CN-121994737-A - Automatic fecal water content detection system, method and input and output balance management method

CN121994737ACN 121994737 ACN121994737 ACN 121994737ACN-121994737-A

Abstract

The application relates to the technical field of intelligent detection, in particular to an automatic fecal water content detection system, an automatic fecal water content detection method and an input and output balance management method, which are used for homogenizing the fecal in a sealed bag by a high-rotation homogenizer after a fecal sample is sealed by the sealed bag so as to uniformly distribute water and solid components, and then extracting a sample from the homogenized sealed bag, injecting the sample into a detection bottle, detecting the weight of the sample through a multi-mode detection unit, transmitting near infrared light through a near infrared spectrometer to penetrate through a quantitative sample in the detection bottle, capturing the absorption spectrum of moisture, calculating the total water content by adopting a pre-trained fusion large model and combining the total weight, and encrypting and transmitting the total water content data to a management system. Through sealed homogeneity automation mechanized operation, avoid medical personnel and excrement and urine direct contact, eliminate peculiar smell, vision pollution and cross infection's risk, quantitative sampling and multimode detection promote the accuracy of testing result after the homogeneity, promote the accuracy that the moisture content detected, promote detection efficiency.

Inventors

ZHOU YOUTIAN
LI LI
XU XI
XIE HONG
CHEN JINYING
HU MEIRONG
DING ZHANGYAN

Assignees

中国人民解放军南部战区总医院

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (12)

1. An automated fecal moisture content detection system, comprising: the split charging unit is used for obtaining and sealing a packaging bag filled with the fecal sample; the mixing unit is used for obtaining the packaging bag and carrying out homogenizing operation so as to uniformly distribute the fecal components; the sampling unit is used for quantitatively sampling the fecal samples which are uniformly distributed to obtain quantitative samples and injecting the quantitative samples into the detection bottle; the multi-mode detection unit is used for acquiring the net total weight of the fecal sample in the packaging bag through the weight detection unit, acquiring spectral characteristic data of the sample in the detection bottle through the near infrared spectrometer, detecting the temperature of the near infrared detection environment in real time through the temperature compensation device and correcting the temperature of the spectral characteristic data; The large model calculation unit inputs the spectral characteristic data after temperature correction into the pre-training model to obtain the water content of the fecal sample, and calculates the total water content of the fecal sample according to the water content and the total net weight of the fecal sample; and the data transmission unit is used for transmitting the total water content of the fecal sample in the packaging bag to the processing module of the input/output balance management system.
2. The automated fecal moisture content detection system of claim 1 wherein the step of obtaining a packaging bag and performing a homogenization operation to uniformly distribute fecal components comprises: the packaging bag is placed into a negative pressure homogenizing cavity, a low-speed rotating motor is arranged in the negative pressure homogenizing cavity, and the packaging bag is driven to rotate through the low-speed rotating motor, so that the fecal components are uniformly distributed.
3. The automated fecal moisture content detection system of claim 1 wherein the packaging bag is provided with a sampling location area where a highly elastic thickened composite membrane or embedded highly elastic reinforcement layer is used to allow for automatic closing and stop sample leakage after withdrawal of the sampling needle.
4. The automated fecal moisture content detection system of claim 3 wherein the fecal sample is quantitatively sampled by a disposable sampling needle through a sampling location area; The disposable sampling needle comprises a self-sealing silica gel rebound cap, the self-sealing silica gel rebound cap comprises a silica gel main body, a built-in stainless steel rebound spring and a sealing lip, when the disposable sampling needle pierces a sampling positioning area of the packaging bag, the silica gel main body is subjected to pressure compression and drives the rebound spring to store energy, the rebound spring releases elasticity when the disposable sampling needle is pulled out to enable the rebound cap to reset, and the sealing lip seals the sampling hole to prevent excrement samples from leaking.
5. The automated fecal moisture content testing system of claim 1 wherein the step of homogenizing the packaging bag with fecal sample by the mixing unit further comprises obtaining a net total weight of fecal sample by the weight testing unit and calculating a time of operation and/or a speed of homogenization of the homogenizing operation based on the net total weight.
6. The automated fecal moisture content testing system of claim 1 wherein the step of obtaining spectral signature data of the sample in the test vial by a near infrared spectrometer comprises: The near infrared spectrometer is a small grating near infrared spectrometer, the spectrum coverage range is 780-1700nm, the resolution is set to 8-16cm -1 , and a single InGaAs detector and a temperature compensation module are configured; The detection bottle is made of quartz glass, a quantitative sample is sampled after homogenization, and the quantitative sample is placed in a constant-temperature environment for standing balance after bubbles are eliminated; When the spectrum is collected, a detection bottle is used as a background to scan to obtain a baseline spectrum, the detection bottle after balance treatment is placed in a sample tank, the center of Ji Guanglu is set with scanning times, integration time and single detection time consumption, the original spectrum is sequentially subjected to normalization pretreatment, spectrum characteristics of 1450nm and 1550-1700nm in a 780-1700nm wave band are extracted, 1450nm is a moisture characteristic peak, 1550-1700nm is a protein characteristic area, the main component analysis is used for reducing the dimension to the first 5 main components, and the main components are converted into a JSON format and transmitted to a large model calculation unit.
7. The automated fecal moisture content detection system of claim 6 wherein the step of normalizing the raw spectra comprises: The method comprises the steps of fitting and smoothing an original spectrum of an acquired fecal sample by adopting a local low-order polynomial with a window size of 5-7 points and a polynomial order of 2, eliminating random noise caused by the differences of the sizes and the viscosity degrees of fecal particles, and keeping the shape and the position of a spectrum characteristic peak; Fitting a baseline by iterative weighting, endowing background fields such as food residues and microorganisms in the fecal sample with high weight, endowing target component peak areas such as moisture and protein with low weight, and inhibiting baseline drift by combining a punishment item to eliminate baseline drift and non-target component interference caused by complex fecal sample background; The first derivative calculation with the difference interval of 5 points is adopted, the resolution of the overlapping peaks of the water and the protein in the fecal sample is enhanced, and the characteristic peak positions are highlighted to eliminate the peak overlapping caused by the mixing of fecal components; Normalization is performed on each stool sample spectrum sample by sample, the spectrum mean value is subtracted and divided by the standard deviation, and the heat dissipation effect caused by the stool viscosity difference is eliminated so that the stool sample spectra in different physical states are comparable.
8. The automated fecal moisture content detection system of claim 1 wherein the temperature correction algorithm for temperature correcting the spectral signature data is: pre-establishing a temperature-spectrum characteristic correction model aiming at a moisture characteristic peak and a protein characteristic region in the fecal sample; the model training step comprises the steps of collecting standard fecal sample spectrum data in a 20-30 ℃ interval, fitting to obtain a 2-order polynomial relation between characteristic peak intensity and temperature, wherein I corr =I raw +a×(T-T0)+b×(T-T0) 2 , I corr is corrected characteristic peak intensity, I raw is original spectrum intensity, T 0 =25 ℃ is reference temperature, and model coefficients obtained by optimizing fecal samples are water peaks a= -0.003, b=0.0001, protein regions a= -0.0025 and b=0.00008; The real-time correction step comprises the step of calculating the characteristic peak intensity of the current spectrum in real time by utilizing a 2-order polynomial relation, and eliminating the offset influence of temperature fluctuation on the spectrum characteristics.
9. The automated fecal moisture content detection system of claim 1 wherein the step of inputting the temperature corrected spectral signature data into a pre-trained model to obtain the moisture content of the fecal sample and calculating the total moisture content of the fecal sample based on the moisture content and the total net weight of the fecal sample comprises: Constructing and deploying a robust prediction model, and inputting spectral characteristic data of a sample to be detected after temperature correction into the model to output a predicted water content W; Calculating the total weight of the fecal sample as the total weight of the packaging bag and the fecal sample-the weight of the packaging bag; the total moisture content of the fecal sample was calculated as the total fecal sample net weight x W/100.
10. The automated fecal moisture content detection system of claim 1 further comprising: The homogenizing unit comprises an electromagnetic vibration homogenizer arranged at the bottom of the collecting tank of the near infrared spectrometer, and the electromagnetic vibration homogenizer is used for bearing the detection bottle and is coaxially aligned with the optical path of the spectrometer; The triggering time of the homogenizing unit is after the weight detection unit obtains the total net weight of the fecal sample in the packaging bag, and before the near infrared spectrometer obtains the spectral characteristic data of the quantitative sample in the detection bottle; The basic vibration frequency of the homogenizing unit is 50Hz plus or minus 2Hz, the amplitude of 0.5mm plus or minus 0.1mm and the duration of 2s plus or minus 0.2s, and the homogenizing unit is kept stand for at least 0.5s and pre-collects 2 spectra; if the variation of the intensity of the 1450nm moisture peak is less than 1%, the method enters into formal spectrum collection, otherwise, the homogenization operation is repeated for 2 times at most.
11. An automated fecal moisture content detection method comprising the steps of: s1, acquiring a packaging bag filled with a fecal sample, and sealing the packaging bag; S2, homogenizing the sealed packaging bag to uniformly distribute fecal components in the packaging bag; s3, quantitatively sampling the fecal sample in the homogenized packaging bag, obtaining a quantitative sample and injecting the quantitative sample into a detection bottle; S4, acquiring the net total weight of the fecal sample in the packaging bag through a weight detection unit, acquiring spectral characteristic data of the quantitative sample in the detection bottle through a near infrared spectrometer, detecting the temperature of a near infrared detection environment in real time through a temperature compensation device, and carrying out temperature correction on the spectral characteristic data by utilizing the temperature; s5, inputting the spectral characteristic data subjected to temperature correction into a pre-training model to obtain the water content of the fecal sample; And S6, sending the total water content of the fecal sample in the packaging bag to a processing module of the input/output balance management system.
12. The method for managing the balance of the input and output quantity is characterized by comprising the following steps: p1, acquiring drinking water intake, food water content and diet type data of an individual in real time through an intelligent terminal; P2, calculating urine discharge amount, drainage amount, vomit and total water content of excrement of an individual, wherein the calculation step of total water content of the excrement comprises the steps of obtaining a packaging bag filled with an excrement sample, sealing the packaging bag, homogenizing the sealed packaging bag to enable excrement components in the packaging bag to be uniformly distributed, quantitatively sampling the excrement sample in the homogenized packaging bag, obtaining a quantitative sample, injecting the quantitative sample into a detection bottle, obtaining the total net weight of the excrement sample in the packaging bag through a weight detection unit, obtaining spectral characteristic data of the quantitative sample in the detection bottle through a near infrared spectrometer, detecting the temperature of a near infrared detection environment in real time through a temperature compensation device, correcting the spectral characteristic data by utilizing the temperature, inputting the spectral characteristic data after temperature correction into a pre-training model to obtain the water content of the excrement sample, calculating the total water content of the excrement sample according to the water content and the total net total weight of the excrement sample, and sending the total water content of the excrement sample in the packaging bag to a processing module of an input and output balance management system; P3, integrating drinking water intake, food water content, diet type data, urine discharge and total fecal water content of an individual through a processing module, constructing a real-time water metabolism balance model, and generating a balance state evaluation result; and P4, outputting a water intake quantity adjustment suggestion or a fecal detection frequency optimization scheme according to the evaluation result of the balance state of the input and output quantity.

Description

Automatic fecal water content detection system, method and input and output balance management method Technical Field The application relates to the technical field of intelligent detection, in particular to an automatic fecal water content detection system and method and an input and output balance management method. Background In clinical diagnosis and treatment and individual health management, the water content of feces is one of the key indexes for evaluating the water metabolism balance, digestive system functions and disease states of human bodies. In the existing detection of the water content of hospital excrement, the traditional drying method needs manual operation, is easy to produce problems of peculiar smell, visual pollution, cross infection risk and the like, and the simple visual identification method is difficult to meet the clinical data accuracy requirement because the excrement component is distributed discretely and the single-point visual observation can only reflect local appearance characteristics and cannot accurately capture the real difference of the whole water content. Therefore, finding an adaptive fecal moisture content detection system is a technical problem to be solved by those skilled in the art. Disclosure of Invention Based on this, in order to solve the problems of difficult operation and low accuracy of the fecal moisture content detection in the prior art, the application provides an automatic fecal moisture content detection system, an automatic fecal moisture content detection method and an input/output balance management method, and mainly aims to disclose an accurate and scientific input/output balance management method, which is an important link in the input/output balance management method, and the accurate and simple acquisition of the fecal moisture content is a key link to be solved urgently at present. In a first aspect, the present application provides an automated fecal moisture content detection system comprising: the split charging unit is used for obtaining and sealing a packaging bag filled with the fecal sample; the mixing unit is used for obtaining the packaging bag and carrying out homogenizing operation so as to uniformly distribute the fecal components; the sampling unit is used for quantitatively sampling the fecal samples which are uniformly distributed to obtain quantitative samples and injecting the quantitative samples into the detection bottle; the multi-mode detection unit is used for acquiring the net total weight of the fecal sample in the packaging bag through the weight detection unit, acquiring spectral characteristic data of the sample in the detection bottle through the near infrared spectrometer, detecting the temperature of the near infrared detection environment in real time through the temperature compensation device and correcting the temperature of the spectral characteristic data; The large model calculation unit inputs the spectral characteristic data after temperature correction into the pre-training model to obtain the water content of the fecal sample, and calculates the total water content of the fecal sample according to the water content and the total net weight of the fecal sample; and the data transmission unit is used for transmitting the total water content of the fecal sample in the packaging bag to the processing module of the input/output balance management system. Further, the step of obtaining the envelope and performing a homogenization operation to uniformly distribute the fecal component comprises: the packaging bag is placed into a negative pressure homogenizing cavity, a low-speed rotating motor is arranged in the negative pressure homogenizing cavity, and the packaging bag is driven to rotate through the low-speed rotating motor, so that the fecal components are uniformly distributed. Further, the packaging bag is provided with a sampling positioning area, and a high-elasticity thickened composite film or a high-elasticity reinforcing layer is embedded in the sampling positioning area to allow the sampling needle to be automatically closed after being pulled out and prevent the sample from leaking. Further, quantitatively sampling the fecal sample through the sampling positioning area by a disposable sampling needle; The disposable sampling needle comprises a self-sealing silica gel rebound cap, the self-sealing silica gel rebound cap comprises a silica gel main body, a built-in stainless steel rebound spring and a sealing lip, when the disposable sampling needle pierces a sampling positioning area of the packaging bag, the silica gel main body is subjected to pressure compression and drives the rebound spring to store energy, the rebound spring releases elasticity when the disposable sampling needle is pulled out to enable the rebound cap to reset, and the sealing lip seals the sampling hole to prevent excrement samples from leaking. Further, the step of homogenizing the packaging bag with the fecal sa