CN-115205249-B - Leakage detection method, device and medium for insulin

CN115205249BCN 115205249 BCN115205249 BCN 115205249BCN-115205249-B

Abstract

A leakage detection method, apparatus, and medium for insulin are described, the method including adsorbing leakage of a target location related to leakage of liquid with an adsorbent material after injection of insulin by a patient and acquiring an image of the adsorbent material adsorbed with the leakage of liquid to obtain a leakage image, contrast enhancing the leakage image to obtain an enhanced image, performing two classifications based on pixels in the enhanced image to obtain two classes of pixels, selecting a target class of pixels from the two classes of pixels based on a first gray scale interval before adsorption of insulin by the adsorbent material and a second gray scale interval after adsorption of insulin by the adsorbent material, performing convex hull fitting on an area corresponding to the target class of pixels to determine a leakage area, determining a leakage liquid volume based on a resolution of the leakage image and the leakage liquid area, and determining a leakage liquid volume corresponding to the leakage liquid volume based on adsorption capacity and the leakage liquid volume. Thus, the leakage amount can be obtained efficiently, conveniently and noninvasively.

Inventors

GAO BIN
XU XIAYU
SUN FEI
TIAN LI
LI SIMIN

Assignees

中国人民解放军空军军医大学

Dates

Publication Date: 20260508
Application Date: 20220714

Claims (10)

1. A leakage detection method for insulin is characterized by comprising the steps of obtaining a leakage image, wherein the leakage image is obtained by absorbing leakage of a target position related to the leakage by an absorption material after insulin is injected into a patient to collect leakage and collecting an image of the absorption material absorbed with the leakage, the absorption material has absorption capacity of absorbing insulin, carrying out contrast enhancement on the leakage image to obtain an enhanced image, carrying out two classification on the basis of pixel points in the enhanced image to obtain two types of pixel points, selecting a target type of pixel point from the two types of pixel points on the basis of a first gray scale interval before the absorption of insulin by the absorption material and a second gray scale interval after the absorption of insulin by the absorption material, carrying out connectivity analysis on a region corresponding to the target type of pixel point to determine at least one connection component, obtaining a convex hull of each connection component in the at least one connection component, carrying out liquid level correspondence between the connection component and the leakage region as the leakage area, and determining the leakage area based on the leakage area.
2. The liquid leakage detection method according to claim 1, wherein: And selecting the pixel points of the target class from the pixel points of the two classes, wherein if the first gray scale interval is larger than the second gray scale interval, the class with the small gray scale average value of the pixel points of the two classes is used as the target class, otherwise, the class with the large gray scale average value of the pixel points of the two classes is used as the target class.
3. The liquid leakage detection method according to claim 2, wherein: the adsorption material is a nitrocellulose membrane, the nitrocellulose membrane is white in color, and the first gray scale interval is larger than the second gray scale interval.
4. The liquid leakage detection method according to claim 1, wherein: the connectivity analysis is eight-neighborhood connectivity analysis, and the convex hull is the minimum convex polygon corresponding to each connection component.
5. The liquid leakage detection method according to claim 1, wherein: In the contrast enhancement, converting a color space of the liquid leakage image into a target color space with a brightness component, performing histogram equalization on a gray value of the brightness component, and converting the gray value of the brightness component back into an RGB space through the target color space after the histogram equalization to obtain the enhanced image, wherein a cumulative distribution function corresponding to the histogram equalization satisfies a formula: wherein k represents a gray value of the brightness component, N k represents the number of pixels with the gray value of k of the brightness component in the liquid leakage image, N represents the total number of pixels of the liquid leakage image, and S k represents the gray value of k of the brightness component after the histogram equalization.
6. The liquid leakage detection method according to claim 1, wherein: The two classification is K-means clustering, in the K-means clustering, let K be 2, the pixel point in the enhanced image be a sample, the gray values of three channels of RGB space in each sample are characteristic, and unsupervised clustering is carried out on the sample to obtain the two classification.
7. The liquid leakage detection method according to claim 1, wherein: the liquid leakage image comprises a pattern of a reference object with a known size, the number of pixel points in the liquid leakage area is determined based on the resolution of the liquid leakage image, the pixel point area is determined based on the pattern of the reference object in the liquid leakage image, and the liquid leakage area is determined based on the number of pixel points and the pixel point area.
8. The liquid leakage detection method according to claim 1, wherein: the target location includes a skin surface at an insulin outlet of an injection device for injecting insulin and around an injection location of a patient.
9. An electronic device comprising at least one processing circuit configured to perform the leakage detection method of any one of claims 1 to 8.
10. A computer readable storage medium storing at least one instruction that when executed by a processor implements the leak detection method of any one of claims 1 to 8.

Description

Leakage detection method, device and medium for insulin Technical Field The present disclosure relates generally to the medical field of health related information systems, and more particularly to a leak detection method, apparatus, and medium for insulin. Background In the medical field (e.g., in the field of diabetes or vaccination), the aim of treating or preventing a disease is often achieved by strictly controlling the injection dose. Taking the field of diabetes as an example, diabetes is a lifelong chronic endocrine disease mainly characterized by insulin resistance and islet function reduction, and can cause a series of metabolic disorder syndromes such as high protein, high fat and the like, the islet function of a diabetic patient is generally poor, the instability of blood sugar level is easy to cause, and some patients often need to control the blood sugar level of the diabetic patient by injecting insulin. Currently, insulin injection has become an important hypoglycemic approach to treating diabetes (e.g., type 2 diabetes), and by injecting a dose of insulin, a patient can control his or her blood glucose level. However, related medical clinical studies have shown that problems such as drug remaining at the outlet of an injection device or leakage of liquid medicine after injection easily occur during insulin injection, which results in a decrease in the dosage of insulin injected into a patient, and easily causes unstable blood glucose control of the patient, thereby reducing the blood glucose control effect, and there has been no related study on how to quantitatively detect leakage. Disclosure of Invention The present disclosure has been made in view of the above-described circumstances, and an object thereof is to provide a leakage detection method, apparatus, and medium for insulin that can effectively, conveniently, and non-invasively obtain leakage. The first aspect of the present disclosure provides a leakage detection method for insulin, for detecting leakage of a patient in an insulin injection process, the leakage detection method comprising obtaining a leakage image, wherein the leakage image is obtained by adsorbing leakage of a target position related to the leakage by an adsorption material after the insulin is injected into the patient and collecting an image of the adsorption material adsorbed with the leakage, the adsorption material has adsorption capacity for adsorbing the insulin, performing contrast enhancement on the leakage image to obtain an enhanced image, performing two classification based on pixel points in the enhanced image to obtain two types of pixel points, selecting a target type of pixel point from the two types of pixel points based on a first gray scale interval before the adsorption of the insulin by the adsorption material and a second gray scale interval after the adsorption of the insulin by the adsorption material, performing connectivity analysis on a region corresponding to the target type of pixel point to determine at least one connection component, obtaining a convex hull of each connection component in the at least one connection component, determining the liquid level of the connection component as the leakage region based on the liquid leakage and the liquid leakage volume area. Under the condition, the liquid leakage amount is effectively collected by utilizing the adsorption material, the pixel points of the target class are determined according to the characteristics of gray level change before and after the adsorption material adsorbs insulin, the area corresponding to the pixel points of the target class is adjusted by utilizing convex hull fitting so as to intelligently and conveniently determine the regular and accurate liquid leakage area, and the liquid leakage condition is quantitatively determined based on the liquid leakage area and the adsorption capacity of the adsorption material. Thus, the quantitative leakage condition can be effectively, conveniently and noninvasively obtained. In addition, in the liquid leakage detection method according to the first aspect of the present disclosure, optionally, in selecting the pixel point of the target class from the pixel points of the two classes, if the first gray scale interval is greater than the second gray scale interval, a class with a small gray scale average value of the pixel points of the two classes is taken as the target class, and otherwise, a class with a large gray scale average value of the pixel points of the two classes is taken as the target class. Thus, the pixel point of the target class can be determined according to the gray scale change before and after the adsorption material adsorbs insulin. In the liquid leakage detection method according to the first aspect of the present disclosure, the adsorption material may be a nitrocellulose membrane, the nitrocellulose membrane may be white in color, and the first gray scale section may be larger than the second gray s