A method for detecting excrement hemoglobin content based on multispectrum

Abstract

The application discloses a kind of based on multispectral excrement hemoglobin content detection method, it is related to hemoglobin detection technical field.The method includes: obtaining the reflectivity sequence of target excrement sample;Reflectivity sequence includes the characteristic measured reflectivity of multiple hemoglobin characteristic wave bands, and the baseline measured reflectivity of multiple baseline reference wave bands;Based on the baseline measured reflectivity of multiple baseline reference wave bands, determine the first base estimate of multiple hemoglobin characteristic wave bands from base shape dictionary;Based on the hemoglobin fixed absorption ratio of multiple hemoglobin characteristic wave bands and characteristic measured reflectivity, the first base estimate of multiple hemoglobin characteristic wave bands is corrected, obtains the second base estimate of multiple hemoglobin characteristic wave bands;Based on the second base estimate of multiple hemoglobin characteristic wave bands, determine the target hemoglobin concentration of target excrement sample.The application can improve the detection precision of excrement hemoglobin content.

Description

Technical Field

[0001] This invention relates to the field of hemoglobin detection technology, specifically to a method for detecting hemoglobin content in excrement based on multispectral methods. Background Technology

[0002] Fecal hemoglobin testing is an important method for screening digestive tract diseases. Accurate detection of hemoglobin levels in feces is of crucial clinical significance for the early screening and diagnosis of related conditions. Currently, there is a high demand in clinical practice for the quantitative and accurate detection of fecal hemoglobin, necessitating highly efficient detection technologies that are interference-resistant, low-cost, and require no pretreatment.

[0003] Currently, the mainstream clinical testing methods are chemical test strips and immunochromatography, while multispectral detection technology is also gradually being applied. Chemical test strips rely on chemical reactions for detection, immunochromatography is based on the principle of immune recognition, while multispectral technology can calculate hemoglobin concentration by collecting multi-band reflectance spectra of samples and combining them with a multivariate calibration model, thus achieving label-free detection.

[0004] However, chemical test strips are easily affected by dietary factors, resulting in poor accuracy; immunochromatography is costly and cannot achieve quantitative detection. Conventional multispectral detection relies on fixed statistical models, which are difficult to adapt to the diverse morphologies of clinical excrement samples. Furthermore, the baseline differences in the excrement's base spectrum are much greater than the fluctuations in the hemoglobin signal, causing the hemoglobin signal to be submerged by baseline noise, leading to a significant decrease in detection accuracy. Summary of the Invention

[0005] To address the technical problem that existing methods suffer from a significant decrease in detection accuracy due to baseline noise overwhelming the hemoglobin signal, this invention provides a multispectral-based method for detecting hemoglobin content in excrement, which can improve the detection accuracy of hemoglobin content in excrement.

[0006] A first aspect of this invention provides a method for detecting hemoglobin content in excrement based on multispectral methods, comprising:

[0007] The reflectance sequence of the target excrement sample is obtained. The reflectance sequence includes the characteristic measured reflectance of the target excrement sample in multiple hemoglobin characteristic bands, and the baseline measured reflectance of the target excrement sample in multiple baseline reference bands. The hemoglobin characteristic bands refer to the bands that are sensitive to changes in hemoglobin concentration, and the baseline reference bands refer to the bands that are not sensitive to changes in hemoglobin concentration and whose reflectance only reflects the optical properties of the excrement substrate itself.

[0008] Based on the measured baseline reflectance of multiple baseline reference bands, the first substrate estimate of the target excrement sample in multiple hemoglobin characteristic bands is determined from the substrate shape dictionary; the substrate shape dictionary is used to characterize the proportional relationship between the reflectance of each hemoglobin characteristic band under different excrement substrate conditions.

[0009] Based on the fixed absorption ratio of hemoglobin in multiple hemoglobin characteristic bands and the measured characteristic reflectance of multiple hemoglobin characteristic bands, the first basis estimate of multiple hemoglobin characteristic bands is corrected to obtain the second basis estimate of multiple hemoglobin characteristic bands.

[0010] The target hemoglobin concentration of the target excrement sample is determined based on the second basis estimate of multiple hemoglobin characteristic bands.

[0011] Furthermore, the present invention also proposes obtaining the reflectance sequence of a target excrement sample, including:

[0012] Obtain the dark current value, median light intensity, and reference light intensity of the target excrement sample in the target wavelength band; the target wavelength band is any one of multiple hemoglobin characteristic wavelength bands and multiple baseline reference wavelength bands;

[0013] Based on the first difference between the median light intensity and the dark current value, and the second difference between the reference light intensity and the dark current value, the measured reflectance of the target excrement sample in the target band is determined.

[0014] Based on the measured reflectance of the target excrement sample in each target band, the reflectance sequence of the target excrement sample is determined.

[0015] Furthermore, the present invention also proposes that the measured baseline reflectance of multiple baseline reference bands includes the first measured baseline reflectance of a first baseline reference band and the second measured baseline reflectance of a second baseline reference band, wherein the first baseline reference band is located on the shortwave side of the second baseline reference band.

[0016] Based on the measured baseline reflectance of multiple baseline reference bands, the first basis estimate of the target excrement sample in multiple hemoglobin characteristic bands is determined from the basis shape dictionary, including:

[0017] Based on the measured reflectance of the first baseline and the measured reflectance of the second baseline, the target wavelength ratio of the target excrement sample is determined.

[0018] The target wavelength ratio is matched with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands; the target feature ratio is used to characterize the reflectance level of the substrate of the target excrement sample in the hemoglobin feature bands relative to the first baseline reference band;

[0019] Based on the measured reflectance of the first baseline and the target feature ratio of multiple hemoglobin feature bands, the first baseline estimate of multiple hemoglobin feature bands is determined.

[0020] Furthermore, the present invention also proposes to match the target wavelength ratio with a base shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin characteristic bands, including:

[0021] The target wavelength ratio is matched with the base shape dictionary to obtain K candidate feature ratio groups that are closest to the target wavelength ratio; the candidate feature ratio groups include candidate feature ratios in multiple hemoglobin feature bands, where K is a positive integer;

[0022] Based on the K groups of candidate feature proportions, the target feature proportions of multiple hemoglobin feature bands are determined.

[0023] Furthermore, the present invention also proposes, before matching the target wavelength ratio with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands, the method further includes:

[0024] The reference wavelength ratio of the reference excrement sample is determined based on the first reference reflectance of the first baseline reference band and the second reference reflectance of the second baseline reference band.

[0025] Based on the reference reflectance of the reference excrement sample in multiple hemoglobin characteristic bands, and the first reference reflectance, the reference feature ratio of the reference excrement sample in multiple hemoglobin characteristic bands is determined.

[0026] A base shape dictionary is constructed based on the reference wavelength ratio of each reference excrement sample and the reference feature ratio of multiple hemoglobin characteristic bands.

[0027] Furthermore, the present invention also proposes to correct the first basis estimate of multiple hemoglobin characteristic bands based on the fixed absorption ratio of hemoglobin in multiple hemoglobin characteristic bands and the measured characteristic reflectance of multiple hemoglobin characteristic bands, to obtain the second basis estimate of multiple hemoglobin characteristic bands, including:

[0028] Based on the fixed absorption ratio of hemoglobin, the measured characteristic reflectance, and the first basis estimate, the candidate hemoglobin characteristic intensities of multiple hemoglobin characteristic bands are determined.

[0029] In response to the fact that the candidate hemoglobin feature intensities of multiple hemoglobin feature bands do not meet the preset convergence conditions, the first basis estimates of multiple hemoglobin feature bands are corrected until the candidate hemoglobin feature intensities of multiple hemoglobin feature bands meet the preset convergence conditions, and the second basis estimates of multiple hemoglobin feature bands are obtained.

[0030] Furthermore, the present invention also proposes to correct the first basis estimate of multiple hemoglobin characteristic bands, including:

[0031] Obtain the feature intensity of candidate hemoglobins in multiple hemoglobin feature bands and the feature intensity difference between them and the median of hemoglobin feature intensity;

[0032] Based on the characteristic intensity difference, the base correction direction of the characteristic band of hemoglobin is determined;

[0033] Based on the basis correction direction, the first basis estimate of the hemoglobin characteristic band is corrected.

[0034] Furthermore, the present invention also proposes that the first baseline estimate includes a first band baseline estimate of a first hemoglobin characteristic band, a second band baseline estimate of a second hemoglobin characteristic band, and a third band baseline estimate of a third hemoglobin characteristic band; the first hemoglobin characteristic band is located on the shortwave side of the second hemoglobin characteristic band, and the second hemoglobin characteristic band is located on the shortwave side of the third hemoglobin characteristic band;

[0035] Based on the basis correction direction, the first basis estimate of the hemoglobin characteristic bands is corrected, including:

[0036] Based on the first characteristic intensity difference of the first hemoglobin characteristic band and the third characteristic intensity difference of the third hemoglobin characteristic band, the expected characteristic intensity difference of the second hemoglobin characteristic band is determined.

[0037] Based on the expected feature intensity difference of the second hemoglobin feature band and the second feature intensity difference of the second hemoglobin feature band, the band smoothing coefficient of the second hemoglobin feature band is determined.

[0038] Based on the baseline estimates of the first and third bands, the expected band baseline estimates of the second hemoglobin characteristic bands are determined.

[0039] The second band basis estimate is corrected based on the band smoothing coefficient and the expected band basis estimate.

[0040] Furthermore, the present invention also proposes determining the target hemoglobin concentration of a target excrement sample based on a second basis estimate of multiple hemoglobin characteristic bands, including:

[0041] Based on the second basis estimate of multiple hemoglobin characteristic bands, the target hemoglobin characteristic intensity of multiple hemoglobin characteristic bands is determined;

[0042] The target comprehensive hemoglobin characteristic intensity of the target excrement sample is obtained by weighted summation of the target hemoglobin characteristic intensities of multiple hemoglobin characteristic bands.

[0043] The target hemoglobin characteristic intensity is matched with the target concentration curve to obtain the target hemoglobin concentration of the target excrement sample; the target concentration curve is used to characterize the mapping relationship between the comprehensive hemoglobin characteristic intensity and the hemoglobin concentration of the excrement sample.

[0044] Furthermore, the present invention also proposes that the target concentration curve includes concentration sub-curves of various excrement forms;

[0045] By matching the target hemoglobin characteristic intensity with the target concentration curve, the target hemoglobin concentration of the target excrement sample is obtained, including:

[0046] Obtain the target excrement morphology from the target excrement sample;

[0047] The target hemoglobin concentration is obtained by matching the target comprehensive hemoglobin characteristic intensity with the target concentration sub-curve of the target excrement morphology.

[0048] A second aspect of the present invention provides a multispectral-based fecal hemoglobin content detection system, comprising: a memory and a processor; the memory being connected to the processor; the memory being used to store program instructions; and the processor being used to implement the above-described multispectral-based fecal hemoglobin content detection method when the program instructions are executed.

[0049] The present invention has the following beneficial effects: In the multispectral-based method for detecting hemoglobin content in excrement provided in the embodiments of the present invention, firstly, the characteristic measured reflectance of the hemoglobin characteristic band, which is sensitive to hemoglobin concentration, is distinguished and collected, as well as the baseline measured reflectance of the baseline reference band, which only reflects the optical properties of the excrement substrate and is not affected by hemoglobin concentration, thereby achieving physical separation between the target signal and the substrate interference signal; secondly, relying on the substrate shape dictionary that can characterize the reflectance ratio of the hemoglobin characteristic band under different excrement substrate conditions, the hemoglobin content is accurately obtained by combining the baseline reference band data. The first baseline estimate of the characteristic bands is adapted to the variable baseline state of clinical excrement, solving the shortcomings of traditional fixed statistical models that cannot adapt to the baseline distribution of multiple samples and the baseline prediction distortion. Finally, the first baseline estimate is corrected a second time by using the fixed absorption ratio of hemoglobin combined with the measured reflectance of the characteristic bands of hemoglobin, resulting in a high-precision second baseline estimate. This can accurately separate the small effective signal of hemoglobin from strong baseline noise. Finally, the hemoglobin concentration is calculated based on the accurately corrected second baseline estimate, which significantly improves the accuracy and precision of excrement hemoglobin content detection. Attached Figure Description

[0050] Figure 1This is a schematic flowchart of a method for detecting hemoglobin content in excrement based on multispectral methods provided by the present invention. Detailed Implementation

[0051] In traditional multispectral detection techniques, the baseline difference in the excrement substrate spectral density exceeds the fluctuation range of the hemoglobin characteristic signal, causing the hemoglobin characteristic signal to be submerged by substrate noise. This problem arises from the variable state of excrement samples, making it difficult for fixed statistical models to separate the hemoglobin signal from the substrate background, thus reducing detection accuracy and reliability. Specifically, the significant differences in the optical properties of the excrement substrate affect the stability of reflectance measurements; the signal fluctuations in the hemoglobin characteristic bands are obscured by the substrate background, making it impossible to effectively extract target information.

[0052] Based on this, the present invention provides a specific embodiment of a multispectral method for detecting excrement hemoglobin content, which can improve the detection accuracy of excrement hemoglobin content.

[0053] like Figure 1 The diagram shows a flowchart of a method for detecting fecal hemoglobin content based on multispectral methods. This method can be applied to electronic devices and includes the following steps S110 to S140:

[0054] S110, Obtain the reflectance sequence of the target excrement sample; the reflectance sequence includes the characteristic measured reflectance of the target excrement sample in multiple hemoglobin characteristic bands, and the baseline measured reflectance of the target excrement sample in multiple baseline reference bands; the hemoglobin characteristic bands refer to bands that are sensitive to changes in hemoglobin concentration, and the baseline reference bands refer to bands that are not sensitive to changes in hemoglobin concentration and whose reflectance only reflects the optical properties of the excrement substrate itself;

[0055] S120, based on the measured baseline reflectance of multiple baseline reference bands, determines the first substrate estimate of the target excrement sample in multiple hemoglobin characteristic bands from the substrate shape dictionary; the substrate shape dictionary is used to characterize the proportional relationship between the reflectance of each hemoglobin characteristic band under different excrement substrate conditions.

[0056] S130, based on the fixed absorption ratio of hemoglobin in multiple hemoglobin characteristic bands and the measured characteristic reflectance of multiple hemoglobin characteristic bands, the first basis estimate of multiple hemoglobin characteristic bands is corrected to obtain the second basis estimate of multiple hemoglobin characteristic bands.

[0057] S140, based on the second basis estimate of multiple hemoglobin characteristic bands, determines the target hemoglobin concentration of the target excrement sample.

[0058] For ease of understanding, the following explains some key terms in this embodiment:

[0059] Multispectral methods for detecting hemoglobin content in excrement refer to a technical solution that uses multispectral technology to quantitatively detect the hemoglobin content in excrement samples by analyzing the reflectance spectral data at different wavelengths. This method aims to overcome the limitations of traditional detection methods and achieve accurate and efficient detection of hemoglobin in excrement.

[0060] A reflectance sequence refers to a series of reflectance data obtained by performing spectral measurements on a target excrement sample within a specific wavelength range. These data reflect the sample's ability to reflect light at different wavelengths and form the basis for subsequent hemoglobin content analysis.

[0061] The characteristic wavelength bands of hemoglobin refer to the wavelengths within the spectral range where hemoglobin exhibits significant absorption or reflection of light. Changes in reflectance in these bands show a sensitive correlation with changes in hemoglobin concentration, and are therefore selected as key regions for analyzing hemoglobin signals. For example, three characteristic absorption peaks of hemoglobin (415 nm, 540 nm, and 575 nm) can be selected to capture the target signal, i.e., as the characteristic wavelength bands of hemoglobin.

[0062] A baseline reference band refers to a band within the spectral range where the excrement substrate exhibits relatively stable light reflectance, and changes in reflectance within these bands are insensitive to changes in hemoglobin concentration. The reflectance of the baseline reference band primarily reflects the optical properties of the excrement substrate itself and is used to separate and eliminate substrate interference with the hemoglobin signal. For example, two reference wavelengths (650 nm and 780 nm) that do not respond to hemoglobin but reflect the substrate state can be selected for subsequent baseline estimation, i.e., as the baseline reference band.

[0063] A substrate shape dictionary is a pre-built dataset that stores the proportional relationships between the reflectance of various hemoglobin characteristic bands under different substrate conditions of excrement. This substrate shape dictionary is used to characterize the inherent morphology of the excrement substrate spectrum, enabling accurate estimation of substrate effects during detection.

[0064] The first basis estimate refers to the basis reflectance of the target excrement sample in the hemoglobin characteristic band, calculated from the basis shape dictionary based on the measured baseline reflectance of the baseline reference band during the preliminary analysis stage. This estimate serves as the initial reference for subsequent correction of the hemoglobin signal.

[0065] The fixed absorption ratio of hemoglobin refers to the fixed proportional relationship between the absorption coefficients of hemoglobin in different characteristic wavelength bands of hemoglobin. This fixed absorption ratio is an inherent optical property of hemoglobin and is used to quantify the contribution of hemoglobin to the spectrum during correction processes.

[0066] The second basis estimate is a more accurate estimate of basis reflectance obtained after correcting the first basis estimate. This correction process takes into account the actual absorption effect of hemoglobin, enabling the second basis estimate to more accurately reflect the true spectral characteristics of the excrement basis.

[0067] The target hemoglobin concentration refers to the quantitative value of hemoglobin contained in the target excrement sample, determined through the aforementioned series of processes and analyses. This concentration value is an important basis for diagnosing gastrointestinal diseases.

[0068] This embodiment provides a multispectral-based method for detecting hemoglobin content in excrement. This method aims to solve the problem of interference of excrement substrate spectral differences on hemoglobin signal detection in the prior art, thereby improving the accuracy and reliability of detection.

[0069] First, the reflectance sequence of the target excrement sample is obtained. This reflectance sequence includes the characteristic measured reflectance of the target excrement sample in multiple hemoglobin characteristic bands, and the baseline measured reflectance of the target excrement sample in multiple baseline reference bands. Hemoglobin characteristic bands refer to bands sensitive to changes in hemoglobin concentration, while baseline reference bands are bands insensitive to changes in hemoglobin concentration, where reflectance only reflects the optical properties of the excrement substrate itself. In practice, the target excrement sample can be directly scanned using a spectrometer to obtain its reflectance spectral data within a preset wavelength range. For example, the target excrement sample can be placed on the measurement platform of the spectrometer, illuminated by a light source, and the reflected light can be received. The received light signal is then converted into an electrical signal and further processed to obtain the reflectance of each band.

[0070] Secondly, based on the measured baseline reflectance of the multiple baseline reference bands, the first substrate estimate of the target excrement sample in the multiple hemoglobin characteristic bands is determined from the substrate shape dictionary. This substrate shape dictionary characterizes the proportional relationship between the reflectance of each hemoglobin characteristic band under different excrement substrate states. In practice, a database containing multiple typical excrement substrate states can be pre-established as the substrate shape dictionary. After obtaining the measured baseline reflectance of the target excrement sample, the reflectance ratios between these measured baseline reflectances can be calculated and compared with the pre-stored data in the substrate shape dictionary. For example, the reflectance ratios between the measured baseline reflectances of the baseline reference bands can be calculated, and using this as a benchmark, the substrate state closest to this reflectance ratio can be found in the substrate shape dictionary, thereby obtaining the proportional relationship of the reflectance of this substrate state in the hemoglobin characteristic bands. Then, by further combining the measured baseline reflectance of the target excrement sample in the baseline reference bands, its first substrate estimate in the hemoglobin characteristic bands is calculated.

[0071] Next, based on the fixed absorption ratio of hemoglobin in the multiple hemoglobin characteristic bands and the measured characteristic reflectance of these multiple hemoglobin characteristic bands, the first basis estimate of the multiple hemoglobin characteristic bands is corrected to obtain the second basis estimate of the multiple hemoglobin characteristic bands. For example, by continuously adjusting the first basis estimate until a preset convergence condition is reached, the preset convergence condition is determined by the fixed absorption ratio of hemoglobin in the hemoglobin characteristic bands and the measured characteristic reflectance of the multiple hemoglobin characteristic bands.

[0072] Finally, based on the second basis estimates of these multiple hemoglobin characteristic bands, the target hemoglobin concentration of the target excrement sample is determined. After obtaining accurate second basis estimates, the spectral signal purely caused by hemoglobin absorption can be derived from them. For example, this can be achieved through conversion using a pre-established standard curve or model, thereby determining the target hemoglobin concentration of the target excrement sample based on the second basis estimates of multiple hemoglobin characteristic bands.

[0073] This embodiment first distinguishes and collects the measured reflectance of the characteristic bands of hemoglobin, which are sensitive to hemoglobin concentration, and the measured reflectance of the baseline reference bands, which only reflect the optical properties of the excrement substrate and are unaffected by hemoglobin concentration, thus achieving physical separation between the target signal and the substrate interference signal. Second, relying on a substrate shape dictionary that characterizes the reflectance ratio of the characteristic bands of hemoglobin under different excrement substrate conditions, a first substrate estimate of the characteristic bands of hemoglobin is accurately fitted using baseline reference band data. This adapts to the variable substrate conditions of clinical excrement and solves the shortcomings of traditional fixed statistical models that cannot adapt to multi-sample baseline distributions and suffer from baseline prediction distortion. Finally, using the fixed absorption ratio of hemoglobin combined with the measured reflectance of the characteristic bands of hemoglobin, the first substrate estimate is corrected a second time to obtain a high-precision second substrate estimate. This allows for the accurate separation of the small effective signal of hemoglobin from strong baseline noise. Finally, the hemoglobin concentration is calculated based on the accurately corrected second substrate estimate, significantly improving the accuracy and precision of excrement hemoglobin content detection.

[0074] In some embodiments, the present invention further proposes that the process of obtaining the reflectance sequence of the target excrement sample in step S110 above specifically includes:

[0075] Obtain the dark current value, median light intensity, and reference light intensity of the target excrement sample in the target wavelength band; the target wavelength band is any one of multiple hemoglobin characteristic wavelength bands and multiple baseline reference wavelength bands;

[0076] Based on the first difference between the median light intensity and the dark current value, and the second difference between the reference light intensity and the dark current value, the measured reflectance of the target excrement sample in the target band is determined.

[0077] Based on the measured reflectance of the target excrement sample in each target band, the reflectance sequence of the target excrement sample is determined.

[0078] In this embodiment, the dark current value refers to the current signal generated by the sensor itself when there is no light input, which represents the inherent noise of the system. The dark current value can be obtained by taking multiple readings of the sensor in a completely dark environment and taking the average or median value.

[0079] The median light intensity refers to the value taken as the middle position after multiple light intensity measurements of a target excrement sample in a specific target wavelength band are performed, sorted, and then used. Using the median instead of the average value can effectively reduce the impact of occasional noise or transient interference on the light intensity readings during the measurement process, thereby improving the stability of the measurement results.

[0080] Reference light intensity refers to the value obtained by measuring the light intensity of a standard reference object with known reflectivity (e.g., a standard white board) while simultaneously measuring the light intensity reflected from a target excrement sample. Reference light intensity is used to calibrate for fluctuations in light source intensity and changes in sensor response, ensuring the accuracy and comparability of measurement results.

[0081] The target band refers to the specific wavelength range for which reflectance measurement is required. It may include the hemoglobin characteristic band, which is sensitive to changes in hemoglobin concentration, and the baseline reference band, which is not sensitive to changes in hemoglobin concentration and only reflects the optical properties of the excrement substrate itself.

[0082] This invention systematically considers various influencing factors during the measurement process to ensure the accuracy of reflectance data. First, by acquiring the dark current value, the inherent noise of the sensor can be accurately subtracted, ensuring that subsequent light intensity calculations are based on the true net light signal. Second, using the median light intensity instead of a simple average effectively suppresses transient noise or outliers that may occur during measurement, thus obtaining a more representative sample reflected light intensity. Simultaneously, by acquiring a reference light intensity, real-time calibration can be performed to account for external environmental factors such as fluctuations in light source intensity and uneven sensor response, ensuring the comparability of data under different measurement conditions. Based on these corrected light intensity data—the first difference between the median light intensity and the dark current value, and the second difference between the reference light intensity and the dark current value—the measured reflectance of the target excrement sample in each target wavelength band can be accurately calculated. This calculation method eliminates errors caused by changes in ambient light and equipment drift, resulting in stable and reliable measured reflectance. Finally, by combining the measured reflectance of all target bands in wavelength order, an accurate reflectance sequence is formed, providing high-quality input data for subsequent multispectral hemoglobin content detection, and significantly improving the reliability and accuracy of the entire detection method.

[0083] As an example, for each target wavelength band, with the light source off, the dark current intensity is continuously collected 10 times using a photodiode, and the arithmetic mean is taken as the dark current value of that target wavelength band. During measurement, the target excrement sample is laid flat on the measurement window, and the multispectral probe is vertically aligned with the sample surface. Three different positions (at least 5 mm apart) are randomly selected on the target excrement sample, and the LED light sources of five target wavelength bands are turned on sequentially at each position. The reflected light intensity is received by the photodiode. The arithmetic mean is taken three times for each position of each target wavelength band to obtain the reflected light intensity count of that position for that target wavelength band. The median of the measurement results at the above three positions is taken for each target wavelength band to obtain the median light intensity of the target excrement sample in that target wavelength band. The reflectance is calculated using a standard white board for calibration. Before measurement, the standard white board (reflectance ≈ 0.99) is calibrated, and the average value is taken three times for each of the five target wavelength bands to obtain the reference light intensity of the target excrement sample in that target wavelength band.

[0084] Finally, the measured reflectance of the target excrement sample in the target wavelength band is determined using the following formula 1:

[0085] Formula 1

[0086] In formula 1, The measured reflectance of the target excrement sample in the j-th target band is used to characterize the actual reflectance. This is used to characterize the median light intensity of the target excrement sample in the j-th target wavelength band. Used to characterize the dark current value of the target excrement sample in the j-th target wavelength band. The reference light intensity used to characterize the target excrement sample in the j-th target wavelength band. The preset parameter adjustment factor is greater than zero (for example, it can be 0.001). By multiplying the parameter adjustment factor with the dark current value which has the dimension of light intensity, a compensation term with the dimension of light intensity is obtained. Then, it is added to the original light intensity difference in the denominator. This avoids the denominator being 0 while ensuring that the dimensions of all calculation terms are consistent.

[0087] Through the above technical solution, this invention effectively eliminates the influence of adverse factors such as ambient light fluctuations, sensor dark current, and instantaneous measurement noise when acquiring the reflectance sequence of target excrement samples. This makes the acquired reflectance data more accurate and stable, providing high-quality input for subsequent hemoglobin concentration calculations. Therefore, this invention can significantly improve the overall accuracy and reliability of multispectral-based excrement hemoglobin content detection methods, thereby enhancing the accuracy of diagnostic results.

[0088] In some embodiments, the present invention further proposes that the measured baseline reflectance of multiple baseline reference bands includes the measured baseline reflectance of a first baseline reference band and the measured baseline reflectance of a second baseline reference band, wherein the first baseline reference band is located on the shortwave side of the second baseline reference band.

[0089] The process of determining the first substrate estimate of the target excrement sample in multiple hemoglobin characteristic bands based on the measured baseline reflectance of multiple baseline reference bands in step S120 above specifically includes:

[0090] Based on the measured reflectance of the first baseline and the measured reflectance of the second baseline, the target wavelength ratio of the target excrement sample is determined.

[0091] The target wavelength ratio is matched with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands; the target feature ratio is used to characterize the reflectance level of the substrate of the target excrement sample in the hemoglobin feature bands relative to the first baseline reference band;

[0092] Based on the measured reflectance of the first baseline and the target feature ratio of multiple hemoglobin feature bands, the first baseline estimate of multiple hemoglobin feature bands is determined.

[0093] In this embodiment, the first baseline reference band and the second baseline reference band are specific spectral regions used to characterize the optical properties of the excrement substrate. They are selected as bands insensitive to changes in hemoglobin concentration, and their reflectance primarily reflects the inherent optical properties of the excrement substrate itself. The first baseline reference band is located on the shorter wavelength side of the second baseline reference band, meaning that the first baseline reference band has a shorter wavelength. For example, the first baseline reference band could be selected as 650 nm, and the second baseline reference band could be selected as 780 nm.

[0094] The target wavelength ratio is a ratio calculated based on the measured reflectance of a first baseline and the measured reflectance of a second baseline (e.g., dividing the measured reflectance of the second baseline by the measured reflectance of the first baseline). This ratio effectively quantifies the reflectance variation trend of the excrement substrate within a specific spectral region, i.e., the "shape" characteristic of the substrate. By calculating the wavelength ratio, complex baseline reflectance data can be simplified into a representative value, facilitating subsequent matching with a substrate shape dictionary.

[0095] The substrate shape dictionary is a pre-built database that stores the ratio of reflectance between the first baseline reference band and the hemoglobin characteristic band under different excrement substrate conditions. By matching the calculated target wavelength ratio with entries in the dictionary, the data that best represents the substrate characteristics of the current target excrement sample can be found. Upon successful matching, the target feature ratio is obtained, which reflects the reflectance level of the target excrement sample's substrate in the hemoglobin characteristic band relative to the reflectance level of the first baseline reference band.

[0096] Once the target feature ratio is obtained, the first basis estimate for each hemoglobin characteristic band can be calculated by combining it with the measured reflectance of the first baseline. Since the target feature ratio characterizes the reflectance level of the hemoglobin characteristic band relative to the first baseline reference band, a simple multiplication operation can be performed to multiply the measured reflectance of the first baseline by the target feature ratio, thereby obtaining the basis reflectance estimate for each hemoglobin characteristic band without the influence of hemoglobin absorption. For example, if the target feature ratio is the ratio of the reflectance of the hemoglobin characteristic band to the reflectance of the first baseline reference band, then the first basis estimate can be obtained by multiplying the measured reflectance of the first baseline by the corresponding target feature ratio.

[0097] The present invention introduces a first baseline reference band and a second baseline reference band, specifying that the first baseline reference band is located on the shortwave side of the second baseline reference band. This allows for the calculation of the target wavelength ratio (WVR) of the target excrement sample based on the measured reflectance of these two bands. This target WVR serves as a quantitative indicator of the optical properties of the excrement substrate, providing a more precise characterization of the substrate shape. Subsequently, this target WVR is matched with a pre-constructed substrate shape dictionary to accurately obtain the target feature proportions of the target excrement sample in multiple hemoglobin characteristic bands. These target feature proportions directly reflect the reflectance level of the substrate in the hemoglobin characteristic bands relative to the first baseline reference band. Finally, by combining the measured reflectance of the first baseline and the obtained target feature proportions, the first substrate estimate for multiple hemoglobin characteristic bands can be accurately determined. This method avoids fuzzy matching by directly using the raw reflectance of multiple baseline reference bands. Instead, it introduces two baseline reference bands with a spectral positional relationship and calculates their ratio, thereby providing a more discriminative feature. This makes the base shape information retrieved from the base shape dictionary more accurate, thus improving the accuracy of the first base estimate.

[0098] Through the above technical solution, this invention introduces a first baseline reference band and a second baseline reference band with a specific spectral positional relationship, and calculates the target wavelength ratio based on their respective measured reflectance, thereby providing a more representative and discriminative substrate shape feature. This quantification method enables more precise location of the entry that best matches the substrate characteristics of the target excrement sample when matching from the substrate shape dictionary, thus obtaining a more accurate target feature ratio. Ultimately, the accuracy of the first substrate estimate determined based on the accurate target feature ratio and the measured reflectance of the first baseline is significantly improved, effectively solving the technical problem of difficulty in accurately obtaining substrate estimates under complex and variable excrement substrate optical characteristics, and laying a solid foundation for the accurate determination of subsequent hemoglobin concentration.

[0099] In some embodiments, the present invention further proposes matching the target wavelength ratio with a substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands, specifically including:

[0100] The target wavelength ratio is matched with the base shape dictionary to obtain K candidate feature ratio groups that are closest to the target wavelength ratio; the candidate feature ratio groups include candidate feature ratios in multiple hemoglobin feature bands, where K is a positive integer;

[0101] Based on the K groups of candidate feature proportions, the target feature proportions of multiple hemoglobin feature bands are determined.

[0102] In this embodiment, the step of "matching the target wavelength ratio with the base shape dictionary to obtain K candidate feature ratio groups that are closest to the target wavelength ratio; the candidate feature ratio groups include candidate feature ratios in multiple hemoglobin characteristic bands, where K is a positive integer" aims to find multiple base shape features from the base shape dictionary that are most similar to the target wavelength ratio of the current target excrement sample. By not limiting itself to a single best match, but obtaining a set of K candidate results with the highest similarity, the uncertainty caused by measurement error and sample diversity can be effectively addressed. This step can be achieved by calculating the absolute value of the difference between the target wavelength ratio and all reference wavelength ratios in the base shape dictionary, and then selecting the K reference wavelength ratios with the smallest absolute difference and their corresponding feature ratio groups as candidate feature ratio groups. K represents the number of candidate feature ratio groups, which is generally 5-15 (the principle is to ensure that the standard deviation of the reference wavelength ratio of each selected candidate feature ratio group does not exceed 0.05; if there is no prior information, K=7 is used by default).

[0103] The step of "determining the target feature ratio of multiple hemoglobin feature bands based on K sets of candidate feature ratios" involves comprehensively processing the K sets of candidate feature ratios obtained from the basis shape dictionary to obtain a more stable and accurate final target feature ratio. By fusing multiple similar results, the random error of a single matching result can be reduced, and the accuracy of basis estimation can be improved. This step can average the candidate feature ratios of each hemoglobin feature band in the K sets of candidate feature ratios, such as arithmetic average or weighted average, to obtain the final target feature ratio of that hemoglobin feature band. In the weighted average, the weights can be determined by normalizing the matching degree between the candidate feature ratio set and the target wavelength ratio (e.g., the reciprocal of the absolute value of the difference), where a preset parameter adjustment factor greater than zero can be added to the denominator (for example, a value of 0.001) to prevent the denominator from being 0.

[0104] The present invention, when determining the target feature ratio of a target excrement sample across multiple hemoglobin characteristic bands, no longer relies solely on a single match result closest to the target wavelength ratio. Instead, it first identifies K candidate feature ratio groups from a substrate shape dictionary that are closest to the target wavelength ratio. These candidate groups represent multiple possibilities for the reflectance levels of hemoglobin characteristic bands relative to a first baseline reference band under different excrement substrate conditions. By acquiring a set rather than a single value, this approach can more comprehensively capture the potential characteristics of the target excrement sample substrate and effectively address uncertainties caused by measurement noise or sample heterogeneity. Subsequently, based on these K candidate feature ratio groups, the approach further determines the final target feature ratio through comprehensive processing (e.g., averaging, median, or weighted fusion). This processing mechanism smooths out potential errors from single matches, enhancing the robustness and accuracy of the results. In this way, the approach can more accurately characterize the reflectance level of the target excrement sample substrate in the hemoglobin characteristic bands, laying a solid foundation for the accurate separation and concentration determination of subsequent hemoglobin absorption, thereby improving the reliability of the entire detection method.

[0105] The above technical solution eliminates the reliance on a single best match when determining the reflectance level of the substrate in the hemoglobin characteristic band of a target excrement sample. Instead, this solution acquires K candidate feature ratio groups closest to the target wavelength ratio and performs comprehensive processing based on these candidate groups. This effectively reduces matching inaccuracies caused by measurement noise, individual sample differences, or limited substrate shape dictionaries. This multi-candidate fusion strategy significantly improves the robustness and accuracy of target feature ratio determination, thereby providing a more reliable substrate estimate for subsequent precise separation of hemoglobin absorption and ultimately enhancing the overall accuracy and stability of excrement hemoglobin content detection.

[0106] In some embodiments, the present invention further includes, before matching the target wavelength ratio with a substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands:

[0107] The reference wavelength ratio of the reference excrement sample is determined based on the first reference reflectance of the first baseline reference band and the second reference reflectance of the second baseline reference band.

[0108] Based on the reference reflectance of the reference excrement sample in multiple hemoglobin characteristic bands, and the first reference reflectance, the reference feature ratio of the reference excrement sample in multiple hemoglobin characteristic bands is determined.

[0109] A base shape dictionary is constructed based on the reference wavelength ratio of each reference excrement sample and the reference feature ratio of multiple hemoglobin characteristic bands.

[0110] In this embodiment, the reference excrement sample refers to the data sample used to construct the base shape dictionary, whose characteristics (such as base state, hemoglobin concentration, etc.) are known or controllable. For example, a large number (e.g., 200-300) of bloodless excrement samples from healthy individuals can be collected, covering various common forms such as loose stools, formed stools, hard stools, and watery stools. To ensure that the samples are bloodless, each sample is first tested using a clinical chemiluminescence immunoassay (reference method, sensitivity 0.05 mg / g) to confirm that the hemoglobin concentration is below 0.05 mg / g. Samples with abnormally low reflectance in the measured spectrum (e.g., reflectance at 415 nm is 2 times lower than the mean of all samples of the same morphology) are considered to be possibly bloody and are removed from the base shape dictionary.

[0111] The first and second reference reflectance refer to the reflectance values ​​measured on a reference excrement sample at a specific baseline reference band (i.e., a band insensitive to changes in hemoglobin concentration). These reflectance values ​​are used to characterize the optical properties of the excrement substrate itself and are the basic data for calculating the reference wavelength ratio.

[0112] The reference longwave ratio refers to the ratio between the second reference reflectance of a reference excrement sample in the second baseline reference band and the first reference reflectance in the first baseline reference band. This ratio is used to quantify the spectral shape characteristics of the excrement substrate in the longwave direction, serving as an index or classification basis for a substrate shape dictionary. It can be determined by directly dividing the second reference reflectance by the first reference reflectance.

[0113] The reference feature ratio refers to the ratio between the reference reflectance of a reference excrement sample in multiple hemoglobin feature bands and the first reference reflectance in a first baseline reference band. This ratio characterizes the relative level of the reflectance of the hemoglobin feature bands relative to the baseline reference band under a specific substrate condition and is core data in the substrate shape dictionary. Specifically, for each hemoglobin feature band, its reference reflectance can be divided by the first reference reflectance to obtain the corresponding reference feature ratio.

[0114] A substrate shape dictionary is a data structure that stores the mapping relationship between the reference wavelength ratio and the reference feature proportions of multiple hemoglobin characteristic bands for different substrate conditions of excrement. This dictionary is used in actual detection to quickly look up and obtain the substrate reflectance ratio of a target excrement sample in the hemoglobin characteristic bands based on its target wavelength ratio, thereby accurately estimating the substrate reflectance. This dictionary can be a look-up table (LUT), where the key is the reference wavelength ratio and the corresponding value is the reference feature proportion of multiple hemoglobin characteristic bands.

[0115] The present invention effectively solves the problem of unclear source of the substrate shape dictionary when determining the first substrate estimate by introducing a substrate shape dictionary construction process. Specifically, before detecting the hemoglobin content of the target excrement sample, a series of representative reference excrement samples are first collected, and multispectral reflectance measurements are performed on these samples. For each reference excrement sample, its first reference reflectance in the first baseline reference band and its second reference reflectance in the second baseline reference band are obtained, and then the reference wavelength ratio characterizing its substrate spectral shape is calculated. At the same time, the reference reflectance of the sample in multiple hemoglobin characteristic bands is obtained, and the reference feature ratio reflecting the reflectance level of the hemoglobin characteristic bands relative to the baseline band is calculated by combining the first reference reflectance. By associating and storing the reference wavelength ratios of a large number of reference excrement samples with different substrate states and their corresponding reference feature ratios, a comprehensive and accurate substrate shape dictionary can be constructed. This dictionary can characterize the proportional relationship between the reflectance of each hemoglobin characteristic band under different excrement substrate states. When actually testing target excrement samples, this pre-constructed base shape dictionary can be used to match or deduce the base reflectance ratio in the hemoglobin characteristic band based on the target wavelength ratio of the excrement sample itself. This provides a reliable basis for the subsequent calculation of the first base estimate. This pre-constructed dictionary method enables more accurate matching results when determining the target feature ratio from the base shape dictionary based on the target wavelength ratio, thereby improving the accuracy of the first base estimate and laying the foundation for the accurate determination of the final hemoglobin concentration.

[0116] Through the above technical solution, this invention provides a systematic and reliable method for constructing a substrate shape dictionary. By utilizing a series of reference excrement samples, and determining the reference wavelength ratio based on their reflectance in the baseline reference band, and determining the reference feature ratio based on their reflectance in the hemoglobin characteristic band, a substrate shape dictionary that comprehensively reflects the spectral characteristics of different excrement substrate states can be constructed. This enables the accurate matching or deduction of the substrate reflectance ratio in the hemoglobin characteristic band from the pre-constructed dictionary when subsequently detecting target excrement samples based on their own spectral characteristics, thereby significantly improving the accuracy of the first substrate estimate.

[0117] In some embodiments, the present invention further proposes that the process of correcting the first basis estimate of multiple hemoglobin characteristic bands based on the fixed absorption ratio of hemoglobin in multiple hemoglobin characteristic bands and the measured characteristic reflectance of multiple hemoglobin characteristic bands in step S130 above, to obtain the second basis estimate of multiple hemoglobin characteristic bands, specifically includes:

[0118] Based on the fixed absorption ratio of hemoglobin, the measured characteristic reflectance, and the first basis estimate, the candidate hemoglobin characteristic intensities of multiple hemoglobin characteristic bands are determined.

[0119] In response to the fact that the candidate hemoglobin feature intensities of multiple hemoglobin feature bands do not meet the preset convergence conditions, the first basis estimates of multiple hemoglobin feature bands are corrected until the candidate hemoglobin feature intensities of multiple hemoglobin feature bands meet the preset convergence conditions, and the second basis estimates of multiple hemoglobin feature bands are obtained.

[0120] In this embodiment, the fixed absorption ratio of hemoglobin refers to the inherent absorption characteristics of hemoglobin in different wavelength bands, usually existing in the form of absorption coefficients, and its proportional relationship between specific wavelength bands is relatively fixed. For example, the fixed absorption ratio of hemoglobin is a constant, determined by the electronic transitions of the hemoglobin molecule, and is independent of concentration and sample; this fixed ratio is denoted as... ,in (Corresponding to the first hemoglobin characteristic band) (Corresponding to the second hemoglobin characteristic band) and (Corresponding to the third hemoglobin characteristic band) is a positive number less than 1. and The specific values ​​can be obtained by measuring pure hemoglobin solution when the equipment leaves the factory, and once determined, they are used permanently; for example, typical values. , .

[0121] The characteristic measured reflectance is the reflectance value of the target excrement sample actually measured in the characteristic band of hemoglobin. It includes the reflection of the excrement substrate and the absorption effect of hemoglobin.

[0122] The first substrate estimate is a preliminary prediction of the excrement substrate reflectance in the characteristic bands of hemoglobin, based on the baseline measured reflectance of the baseline reference band and the substrate shape dictionary.

[0123] The candidate hemoglobin characteristic intensity is a provisional hemoglobin absorption intensity index calculated during the correction process based on the first basis estimate, the fixed absorption ratio of hemoglobin, and the measured characteristic reflectance. For example, the candidate hemoglobin characteristic intensity can be determined using the following formula 2:

[0124] Formula 2

[0125] In formula 2, Used to characterize the intensity of candidate hemoglobin features The first basis estimate is used to characterize x, the measured reflectance is used to characterize the characteristic, and r is used to characterize the fixed absorption ratio of hemoglobin.

[0126] The preset convergence criterion is a standard used to determine whether the iterative correction process has reached sufficient accuracy or a stable state. When the feature intensity of the candidate hemoglobins meets this condition, it indicates that the current basis estimate is accurate enough, and the correction can be stopped. For example, this condition can be set as follows: if the range of the feature intensity of the three candidate hemoglobins... If the difference is less than a preset range threshold, the feature intensities of the three candidate hemoglobins are considered sufficiently consistent, and the loop correction ends. Simultaneously, if the number of iterations reaches a preset iteration threshold, the loop correction also ends to avoid infinite loops. The characteristic intensity of the first candidate hemoglobin. The characteristic intensity of the second candidate hemoglobin. For the third candidate hemoglobin feature intensity, max represents the function that takes the maximum value and min represents the function that takes the minimum value.

[0127] Correction refers to the process of adjusting the first basis estimate based on the relationship between the candidate hemoglobin feature intensity and a preset convergence condition. The aim is to make the basis estimate more accurately reflect the true reflectance of the excrement basis in the hemoglobin feature band. Correction can employ iterative optimization algorithms, such as gradient descent, to gradually adjust the first basis estimate (each adjustment increment can be set to a preset step size, for example, a preset step size of...). This value can be equal to 1 / 100 of the maximum reflectance reduction that hemoglobin may cause (approximately 0.05), and can be kept consistent with the device reflectance measurement noise (approximately 0.0005) to ensure stable convergence of the iteration, so that it approaches the true substrate reflectance.

[0128] The second basis estimate is a more accurate estimate of the excrement basis reflectance in the characteristic bands of hemoglobin, obtained after iterative correction. It is considered to more accurately represent the true basis spectral shape under the influence of hemoglobin absorption. When the iterative correction process meets the preset convergence condition, the basis estimate obtained in the last iteration is the second basis estimate.

[0129] The present invention refines the initially obtained first basis estimate by introducing an iterative correction mechanism. First, based on the current first basis estimate, the fixed absorption ratio of hemoglobin, and the measured characteristic reflectance, candidate hemoglobin characteristic intensities are calculated for each hemoglobin characteristic band. These candidate intensities reflect the degree of hemoglobin absorption under the current basis assumptions. Ideally, if the basis estimate is accurate, the calculated hemoglobin characteristic intensities across different hemoglobin characteristic bands should exhibit good consistency. Therefore, the system checks whether these candidate hemoglobin characteristic intensities meet preset convergence conditions. If they do not meet the convergence conditions, it indicates that the current first basis estimate still has a deviation. In this case, the system corrects the first basis estimate based on this deviation. The purpose of the correction is to adjust the basis estimate so that it can generate more consistent or stable candidate hemoglobin characteristic intensities in the next iteration. This process of calculation, checking, and correction is repeated, forming a closed-loop iteration. Each iteration brings the baseline estimate closer to the true reflectance of the excrement substrate in the hemoglobin characteristic band, thus more effectively separating the hemoglobin absorption signal from the total reflectance. The iteration stops when the candidate hemoglobin characteristic intensity meets the preset convergence condition, and the resulting baseline estimate is determined as the second baseline estimate. This iterative correction method ensures the accuracy of the baseline estimate, providing a reliable basis for subsequent accurate calculations of hemoglobin concentration.

[0130] Through the above technical solution, this invention introduces an iterative correction mechanism, enabling continuous optimization of the baseline estimation values ​​for hemoglobin characteristic bands. By repeatedly calculating the characteristic intensities of candidate hemoglobins and adjusting the baseline estimation values ​​based on their convergence, the coupling effect between hemoglobin absorption and excrement baseline reflectivity can be more effectively isolated. This significantly improves the accuracy of the baseline estimation, thus laying a solid foundation for the subsequent accurate determination of the target hemoglobin concentration, effectively avoiding hemoglobin concentration detection errors caused by inaccurate baseline estimation, and improving the reliability and accuracy of the detection results.

[0131] In some embodiments, the present invention further proposes to correct the first basis estimate of multiple hemoglobin characteristic bands, including:

[0132] Obtain the feature intensity of candidate hemoglobins in multiple hemoglobin feature bands and the feature intensity difference between them and the median of hemoglobin feature intensity;

[0133] Based on the characteristic intensity difference, the base correction direction of the characteristic band of hemoglobin is determined;

[0134] Based on the basis correction direction, the first basis estimate of the hemoglobin characteristic band is corrected.

[0135] In this embodiment, the difference between the candidate hemoglobin feature intensity of multiple hemoglobin feature bands and the median of hemoglobin feature intensity is obtained. This aims to quantify the degree of deviation between the candidate hemoglobin feature intensity of each hemoglobin feature band in the current iteration and the overall trend (represented by the median).

[0136] The candidate hemoglobin feature intensity is calculated based on the current first basis estimate, the fixed absorption ratio of hemoglobin, and the measured feature reflectance. It reflects the absorption intensity of hemoglobin in this band under the current basis estimate. The median of the hemoglobin feature intensity can represent the central trend of the candidate hemoglobin feature intensity across all current hemoglobin feature bands, effectively avoiding interference from outliers in determining the correction direction.

[0137] The difference in feature intensity can be obtained by subtracting the median of the candidate hemoglobin feature intensity of all hemoglobin feature bands from the candidate hemoglobin feature intensity of each hemoglobin feature band.

[0138] Based on the feature intensity difference, the basis correction direction of the hemoglobin feature bands is determined. This step uses the feature intensity difference to guide the adjustment direction of the first basis estimate:

[0139] ①If (That is, the intensity difference of the j-th feature is greater than 0), indicating that the j-th feature is greater than 0. The intensity of candidate hemoglobin features in each hemoglobin characteristic band is greater than the median; since the median represents the current baseline intensity, This means that the calculated drop for this wavelength is too large relative to the baseline; in other words, the initial baseline estimate for this wavelength has been overestimated and needs to be reduced.

[0140] ②If This indicates that the initial basis estimate for this wavelength was underestimated and needs to be increased.

[0141] ③If (Or within a certain preset floating range), it indicates that the basis estimation of this wavelength is already reasonable and does not need to be adjusted for the time being.

[0142] Thus, this invention can determine the bias direction of the basis estimation at each wavelength simply by comparing the magnitude of the characteristic intensities of the three candidate hemoglobins, without needing to know the specific value of the bias.

[0143] Based on the basis correction direction, the first basis estimate for the hemoglobin characteristic bands is corrected. This step adjusts the first basis estimate in the current iteration according to the basis correction direction determined in the previous step. The purpose of the correction is to make the candidate hemoglobin feature intensities of each hemoglobin characteristic band more consistent, so that the first basis estimate more accurately reflects the true optical properties of the excrement substrate. The correction can be achieved by incrementally adjusting the first basis estimate according to the basis correction direction, for example, by adding or subtracting an adjustment step size from the first basis estimate. This correction method makes the adjustment of the first basis estimate no longer a blind iteration, but rather a directional approach towards the convergence target.

[0144] The present invention, when correcting the first basis estimate of multiple hemoglobin characteristic bands, first obtains the feature intensity difference between the candidate hemoglobin feature intensity and the median hemoglobin feature intensity of each hemoglobin characteristic band. This difference intuitively reflects the degree of deviation of the hemoglobin absorption intensity of each band from the overall average level. Subsequently, based on these feature intensity differences, the basis correction direction for each hemoglobin characteristic band is systematically determined. Finally, according to the determined basis correction direction, the first basis estimate of the corresponding hemoglobin characteristic band is specifically corrected. In this way, each iterative correction can intelligently adjust the first basis estimate according to the current distribution of hemoglobin absorption, so that the hemoglobin feature intensity of each band gradually tends to be consistent, thereby ensuring that the first basis estimate can more accurately isolate the influence of hemoglobin absorption, laying a solid foundation for subsequent determination of target hemoglobin concentration. This iterative correction mechanism based on feature intensity difference and correction direction significantly improves the accuracy and convergence efficiency of basis estimation, avoiding the oscillation or slow convergence problems that may be caused by blind iteration.

[0145] The above technical solution addresses the issue of base course correction when detecting hemoglobin levels in target excrement samples. Instead of simple iteration, when the first base estimate needs correction to meet convergence conditions, the method obtains the difference between the candidate hemoglobin characteristic intensity and the median hemoglobin characteristic intensity for each hemoglobin characteristic band. Based on this difference, the base course correction direction is determined, allowing for targeted and directional adjustments to the first base estimate. This correction mechanism effectively guides the first base estimate towards the true base reflectance curve, avoiding the slow convergence, oscillations, or non-convergence issues that can result from blind iteration. Therefore, this solution significantly improves the correction efficiency and accuracy of the first base estimate, thereby enhancing the detection precision and stability of the final target hemoglobin concentration.

[0146] In some embodiments, the present invention further proposes that the first baseline estimate includes a first band baseline estimate of a first hemoglobin characteristic band, a second band baseline estimate of a second hemoglobin characteristic band, and a third band baseline estimate of a third hemoglobin characteristic band; the first hemoglobin characteristic band is located on the shortwave side of the second hemoglobin characteristic band, and the second hemoglobin characteristic band is located on the shortwave side of the third hemoglobin characteristic band;

[0147] Based on the basis correction direction, the first basis estimate of the hemoglobin characteristic bands is corrected, including:

[0148] Based on the first characteristic intensity difference of the first hemoglobin characteristic band and the third characteristic intensity difference of the third hemoglobin characteristic band, the expected characteristic intensity difference of the second hemoglobin characteristic band is determined.

[0149] Based on the expected feature intensity difference of the second hemoglobin feature band and the second feature intensity difference of the second hemoglobin feature band, the band smoothing coefficient of the second hemoglobin feature band is determined.

[0150] Based on the baseline estimates of the first and third bands, the expected band baseline estimates of the second hemoglobin characteristic bands are determined.

[0151] The second band basis estimate is corrected based on the band smoothing coefficient and the expected band basis estimate.

[0152] In this embodiment, the first, second, and third hemoglobin characteristic bands are specific spectral bands used for detecting hemoglobin. These bands were selected because hemoglobin exhibits significant absorption characteristics within these bands and is sensitive to changes in hemoglobin concentration. These bands are typically located in the visible or near-infrared spectral regions; for example, 415 nm, 540 nm, and 575 nm can be selected as the three hemoglobin characteristic bands.

[0153] The first, second, and third band floor estimates are preliminary estimates of the floor reflectance of excrement corresponding to the three characteristic hemoglobin bands mentioned above. These floor estimates represent the optical properties of the excrement itself in the absence of hemoglobin absorption. These estimates can be obtained through spectral modeling, curve fitting, or by looking up values ​​in a pre-defined dictionary.

[0154] The first hemoglobin characteristic band is located on the short-wave side of the second hemoglobin characteristic band, and the second hemoglobin characteristic band is located on the short-wave side of the third hemoglobin characteristic band. This describes the relative positional relationship of the three hemoglobin characteristic bands in the spectrum, that is, they are arranged in ascending order of wavelength.

[0155] The first feature intensity difference, the third feature intensity difference, and the second feature intensity difference are the differences between the candidate hemoglobin feature intensity and the median hemoglobin feature intensity corresponding to the first hemoglobin feature band, the third hemoglobin feature band, and the second hemoglobin feature band.

[0156] Determining the expected characteristic intensity difference of the second hemoglobin characteristic band refers to predicting the characteristic intensity difference that the second hemoglobin characteristic band should have based on the first characteristic intensity difference of the first hemoglobin characteristic band and the third characteristic intensity difference of the third hemoglobin characteristic band. For example, it can be obtained by adding the first characteristic intensity difference of the first hemoglobin characteristic band and the third characteristic intensity difference of the third hemoglobin characteristic band and then dividing by two.

[0157] Determining the band smoothing coefficient for the second hemoglobin characteristic band refers to calculating a coefficient used to adjust the correction amplitude based on the relationship between the expected characteristic intensity difference and the actual second characteristic intensity difference of the second hemoglobin characteristic band. This coefficient reflects the degree of deviation between the actual measured value and the smoothing trend. The specific calculation method is as follows: First, solve for the absolute difference between the expected characteristic intensity difference and the second characteristic intensity difference. Then, use this absolute difference as the numerator and the average of the absolute values ​​of the first and third characteristic intensity differences as the denominator. Calculate the ratio of the numerator to the denominator to obtain the dimensionless deviation index. To avoid the denominator being zero, a preset parameter adjustment factor greater than zero needs to be added to the denominator. The dimensions of the parameter adjustment factor are consistent with those of the denominator, and the value of the parameter adjustment factor can be set to 0.001. Finally, based on the deviation index threshold, the band smoothing coefficient in the 0~1 interval is obtained by segmented mapping: when the deviation index is less than or equal to 30%, it means that there is no obvious local curvature of the substrate, and the band smoothing coefficient is taken as 0.1; when the deviation index is greater than 30%, it is determined that there is significant local curvature of the substrate and a deviation from the smoothness trend, and the band smoothing coefficient is taken as 0.3.

[0158] Determining the expected base estimate for the second hemoglobin characteristic band refers to predicting the base estimate that the second hemoglobin characteristic band should have based on the base estimates for the first and third bands. For example, the base estimates for the first and third bands can be added together and then divided by two.

[0159] Correcting the second band base estimate involves adjusting the original base estimate of the second hemoglobin characteristic band using the calculated band smoothing coefficient and the expected band base estimate, making it more consistent with the smoothing trend of the spectrum. For example, the second band base estimate can be corrected using the following formula 3:

[0160] Formula 3

[0161] In formula 3, This is used to characterize the result obtained after correcting the second-band basis estimate. Used to characterize the second band basis estimate. Used to characterize the expected band basis estimate. Used to characterize the band smoothness coefficient.

[0162] In the above method, to improve the accuracy and stability of the first basis estimate, this invention introduces a correction mechanism based on spectral smoothing. This mechanism first identifies three adjacent hemoglobin characteristic bands: the first, second, and third hemoglobin characteristic bands, arranged in ascending order of wavelength in the spectrum. For each of these three bands, there exist corresponding basis estimates for the first, second, and third bands, as well as their respective feature intensity differences. The solution of this invention optimizes the basis estimate of the intermediate band (the second hemoglobin characteristic band) by utilizing the continuity and smoothness of the spectral data. Specifically, it first predicts the expected feature intensity difference of the second hemoglobin characteristic band based on the first feature intensity difference of the first hemoglobin characteristic band and the third feature intensity difference of the third hemoglobin characteristic band. This expected value reflects the level of feature intensity difference that the second band should have under ideally smoothed spectra. Next, this expected feature intensity difference is compared with the actual measured second feature intensity difference in the second hemoglobin feature band to determine a band smoothing coefficient. This coefficient quantifies the deviation between the actual measured value of the second band and the smoothing trend. Simultaneously, based on the first and third band baseline estimates, the scheme also determines the expected band baseline estimate for the second hemoglobin feature band. This expected value represents the level the second band baseline should have under the spectral smoothing assumption. Finally, using the previously calculated band smoothing coefficient, combined with the expected band baseline estimate for the second band, the original baseline estimate for the second band is corrected. This correction is not a simple replacement, but rather an adjustment to the smoothing trend through weighted averaging and other methods. This effectively suppresses the influence of noise and local fluctuations on the baseline estimate while preserving actual measurement information, making the corrected baseline estimate more spectrally smooth and reasonable. In this way, the scheme of this invention can more accurately separate the absorption signal of hemoglobin from the reflection signal of the excrement baseline, laying the foundation for the subsequent accurate determination of hemoglobin concentration.

[0163] Through the above technical solution, when correcting the first basis estimate, this invention no longer relies solely on the characteristic intensity difference of a single band, but fully utilizes the spectral continuity and smoothness between adjacent hemoglobin characteristic bands. By introducing the expected characteristic intensity difference and the expected band basis estimate, and combining it with a band smoothing coefficient for weighted correction, the problem of unsmooth or inaccurate basis estimates caused by single-band measurement noise or local fluctuations is effectively avoided. This correction mechanism based on spectral smoothing makes the basis estimates of each hemoglobin characteristic band more coordinated and stable in the spectral dimension, thereby enabling more accurate stripping of the hemoglobin absorption signal and significantly improving the accuracy and reliability of excrement hemoglobin content detection.

[0164] In some embodiments, the present invention further proposes that the process of determining the target hemoglobin concentration of the target excrement sample based on the second basis estimate of multiple hemoglobin characteristic bands in step S140 above specifically includes:

[0165] Based on the second basis estimate of multiple hemoglobin characteristic bands, the target hemoglobin characteristic intensity of multiple hemoglobin characteristic bands is determined;

[0166] The target comprehensive hemoglobin characteristic intensity of the target excrement sample is obtained by weighted summation of the target hemoglobin characteristic intensities of multiple hemoglobin characteristic bands.

[0167] The target hemoglobin characteristic intensity is matched with the target concentration curve to obtain the target hemoglobin concentration of the target excrement sample; the target concentration curve is used to characterize the mapping relationship between the comprehensive hemoglobin characteristic intensity and the hemoglobin concentration of the excrement sample.

[0168] In this embodiment, the target hemoglobin characteristic intensity of multiple hemoglobin characteristic bands is determined based on second basis estimates of multiple hemoglobin characteristic bands. The aim is to quantify the signal intensity caused by hemoglobin absorption in each hemoglobin characteristic band from the basis-corrected spectral data. The second basis estimates have removed the influence of the optical properties of the excrement substrate itself; therefore, the characteristic intensity determined based on this can more accurately reflect the true absorption of hemoglobin. For example, the target hemoglobin characteristic intensity can be obtained by subtracting the measured reflectance of the corresponding hemoglobin characteristic band from the second basis estimate and then dividing by the corresponding fixed absorption ratio of hemoglobin.

[0169] The target hemoglobin characteristic intensities from multiple hemoglobin characteristic bands are weighted and summed to obtain the target comprehensive hemoglobin characteristic intensity of the target excrement sample. The purpose is to integrate multiple hemoglobin characteristic intensities from different hemoglobin characteristic bands into a single, representative comprehensive index. Through weighted summation, different weights can be assigned based on the sensitivity of different bands to changes in hemoglobin concentration, signal-to-noise ratio, or measurement reliability, thereby improving the accuracy and stability of the final concentration determination. For example, the weight of the first hemoglobin characteristic band can be set to 0.5, the weight of the second hemoglobin characteristic band can be set to 0.25, and the weight of the third hemoglobin characteristic band can be set to 0.25.

[0170] Matching the target composite hemoglobin characteristic intensity with the target concentration curve yields the target hemoglobin concentration of the target excrement sample. This step converts the composite characteristic intensity obtained from spectral measurements into an actual hemoglobin concentration value. The target concentration curve establishes a quantitative relationship between the composite spectral signal and the biochemical concentration, which is crucial for the final concentration determination. The target concentration curve can be obtained in advance through experimental calibration. For example, a series of excrement samples with known hemoglobin concentrations can be prepared, their composite hemoglobin characteristic intensities measured, and then a calibration curve between the composite characteristic intensities and hemoglobin concentrations can be plotted. During actual testing, the measured target composite hemoglobin characteristic intensity is mapped to the corresponding hemoglobin concentration through table lookup, interpolation, or curve fitting.

[0171] The target concentration curve characterizes the mapping relationship between the overall hemoglobin characteristic intensity and hemoglobin concentration in excretory samples, clarifying its essence and function. It acts as a bridge, connecting the spectroscopic measurement results (overall hemoglobin characteristic intensity) with the actual quantity of biological or medical interest (hemoglobin concentration). This mapping relationship can be linear, meaning there is a simple proportional relationship between the overall characteristic intensity and concentration, especially in the low concentration range. It can also be non-linear; for example, a saturation effect may occur at higher concentrations, in which case the curve may exhibit a logarithmic or exponential relationship. Establishing the curve typically requires extensive experimental data for fitting and validation.

[0172] The present invention first determines the target hemoglobin feature intensity for each hemoglobin feature band based on second-basis estimates of multiple hemoglobin feature bands, thereby effectively isolating hemoglobin-specific signals. Subsequently, these hemoglobin feature intensities from different hemoglobin feature bands are weighted and summed to generate a single target composite hemoglobin feature intensity. This weighted summation process can be optimized according to factors such as the sensitivity and signal-to-noise ratio of each band, thus more robustly integrating multispectral information and effectively reducing the errors and uncertainties that may arise from single-band measurements. Finally, by matching this target composite hemoglobin feature intensity with a pre-established target concentration curve, the spectral signal can be accurately converted into an actual hemoglobin concentration value. The entire process forms a complete and optimized chain from multispectral data to accurate concentration determination, significantly improving the accuracy and reliability of detection.

[0173] As an example, after obtaining the target composite hemoglobin characteristic intensity of the target excrement sample, a physical plausibility check can be performed. Since the reflectance reduction caused by hemoglobin cannot exceed the base reflectance itself, the target composite hemoglobin characteristic intensity must be less than or equal to the second base estimate of the first hemoglobin characteristic band. If the target composite hemoglobin characteristic intensity is greater than the second base estimate of the first hemoglobin characteristic band, then the target composite hemoglobin characteristic intensity is forcibly set to the second base estimate of the first hemoglobin characteristic band. Furthermore, the second and third hemoglobin characteristic bands should also be checked. .in, This is the second base estimate of the j-th hemoglobin characteristic band (j is either two or three). The target composite hemoglobin characteristic intensity of the target excrement sample. The fixed absorption ratio of hemoglobin in the j-th hemoglobin characteristic band is used. If the condition is not met, the ratio is compressed until the inequality is not violated, thus ensuring that the output result is always within the physically permissible range.

[0174] Through the above technical solution, this invention can effectively integrate spectral information from multiple hemoglobin characteristic bands. By using a weighted summation method, it optimizes the utilization efficiency of data from different bands and reduces the errors and uncertainties that may exist in single-band measurements. Simultaneously, by matching with a pre-calibrated target concentration curve, it achieves precise quantification of the spectral signal to the actual hemoglobin concentration, thereby significantly improving the accuracy and robustness of excrement hemoglobin content detection and solving the problem of how to effectively integrate information from different bands and accurately map it to the actual concentration.

[0175] In some embodiments, the present invention further proposes that the target concentration curve includes concentration sub-curves of various excrement forms;

[0176] By matching the target hemoglobin characteristic intensity with the target concentration curve, the target hemoglobin concentration of the target excrement sample is obtained, including:

[0177] Obtain the target excrement morphology from the target excrement sample;

[0178] The target hemoglobin concentration is obtained by matching the target comprehensive hemoglobin characteristic intensity with the target concentration sub-curve of the target excrement morphology.

[0179] In this embodiment, the target concentration curve includes concentration sub-curves for various excretory forms. This means that the mapping relationship used to convert the overall hemoglobin characteristic intensity into hemoglobin concentration is no longer a single curve, but rather a set of multiple sub-curves pre-established or derived for different excretory forms. Each concentration sub-curve is specifically designed to characterize the correspondence between the overall hemoglobin characteristic intensity and hemoglobin concentration under a specific excretory form. For example, a separate concentration sub-curve can be established for different forms such as loose stools, soft stools, and hard stools.

[0180] The purpose of acquiring the target excrement morphology of a sample is to identify the specific form of the excrement sample to be tested, so as to select an appropriate concentration sub-curve for subsequent concentration calculation. Implementation methods may include, but are not limited to: First, manual observation and judgment, where operators judge the excrement sample based on its visual characteristics (such as color, consistency, shape, etc.) and manually input or select the corresponding excrement morphology into the system; Second, image recognition technology, using a camera or other imaging device to acquire images of the excrement sample, and then using image processing and machine learning algorithms to analyze the image features and automatically identify and classify the morphology of the excrement sample; Third, using physical sensors to assist in judgment, for example, using physical quantity sensors such as viscosity sensors, humidity sensors, or density sensors to measure the physical properties of the excrement sample and determine its morphology based on these measurements.

[0181] Matching the target composite hemoglobin characteristic intensity with the target concentration sub-curve of the target excrement morphology refers to comparing the previously calculated composite hemoglobin characteristic intensity with the concentration sub-curve corresponding to that specific morphology after determining the morphology of the target excrement sample, thereby obtaining the final hemoglobin concentration value. Implementation methods may include, but are not limited to: First, a lookup table method, where each concentration sub-curve is pre-stored as a data table (e.g., a pair of values ​​corresponding to composite hemoglobin characteristic intensity and hemoglobin concentration). Based on the obtained target excrement morphology, the corresponding sub-data table is selected, and then the hemoglobin concentration closest to or corresponding to the target composite hemoglobin characteristic intensity is found from the sub-data table through lookup or interpolation. Second, a function fitting method, where each concentration sub-curve is fitted to a specific mathematical function (e.g., a polynomial function, exponential function, or logarithmic function). Based on the obtained target excrement morphology, the corresponding mathematical function model is selected, and then the target composite hemoglobin characteristic intensity is substituted into the function as input to calculate the hemoglobin concentration.

[0182] When determining the target hemoglobin concentration of a target excrement sample, the present invention first identifies the specific excrement morphology of the sample. This identification process is crucial because it allows the system to personalize the processing based on the actual physical characteristics of the sample. Once the sample morphology is determined, the system precisely selects a concentration sub-curve that matches the specific morphology from a pre-set set of concentration sub-curves for various excrement morphologies. Subsequently, the target composite hemoglobin characteristic intensity calculated through the aforementioned steps is matched with this selected concentration sub-curve optimized for the specific morphology. This matching process is no longer based on a universal, potentially biased, single curve, but rather on a dedicated curve that highly fits the current sample morphology. In this way, the present invention can effectively eliminate or significantly reduce the impact of changes in optical properties caused by differences in excrement morphology on the accuracy of hemoglobin concentration measurement, making the conversion from composite hemoglobin characteristic intensity to hemoglobin concentration more accurate and reliable.

[0183] Through the above technical solution, this invention can provide more accurate hemoglobin concentration detection results for excrement samples of different forms. Because the diversity of excrement forms can lead to significant differences in their optical properties, using a single concentration curve for calculation can easily introduce errors. This invention effectively solves this problem by introducing concentration sub-curves for multiple excrement forms and selecting the appropriate sub-curve for matching based on the actual form of the target excrement sample. This makes the mapping relationship between hemoglobin concentration and overall hemoglobin characteristic intensity more closely reflect the actual situation of the sample, thereby significantly improving the accuracy and reliability of the detection and providing more valuable data for clinical diagnosis.

[0184] This invention also provides a multispectral-based fecal hemoglobin content detection system, comprising: a memory and a processor; the memory and the processor are connected; the memory is used to store program instructions; the processor is used to implement the above-described multispectral-based fecal hemoglobin content detection method when the program instructions are executed.

Claims

1. A method for detecting hemoglobin content in excrement based on multispectral analysis, characterized in that, The method includes: Obtain the reflectance sequence of the target excrement sample; the reflectance sequence includes the characteristic measured reflectance of the target excrement sample in multiple hemoglobin characteristic bands, and the baseline measured reflectance of the target excrement sample in multiple baseline reference bands; the hemoglobin characteristic bands refer to bands that are sensitive to changes in hemoglobin concentration, and the baseline reference bands refer to bands that are not sensitive to changes in hemoglobin concentration and whose reflectance only reflects the optical properties of the excrement substrate itself; Based on the measured baseline reflectance of the multiple baseline reference bands, the first substrate estimate of the target excrement sample in the multiple hemoglobin characteristic bands is determined from the substrate shape dictionary; the substrate shape dictionary is used to characterize the proportional relationship between the reflectance of each of the hemoglobin characteristic bands under different excrement substrate conditions. Based on the fixed absorption ratio of hemoglobin in the multiple hemoglobin characteristic bands and the measured characteristic reflectance of the multiple hemoglobin characteristic bands, the first basis estimate of the multiple hemoglobin characteristic bands is corrected to obtain the second basis estimate of the multiple hemoglobin characteristic bands. The target hemoglobin concentration of the target excrement sample is determined based on the second basis estimate of the multiple hemoglobin characteristic bands.

2. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 1, characterized in that, The acquisition of the reflectance sequence of the target excrement sample includes: Obtain the dark current value, median light intensity, and reference light intensity of the target excrement sample in the target wavelength band; the target wavelength band is any one of the plurality of hemoglobin characteristic wavelength bands and the plurality of baseline reference wavelength bands; Based on the first difference between the median light intensity and the dark current value, and the second difference between the reference light intensity and the dark current value, the measured reflectance of the target excrement sample in the target wavelength band is determined. Based on the measured reflectance of the target excrement sample in each of the target wavelength bands, the reflectance sequence of the target excrement sample is determined.

3. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 1, characterized in that, The measured baseline reflectance of the plurality of baseline reference bands includes the first measured baseline reflectance of the first baseline reference band and the second measured baseline reflectance of the second baseline reference band, wherein the first baseline reference band is located on the shortwave side of the second baseline reference band; The determination of the first substrate estimate of the target excrement sample in the multiple hemoglobin characteristic bands based on the measured baseline reflectance of the multiple baseline reference bands from the substrate shape dictionary includes: Based on the measured reflectance of the first baseline and the measured reflectance of the second baseline, the target wavelength ratio of the target excrement sample is determined; The target wavelength ratio is matched with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample in multiple hemoglobin feature bands; the target feature ratio is used to characterize the reflectance level of the substrate of the target excrement sample in the hemoglobin feature bands relative to the first baseline reference band. Based on the measured reflectance of the first baseline and the target feature ratio of the plurality of hemoglobin feature bands, the first baseline estimate of the plurality of hemoglobin feature bands is determined.

4. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 3, characterized in that, The step of matching the target wavelength ratio with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample in the multiple hemoglobin feature bands includes: The target wavelength ratio is matched with the base shape dictionary to obtain K candidate feature ratio groups that are closest to the target wavelength ratio; the candidate feature ratio groups include candidate feature ratios of the plurality of hemoglobin feature bands, where K is a positive integer; Based on the K groups of candidate feature proportions, the target feature proportions of the multiple hemoglobin feature bands are determined.

5. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 3, characterized in that, The step of matching the target wavelength ratio with the substrate shape dictionary to obtain the target feature ratio of the target excrement sample before the target feature ratio of the multiple hemoglobin feature bands further includes: The reference wavelength ratio of the reference excrement sample is determined based on the first reference reflectance of the reference excrement sample in the first baseline reference band and the second reference reflectance of the second baseline reference band. Based on the reference reflectance of the reference excrement sample in the multiple hemoglobin characteristic bands, and the first reference reflectance, the reference feature ratio of the reference excrement sample in the multiple hemoglobin characteristic bands is determined. The base shape dictionary is constructed based on the reference wavelength ratio of each of the reference excrement samples and the reference feature ratio of the multiple hemoglobin characteristic bands.

6. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 1, characterized in that, The process of correcting the first basis estimate of the multiple hemoglobin characteristic bands based on the fixed absorption ratio of hemoglobin in the multiple hemoglobin characteristic bands and the measured characteristic reflectance of the multiple hemoglobin characteristic bands to obtain the second basis estimate of the multiple hemoglobin characteristic bands includes: Based on the fixed absorption ratio of hemoglobin, the measured reflectance of the features, and the estimated value of the first substrate, the candidate hemoglobin feature intensities of the multiple hemoglobin feature bands are determined. In response to the fact that the candidate hemoglobin feature intensities of the plurality of hemoglobin feature bands do not meet the preset convergence condition, the first basis estimate of the plurality of hemoglobin feature bands is corrected until the candidate hemoglobin feature intensities of the plurality of hemoglobin feature bands meet the preset convergence condition, and the second basis estimate of the plurality of hemoglobin feature bands is obtained.

7. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 6, characterized in that, The correction of the first basis estimate of the plurality of hemoglobin characteristic bands includes: Obtain the feature intensity difference between the candidate hemoglobin feature intensity and the median hemoglobin feature intensity of the multiple hemoglobin feature bands; Based on the characteristic intensity difference, the base correction direction of the hemoglobin characteristic band is determined; Based on the said basis correction direction, the first basis estimate of the hemoglobin characteristic band is corrected.

8. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 7, characterized in that, The first baseline estimate includes a first band baseline estimate of a first hemoglobin characteristic band, a second band baseline estimate of a second hemoglobin characteristic band, and a third band baseline estimate of a third hemoglobin characteristic band; the first hemoglobin characteristic band is located on the shortwave side of the second hemoglobin characteristic band, and the second hemoglobin characteristic band is located on the shortwave side of the third hemoglobin characteristic band; The step of correcting the first basis estimate of the hemoglobin characteristic band based on the basis correction direction includes: Based on the first feature intensity difference of the first hemoglobin feature band and the third feature intensity difference of the third hemoglobin feature band, the expected feature intensity difference of the second hemoglobin feature band is determined. Based on the expected feature intensity difference of the second hemoglobin feature band and the second feature intensity difference of the second hemoglobin feature band, the band smoothing coefficient of the second hemoglobin feature band is determined. Based on the first band baseline estimate and the third band baseline estimate, the expected band baseline estimate of the second hemoglobin characteristic band is determined; The second band basis estimate is corrected based on the band smoothing coefficient and the expected band basis estimate.

9. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 1, characterized in that, The determination of the target hemoglobin concentration of the target excrement sample based on the second basis estimation value of the multiple hemoglobin characteristic bands includes: Based on the second basis estimation values ​​of the plurality of hemoglobin characteristic bands, the target hemoglobin characteristic intensity of the plurality of hemoglobin characteristic bands is determined; The target comprehensive hemoglobin feature intensity of the target excrement sample is obtained by weighted summation of the target hemoglobin feature intensities of the multiple hemoglobin feature bands. The target hemoglobin characteristic intensity is matched with the target concentration curve to obtain the target hemoglobin concentration of the target excrement sample; the target concentration curve is used to characterize the mapping relationship between the comprehensive hemoglobin characteristic intensity and the hemoglobin concentration of the excrement sample.

10. The method for detecting excrement hemoglobin content based on multispectral analysis according to claim 9, characterized in that, The target concentration curve includes concentration sub-curves for various excrement forms; The step of matching the target hemoglobin feature intensity with the target concentration curve to obtain the target hemoglobin concentration of the target excrement sample includes: Obtain the target excrement morphology of the target excrement sample; The target hemoglobin characteristic intensity is matched with the target concentration sub-curve of the target excrement morphology to obtain the target hemoglobin concentration of the target excrement sample.

Images