels and n columns of pixels (both m and n are positive integers), the x-axis direction of the specified image coordinate is from top to bottom, and the x-axis coordinate of any pixel point is the number of rows counted from top to bottom; the y-axis direction of the defined image coordinates is from left to right, and the y-axis coordinate of any pixel point is the number of columns counted from left to right. Thus, each pixel in the blood VESSEL binary image VESSEL has a unique image coordinate. In this embodiment, the pixel row number m is 128, the pixel column number n is 222, and the image coordinate is schematically shown in fig. 8, where the image coordinate includes 128 rows and 222 columns of pixels: the x-axis, the y-axis, and the origin O, "number of columns counting from left to right," and "number of rows counting from top to bottom.
Fig. 8 may be understood as an illustrative schematic diagram of the definition of the image coordinates, or may be expressed as "image coordinate system corresponding to the image according to the definition of the image coordinates".
Considering that the size of the blood VESSEL binary image VESSEL is possibly large, if each pixel point is traversed to search the most appropriate injection point position, the calculation time is possibly long; in addition, within a certain range, the differences of the pixel points close in position as injection points are not large, so that before the most suitable injection point is searched, candidate pixel points are collected firstly, and the calculation amount is reduced on the premise of ensuring the accuracy of searching the injection point.
The specific steps of collecting the candidate pixel points according to the vascular binary image VESSEL are as follows: firstly, selecting a line sampling interval k and a column sampling interval l (both k and l are positive integers) according to actual conditions and requirements, and then calculating the number of sampling lines and the number of sampling columns a and b, wherein the specific calculation formula is as follows:
Figure BDA0002948354840000111
Figure BDA0002948354840000112
wherein
Figure BDA0002948354840000113
Indicating a rounding down. Finally, selecting a pixel point with an image coordinate of (ki, lj) as a candidate pixel point, wherein i is 1, 2. j is 1, 2.
In this embodiment, k is equal to l is equal to 4, the sampling row number a is equal to 55, the sampling column number b is equal to 32, and 55 × 32 is total collected to 1760 candidate pixels. An effect schematic diagram of collecting candidate pixel points in this embodiment is shown in fig. 9, where each square represents a pixel point, a gray square represents a pixel point that is not collected, and a gray square represents a pixel point that is collected as a candidate pixel point.
(2) Calculating the size of the square of the range of each candidate pixel point
And calculating the size of the square in the range of each candidate pixel point. The size of the square in the range of any pixel can quantitatively describe the quality degree of the corresponding position of the pixel as an injection point. For each pixel point, the pixel point is taken as the center, the pixel point is taken as the unit, the pixel point is expanded to the periphery to form a square, in the limited vascular binary image VESSEL, a maximum square exists, all the pixel points in the coverage range of the square are white (the gray value is 255), and the square is called as a 'range square' of the pixel point. The range squares corresponding to different pixels have different effects as shown in fig. 10, and the size of the range squares may be quantitatively described by using the pixel side length of the range squares (i.e., the number of pixels corresponding to the side length of the range squares). And for each candidate pixel point, finding a corresponding range square, recording the side length of the pixel of the range square, and completing the calculation of the size of the range square of each candidate pixel point.
(3) Selecting the most suitable injection point
And comparing the size of the square range of each candidate pixel point to obtain the pixel point position which is most suitable for corresponding to one or more specific positions of the injection point, and finishing the selection of the non-intravenous injection treatment injection point.
The requirement of the non-intravenous injection treatment task on the injection point is to be as far away from the blood VESSEL region as possible, and in the blood VESSEL binary image VESSEL, the blood VESSEL is composed of black pixel points (the gray value is 0), so according to the definition of the range square, the larger the range square of the pixel points in the blood VESSEL binary image VESSEL is, the farther the pixel points are away from the blood VESSEL region, and the more suitable the pixel points are as the injection point.
And selecting one or more of the largest (namely the largest side length of the pixel) range squares corresponding to all the candidate pixel points, and finding out the candidate pixel points corresponding to the one or more range squares, wherein the candidate pixel points are the pixel point positions corresponding to the specific positions which are most suitable to be used as injection points.
To this end, selection of a non-intravenous therapy injection site based on near infrared vein images and image processing is accomplished.
The whole operation in the above step S6 can be represented by the pseudo code shown in fig. 11. In the present embodiment, the blood VESSEL binary image VESSEL is input, the sampling interval k is equal to l is equal to 4, the number of the expected injection points to be searched is set to be 5, and the effect of the obtained selection result of the non-intravenous injection treatment injection points is schematically shown in fig. 12, where the pixel positions corresponding to the obtained 5 most suitable injection positions are marked by squares filled with black frame gray in the near-infrared vein image ORIGINAL (gray scale image).
The foregoing detailed description of the preferred embodiments of the utility model has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.

Claims (9)

1. A method for selecting a non-intravenous therapy injection site, comprising the steps of:
s1, imaging the part to be injected of the user by using near-infrared imaging equipment to obtain a near-infrared vein image ORIGINAL of the part to be injected;
s2, cutting and zooming the near-infrared vein image ORIGINAL obtained in S1 to obtain a GRAY image GRAY;
s3, carrying out GRAY scale normalization processing on the GRAY scale image GRAY obtained in the S2 to obtain a normalized GRAY scale image NORM;
s4, carrying out image binarization operation on the normalized grayscale image NORM obtained in S3 to obtain a near-infrared BINARY image BINARY;
s5, carrying out filtering operation on the near-infrared BINARY image BINARY obtained in the S4 to obtain a blood VESSEL BINARY image VESSEL;
s6, collecting candidate pixel points of the blood VESSEL binary image VESSEL obtained in S5, respectively calculating the size of a range square of each candidate pixel point, comparing the size of the range square of each candidate pixel point to obtain pixel point positions corresponding to one or more specific positions which are most suitable for injection points, and completing selection of non-intravenous injection treatment injection points; for each pixel point, the pixel point is taken as the center, the pixel point is taken as the unit, the pixel point is expanded to the periphery to form a square, in the limited vascular binary image VESSEL, a maximum square exists, all the pixel points in the coverage range of the square are white, and the square is called as a 'range square' of the pixel point.
2. The method according to claim 1, wherein in step S1, the near-infrared imaging device is used to image the injection site of the user, and the device imaging result is derived to obtain the near-infrared vein image ORIGINAL of the injection site.
3. The method as claimed in claim 1, wherein in step S2, the near-infrared vein image ORIGINAL obtained in step S1 is cut and scaled to obtain a GRAY-scale image GRAY according to actual needs and specific situations.
4. The method according to claim 1, wherein in step S3, the GRAY image GRAY obtained in step S2 is subjected to GRAY normalization processing to adjust the GRAY value of each pixel point to obtain a normalized GRAY image NORM;
for the GRAY image GRAY, it is assumed that the GRAY value of the pixel in which the GRAY value is the largest is GRAYmaxThe gray value of the pixel with the smallest gray value is graymin(ii) a Then, for any pixel in the GRAY image GRAY, assuming that the GRAY value of the pixel before the GRAY normalization is g, after the GRAY normalization, the GRAY value g' is:
Figure FDA0003508835330000021
wherein round represents rounding; through gray normalizationNormalized gray level image NORM obtained by the quantization operation and having an overall gray level ranging from graymin，graymax]Is adjusted to [0,255]。
5. The method according to claim 1, characterized in that in step S4, a binarization threshold is selected for the normalized grayscale image NORM obtained in S3, and an image binarization operation is performed to obtain a near-infrared BINARY image BINARY;
firstly, selecting a threshold value for image binarization;
the method comprises the following specific steps: traversing each pixel point of the normalized grayscale image NORM, comparing the grayscale value with a selected threshold, and if the grayscale value is less than or equal to the threshold, setting the grayscale value of the pixel point to be 0; otherwise, setting the gray value of the pixel point to be 255, namely, the gray value calculation expression of each pixel point after the image binarization is as follows:
Figure FDA0003508835330000022
where th is the selected threshold and is [0,255 [ ]]G' is the gray value, gray, of a certain pixel point in the normalized gray image NORMbinaryThe gray value of the pixel point after binarization.
6. The method according to claim 1, characterized in that in step S5, a median filtering method is adopted to perform filtering operation on the near-infrared BINARY image BINARY obtained in S4 to reduce noise influence and obtain a blood VESSEL BINARY image VESSEL;
the specific steps of median filtering are as follows: firstly, selecting a filtering template which is a square, wherein the pixel length of the square template is selected according to the situation; then traversing each pixel point in the original image to carry out filtering operation, namely for any pixel point, firstly coinciding the center of a filtering template with the pixel point, then counting the gray values of all the pixel points in the filtering template, and selecting the median of the gray values as the gray value of the pixel point after filtering;
after the near-infrared BINARY image BINARY is subjected to filtering operation, the obtained result is still a BINARY image, which is defined as VESSEL herein.
7. The method as claimed in claim 6, wherein, assuming that the vascular binary image VESSEL has m rows of pixels, n columns of pixels, m and n are positive integers, the x-axis direction of the image coordinate is defined from top to bottom, and the x-axis coordinate of any pixel point is the number of rows from top to bottom; the y-axis direction of the specified image coordinate is from left to right, and the y-axis coordinate of any pixel point is the number of columns counted from left to right; thus, each pixel in the blood VESSEL binary image VESSEL has a unique image coordinate;
the specific steps of collecting the candidate pixel points according to the vascular binary image VESSEL are as follows: firstly, selecting a row sampling interval k and a column sampling interval l according to actual conditions and requirements, wherein both k and l are positive integers, and then calculating the number of sampling rows and the number of sampling columns a and b, wherein the specific calculation formula is as follows:
Figure FDA0003508835330000031
Figure FDA0003508835330000032
wherein
Figure FDA0003508835330000033
Represents rounding down; finally, selecting a pixel point with the image coordinate (ki, lj) as a candidate pixel point, wherein i is 1,2, …, a; j is 1,2, …, b.
8. The method of claim 7, characterized in that, further, for each candidate pixel point, the size of its range square is calculated; quantitatively describing the size of the range square by the number of pixels corresponding to the side length of the range square; and for each candidate pixel point, finding a corresponding range square, recording the side length of the pixel of the range square, and completing the calculation of the size of the range square of each candidate pixel point.
9. The method of claim 7, further comprising selecting one or more specific pixel points as the most suitable injection point location for selecting a non-iv therapy injection point by:
and taking a candidate pixel point corresponding to one or more range squares with the largest pixel side length as the injection point position.
CN202110200103.2A 2021-02-23 2021-02-23 Method for selecting injection point for non-intravenous injection treatment Active CN112927199B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110200103.2A CN112927199B (en) 2021-02-23 2021-02-23 Method for selecting injection point for non-intravenous injection treatment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110200103.2A CN112927199B (en) 2021-02-23 2021-02-23 Method for selecting injection point for non-intravenous injection treatment

Publications (2)

Publication Number Publication Date
CN112927199A CN112927199A (en) 2021-06-08
CN112927199B true CN112927199B (en) 2022-05-13

Family

ID=76170218

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110200103.2A Active CN112927199B (en) 2021-02-23 2021-02-23 Method for selecting injection point for non-intravenous injection treatment

Country Status (1)

Country Link
CN (1) CN112927199B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115620286B (en) * 2022-11-02 2023-05-05 安徽云层智能科技有限公司 Automatic data labeling system and method based on big data

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006085288A2 (en) * 2005-02-14 2006-08-17 Koninklijke Philips Electronics N.V. Apparatus and method for determining an injection point for targeted drug delivery
CN204636311U (en) * 2015-04-29 2015-09-16 吉林大学 Medical near-infrared venae subcutaneae detects imager
CN110694149A (en) * 2019-10-16 2020-01-17 山东大学齐鲁医院 Ultrasonic-assisted muscle identification method and system and auxiliary injection device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11389171B2 (en) * 2006-11-21 2022-07-19 David S. Goldsmith Integrated system for the infixion and retrieval of implants
WO2010056538A1 (en) * 2008-10-29 2010-05-20 Tim Maguire An automated vessel puncture device using three-dimensional(3d) near infrared (nir) imaging and a robotically driven needle
CN109788927B (en) * 2016-09-23 2023-06-23 皇家飞利浦有限公司 Volume rendering for planning the position of an injection site
US11400212B2 (en) * 2019-07-19 2022-08-02 Nexus Medical, Llc Clinical assessment of an intravenous catheter site

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2006085288A2 (en) * 2005-02-14 2006-08-17 Koninklijke Philips Electronics N.V. Apparatus and method for determining an injection point for targeted drug delivery
CN204636311U (en) * 2015-04-29 2015-09-16 吉林大学 Medical near-infrared venae subcutaneae detects imager
CN110694149A (en) * 2019-10-16 2020-01-17 山东大学齐鲁医院 Ultrasonic-assisted muscle identification method and system and auxiliary injection device

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"载顺铂的PLGA/Fe O 原位植入物磁热消融联合化疗增效兔VX2肿瘤实验研究";杨扬;《博士电子期刊》;20171115(第11期);论文全文 *
"近红外人体静脉成像技术的研究";杨小庆;《说是电子期刊》;20180815(第08期);论文全文 *

Also Published As

Publication number Publication date
CN112927199A (en) 2021-06-08

Similar Documents

Publication Publication Date Title
CN110689083A (en) Context pyramid fusion network and image segmentation method
JP4509531B2 (en) Acute cerebral infarction diagnosis and treatment support device
US20060173351A1 (en) System and method for inserting a needle into a blood vessel
Chen et al. Portable robot for autonomous venipuncture using 3D near infrared image guidance
CN107041729A (en) Binocular near infrared imaging system and blood vessel recognition methods
Mansoor et al. Real-time low cost infrared vein imaging system
CN112927199B (en) Method for selecting injection point for non-intravenous injection treatment
CN110448267A (en) A kind of multimode eyeground dynamic imaging analysis system and its method
CN110298273A (en) A kind of 3D based on multispectral image refers to vein extracting method and system
Marathe et al. A novel wireless vein finder
CN111339828A (en) Vein developing identification method based on combination of infrared image and ultrasonic Doppler
Ahmed et al. Real time injecting device with automated robust vein detection using near infrared camera and live video
CN116993628B (en) CT image enhancement system for tumor radio frequency ablation guidance
Brewer et al. Visual vein-finding for robotic IV insertion
CN112075981B (en) Venipuncture robot control method, device and computer-readable storage medium
CN112183518A (en) Vein target point automatic determination method, device and equipment
CN111667905A (en) Examination device and information processing method for traditional Chinese medicine eye diagnosis
May et al. Real Time Vein Visualization using Near-Infrared Imaging
CN104616260A (en) Vein image enhancement method and device
CN114677713A (en) Near-infrared light-based arm vein blood sampling point identification method and system
Cao et al. Venibot: Towards autonomous venipuncture with automatic puncture area and angle regression from nir images
US20130261424A1 (en) System for Inducing Respiration Using Biofeedback Principle
CN112102227A (en) Blood vessel optimal puncture point selection method based on blood vessel size characteristics
CN110585591A (en) Brain vision detection and analysis equipment and method based on nerve feedback
Garduno-Alvarado et al. Fast optic disc segmentation in fundus images

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
GR01 Patent grant
GR01 Patent grant