CN-121994794-A - Multi-algorithm comprehensive analysis method and system for detecting abnormality of test strip

CN121994794ACN 121994794 ACN121994794 ACN 121994794ACN-121994794-A

Abstract

The invention discloses a multi-algorithm comprehensive analysis method and a system for detecting test strip abnormality. The method comprises the steps of obtaining relevant data of a B line and a C line of test paper to be detected, judging whether the test paper to be detected has test paper strip position deviation or not by comparing average value differences at the tail ends of the B line and the C line according to the relevant data of the B line and the C line, if not, analyzing photoelectric signal changes caused by error or excessive sample adding sequences, extracting and verifying an abnormal region to obtain the abnormal region, repairing the abnormal region, restoring a normal form of the curve to obtain a correction result, determining a value window which is not intersected with all the abnormal regions according to the correction result to obtain a value position calculation result, constructing a model according to the abnormal region, the correction result and the value position calculation result, matching with a historical case, and calculating a final detection result and the confidence coefficient of the final detection result. The method not only can effectively identify various abnormal conditions, but also can dynamically optimize the value position to improve the data confidence.

Inventors

XIE SIWEN
ZHANG JUN
MA XIAOYONG

Assignees

杭州宣航科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260410

Claims (10)

1. The multi-algorithm comprehensive analysis method for detecting the abnormality of the test strip is characterized by comprising the following steps: acquiring related data of a B line and a C line of test paper to be detected; According to the related data of the B line and the C line, judging whether the test paper to be detected has test paper strip position deviation or not by comparing the average value difference of the tail ends of the B line and the C line; If the test paper to be detected has the test paper strip position deviation, judging that the detection result is invalid; If the test paper to be detected does not have the test paper strip position deviation, analyzing photoelectric signal change caused by error or excessive sample adding sequence, and extracting and verifying an abnormal section to obtain an abnormal region; repairing the abnormal region, and restoring the normal form of the curve to obtain a correction result; determining a value window which has no intersection with all the abnormal areas according to the correction result to obtain a value position calculation result; And constructing a model according to the abnormal region, the correction result and the value position calculation result, matching the model with a historical case, and calculating a final detection result and the confidence coefficient thereof.
2. The multi-algorithm integrated analysis method for detecting test strip anomalies according to claim 1, wherein the judging whether the test strip to be detected has test strip position deviation by comparing average value differences at the ends of the B line and the C line according to the related data of the B line and the C line comprises the following steps: Calculating the average value of a plurality of data points at the tail of the B line and the C line according to the related data of the B line and the C line; And when the absolute value of the difference between the data of the B line and the C line at the initial moment and the average value of a plurality of data points at the tail of the B line and the C line is not more than a set difference threshold value, determining that the test paper to be detected has test paper strip position deviation.
3. The multi-algorithm integrated analysis method for detecting strip anomalies according to claim 1, characterized in that said analyzing the changes of the photoelectric signal caused by the errors or overdose of the sample-adding sequence, extracting and verifying the anomaly intervals to obtain anomaly areas, comprises: Acquiring related data of test paper to be detected; setting an abnormal interval set, a time span and a threshold key parameter, and initializing an abnormal state as closed; Analyzing the trend and the pre-estimated value of each data point in the related data of the test paper to be detected through a sliding window, and judging whether the condition of starting abnormal judgment is met or not; when the condition of starting the abnormality judgment is met, marking the current index as an abnormality starting point, recording left boundary information, and activating an abnormality judgment mode to obtain an abnormality section; continuously monitoring the dropping rate, the time span, the lowest point position and the corresponding minimum value of the abnormal section to obtain a monitoring result; Adjusting the end position of the abnormal section by using an angle calculation method; Evaluating the effectiveness of the abnormal interval according to the data characteristics of the abnormal interval, and considering the abnormal interval as invalid when the abnormal interval does not accord with the standard; And aiming at the condition that the sequence is at the end or the time span is not met, applying a specified rule to process so as to identify and confirm all effective abnormal sections and obtain an abnormal region.
4. The multi-algorithm comprehensive analysis method for detecting test strip anomalies according to claim 3, wherein the analyzing the trend and the predicted value of each data point in the related data of the test strip to be detected through the sliding window, judging whether the condition for starting anomaly judgment is satisfied, comprises: Establishing a sliding window for calculating a change trend by using a certain data point in the related data of the test paper to be detected, and establishing a sliding window for carrying out abnormality judgment by using the data point at the rear; calculating the average value, average increment and predicted value of the two sliding windows; And according to the relation between the average increment of the sliding window of the calculated change trend and the current value and the predicted value, entering an abnormal judgment mode and recording related information when the relation meets the condition and the descending rate exceeds a set threshold value, wherein the abnormal judgment mode is that the abnormal state is set to be on, and the related information comprises a left side boundary value and a left side boundary index.
5. The multi-algorithm analysis-by-synthesis method for detecting test strip anomalies according to claim 3, wherein the adjusting of the end position of the anomaly interval using an angle calculation method includes: determining the position of the right boundary according to the monitoring result; And optimizing the position of the right boundary by using an angle calculation method to obtain the end position of the abnormal section.
6. The multi-algorithm analysis-by-synthesis method for detecting test strip anomalies according to claim 3, wherein evaluating the validity of the anomaly interval according to the data characteristics of the anomaly interval, when the anomaly interval does not meet a criterion, is regarded as invalid, comprises: Verifying the validity of the ending position of the abnormal section, including checking the descending concentration degree, and calculating whether the descending angle accords with a specific threshold value; and when the ending position of the abnormal section is valid, adding the ending position of the abnormal section into the set, otherwise, resetting the related state variable.
7. The multi-algorithm analysis-by-synthesis method for detecting strip anomalies according to claim 1, wherein repairing the anomaly region and restoring the normal form of the curve to obtain a corrected result comprises: Setting the single increment step length as an integer; calculating the length of the abnormal region; determining a left side boundary index and a right side boundary index of the abnormal region, and recording the left side boundary value and the right side boundary value of the abnormal region; calculating integer increment according to the difference value of the left boundary value and the right boundary value; when the abnormal region is at the end of the curve or the integer increment meets the requirement, correcting all values of the abnormal region by using the LV, wherein the method comprises the following steps: Initializing a left repair value, calculating a margin, adjusting the value in an abnormal section in an arithmetic progression mode when the margin is not 0, setting the value in a specific range as RV if the margin is 0, and fine-tuning the value of a specific position in the abnormal region if the margin is not 0.
8. The multi-algorithm analysis-by-synthesis method for detecting strip anomalies according to claim 1, wherein determining a value window that does not intersect all of the anomaly regions according to the correction result to obtain a value position calculation result includes: Recording the length of the whole sequence of the correction result; traversing forward from the last element of the sequence, and searching a value window; establishing a set containing a plurality of element indexes recently in each iteration; Checking whether the set has an intersection with any of the anomaly regions; when a window without intersection with any abnormal region is found, the window is confirmed to be a valid value window.
9. The multi-algorithm comprehensive analysis method for detecting test strip anomalies according to claim 1, wherein the constructing a model according to the anomaly region, the correction result and the value position calculation result and matching with historical cases, calculating a final detection result and a confidence level thereof, comprises: analyzing and quantifying abnormal distribution, morphology, repair proportion, value position and sample type according to the abnormal region, the correction result and the value position calculation result to obtain multidimensional characteristics; Constructing a comprehensive model based on the multidimensional features to describe the overall features of the current detection situation; searching a historical case model with the similarity meeting the requirement with the comprehensive model, and calculating a final detection result according to an algorithm and parameters; And combining the confidence coefficient of the final detection result according to a preset rule to determine a high confidence coefficient result, a low confidence coefficient suggestion or invalid result prompt content.
10. A multi-algorithm integrated analysis system for detecting test strip anomalies, comprising: the acquisition unit is used for acquiring the related data of the B line and the C line of the test paper to be detected; The difference calculation unit is used for judging whether the test paper to be detected has test paper strip position deviation or not by comparing the average value difference at the tail of the B line and the C line according to the related data of the B line and the C line; the invalidation unit is used for judging that the detection result is invalid if the test paper to be detected has the position deviation of the test paper strip; the analysis unit is used for analyzing photoelectric signal change caused by error or excessive sample adding sequence if the test paper to be detected does not have the position deviation of the test paper strip, and extracting and verifying an abnormal section to obtain an abnormal region; The repairing unit is used for repairing the abnormal region and recovering the normal form of the curve so as to obtain a correction result; The position calculation unit is used for determining a value window which has no intersection with all the abnormal areas according to the correction result so as to obtain a value position calculation result; and the multidimensional analysis unit is used for constructing a model according to the abnormal region, the correction result and the value position calculation result and matching with the historical case to calculate a final detection result and the confidence coefficient thereof.

Description

Multi-algorithm comprehensive analysis method and system for detecting abnormality of test strip Technical Field The invention relates to a test paper abnormality analysis method, in particular to a multi-algorithm comprehensive analysis method and a system for detecting test paper strip abnormality. Background In the field of modern biomedical detection, test strips are widely used as a convenient and rapid detection tool. However, many challenges are often faced during their use. For example, in the practical application process, problems such as position deviation and misoperation often occur when the test paper is matched with an electronic detection instrument, so that the final detection result is affected. In addition, the user may have the situation of wrong sample adding sequence or excessive sample adding amount in the actual operation, and these all can cause repeated chromatographic phenomena on the test strip, so as to further interfere with the normal detection flow. The existing technical means can only identify and process the abnormality of a specific type, and a systematic and comprehensive solution is lacking to cope with all the potential problems. There are some technical solutions on the market that try to solve these problems. For example, some devices employ sensor monitoring techniques to monitor the position change of a test strip in real time and stop detection immediately upon detection of an abnormality, and some systems attempt to identify an abnormality caused by a user's mishandling by analyzing the trend of changes in photoelectric sensor data. Although these methods improve the accuracy of detection to some extent, they generally suffer from drawbacks such as low accuracy, high false positive rate, and inability to effectively repair the identified abnormal region. In addition, most of the existing repair algorithms are based on simple interpolation or smoothing processing, and it is difficult to completely recover the real change characteristics of the original data, so that the subsequent data analysis and result interpretation are affected. Therefore, a new method is necessary to be designed, so that not only can various types of abnormal conditions be effectively identified, but also the value-taking position can be dynamically optimized to improve the data confidence, thereby remarkably improving the accuracy and reliability of the detection result of the biomedical test strip, and solving the defects of low accuracy, high false alarm rate and limited data restoration capability in the prior art when the problems of position deviation, misoperation of a user and the like in the detection process of the biomedical test strip are solved. Disclosure of Invention The invention aims to overcome the defects of the prior art and provides a multi-algorithm comprehensive analysis method and a system for detecting the abnormality of a test strip. In order to achieve the purpose, the invention adopts the following technical scheme that the multi-algorithm comprehensive analysis method for detecting the abnormality of the test strip comprises the following steps: acquiring related data of a B line and a C line of test paper to be detected; According to the related data of the B line and the C line, judging whether the test paper to be detected has test paper strip position deviation or not by comparing the average value difference of the tail ends of the B line and the C line; If the test paper to be detected has the test paper strip position deviation, judging that the detection result is invalid; If the test paper to be detected does not have the test paper strip position deviation, analyzing photoelectric signal change caused by error or excessive sample adding sequence, and extracting and verifying an abnormal section to obtain an abnormal region; repairing the abnormal region, and restoring the normal form of the curve to obtain a correction result; determining a value window which has no intersection with all the abnormal areas according to the correction result to obtain a value position calculation result; And constructing a model according to the abnormal region, the correction result and the value position calculation result, matching the model with a historical case, and calculating a final detection result and the confidence coefficient thereof. The invention also provides a multi-algorithm comprehensive analysis system for detecting the abnormality of the test paper strip, which comprises: the acquisition unit is used for acquiring the related data of the B line and the C line of the test paper to be detected; The difference calculation unit is used for judging whether the test paper to be detected has test paper strip position deviation or not by comparing the average value difference at the tail of the B line and the C line according to the related data of the B line and the C line; the invalidation unit is used for judging that the detection result is invalid if th