Container transmittance difference correction method

Abstract

The invention relates to the technical field of light intensity detection, in particular to a container transmittance difference correction method, which comprises the following steps: calculating blank sample transmittance, collecting blank sample transmission images of blank samples, mapping the blank sample transmittance and the corresponding blank sample transmission images to form a blank sample image transmittance data set, and correcting the blank sample image transmittance data set; inputting the image transmittance into a deep learning neural network model for training to obtain an image transmittance correction model; acquiring a test sample transmission image of the test sample, and inputting the test sample transmission image into the model to obtain the transmittance; and dividing the detected transmission light intensity of the test sample by the transmittance corresponding to the obtained transmission image of the test sample. According to the invention, based on the error correction model of deep learning, the image of the light passing surface is measured to correct the influence of transmittance difference caused by scratches, defects, surface shape difference and the like of the light passing surface on light intensity detection, a blank state of the container is not needed, and the detection reliability is improved while the manufacturing requirement and cost of the container are reduced.

Description

Technical Field

[0001] This invention belongs to the field of light intensity detection technology, and in particular relates to a method for correcting differences in container transmittance. Background Technology

[0002] Light intensity detection is an indispensable core technology in modern analytical chemistry. It is based on the principle that the intensity of light changes after a substance interacts with electromagnetic waves (light), allowing for qualitative or quantitative analysis of the substance. This includes: 1. Ultraviolet-visible absorption spectroscopy is based on the principle that molecules absorb ultraviolet or visible light, causing electronic energy level transitions that lead to a decrease in the intensity of light at a specific wavelength. 2. Molecular fluorescence spectroscopy is based on the fact that after a specific substance (fluorophore) absorbs light of a specific wavelength (excitation light), the molecule transitions to an excited state and emits light of a longer wavelength (emission light) when returning to the ground state. The intensity of the emitted light is detected for analysis. 3. Chemiluminescence is a form of light emitted during chemical reactions or biological metabolism, and its quantitative analysis is performed by detecting the intensity of the emitted light. 4. Based on the fact that the atoms of the analyte absorb light of a specific wavelength under high temperature conditions, the atomic absorption spectrum of the light intensity decreases and the atomic emission spectrum of the analyte atoms emitted light of a specific wavelength when excited under high temperature conditions and returning to the ground state.

[0003] The above technologies all belong to the category of quantitative analysis methods that detect light intensity, and are widely used in fields such as environmental monitoring, food and drug safety analysis, industrial production analysis, and clinical diagnosis.

[0004] When light intensity detection is used for quantitative analysis, the difference in transmittance between different containers can reduce detection accuracy and affect analytical sensitivity. Current common practices primarily mitigate this impact by pre-testing container transmittance and improving transmittance consistency. Pre-testing transmittance requires the container to be in a blank state before testing; without a blank state (such as in a fully automated microfluidic reaction vessel pre-packaged with reagents), the test cannot be performed. Improving transmittance consistency requires high standards throughout the container's manufacturing, packaging, transportation, and usage processes, significantly increasing testing costs.

[0005] Patent No. ZL201810886145.4 discloses a spectrophotometric multi-wavelength detection device and its detection method, which corrects for differences in the transmittance of containers. This method requires the analyte to have a specific absorption peak within a corresponding wavelength range relative to the background material. This limits its application scenarios.

[0006] Therefore, existing methods for addressing the reduced detection accuracy caused by differences in transmittance suffer from limitations in application scenarios and a trade-off between cost and accuracy. Summary of the Invention

[0007] In view of this, the present invention aims to provide a method for correcting the transmittance difference of containers. Based on a deep learning-based error correction model, the method measures the image of the light-transmitting surface to correct the influence of transmittance differences caused by scratches, defects, etc. on the light-transmitting surface of the container. This method can be used even when the container is not blank, reducing the consistency requirements of the container transmittance, ensuring detection accuracy, and greatly reducing detection costs.

[0008] To achieve the above objectives, the technical solution created by this invention is implemented as follows: A method for correcting differences in container permeability includes the following steps: S1: Construct a blank sample set, and detect the incident light intensity I of each blank sample in the blank sample set using a light intensity detection device. 入射 And the transmitted light intensity I of each blank sample 透射 And calculate the transmittance T of the blank sample, using the following formula: T= I 透射 / I 入射 ; S2: Acquire the transmission image of the light-transmitting surface of each blank sample in the blank sample set through the light-transmitting surface image acquisition device, and perform image normalization processing on all acquired blank sample transmission images to obtain blank sample normalized image information. S3: Map the transmittance of blank samples to the normalized image information of blank samples to form an image transmittance dataset. S4: Input the image transmittance dataset into a deep learning neural network model for training to obtain an image transmittance correction model; S5: Acquire the transmission image of the test sample on the light-transmitting surface using the light-transmitting surface image acquisition device, detect the transmitted light intensity of the test sample using the light intensity detection device, and perform image normalization processing on the transmission image of the test sample to obtain the normalized image information of the test sample. S6: Input the normalized image information of the test sample into the image transmittance correction model to obtain the transmittance corresponding to the light-transmitting surface of the test sample; S7: Divide the detected transmitted light intensity of the test sample by the transmittance corresponding to the light-transmitting surface of the test sample to obtain the light intensity data that theoretically does not include the influence of the transmittance of the test sample container.

[0009] A light-transmitting surface image acquisition device is applied to a method for correcting differences in the transmittance of a container. The device is characterized by comprising a first light source, an image acquisition field aperture, and an image detector. The first light source is located at the light-incident end of the sample cup, the image detector is located at the light-outceasing end of the sample cup, and the image acquisition field aperture is disposed between the first light source and the image detector. When the sample cup is empty of solvent, it serves as a blank sample, and the image detector is used to acquire the transmission image of the blank sample. When the sample cup is filled with solvent, it serves as a test sample, and the image detector is used to acquire the transmission image of the test sample from the transmissive surface.

[0010] Furthermore, the image acquisition field stop has a cylindrical or annular structure.

[0011] Furthermore, the corresponding area on the sample cup and located inside the image acquisition field aperture is the image acquisition area.

[0012] A light intensity detection device is used in a method for correcting differences in the transmittance of a container. The device includes a second light source, a light intensity detection field stop, and a light intensity detector. The second light source is located at the light-incident end of the sample cup, the light intensity detector is located at the light-outceasing end of the sample cup, and the light intensity detection field stop is disposed between the first light source and the light intensity detector.

[0013] When the sample cup is not filled with solvent, it serves as a blank sample. The light intensity detector is used to collect the transmitted light intensity of the blank sample on the light-transmitting surface. When the second sample cup is filled with solvent, it serves as a test sample. The image detector is used to detect the transmitted light intensity of the test sample on the light-transmitting surface. When no sample cup is placed, the incident light intensity is detected.

[0014] Furthermore, the area on the sample cup that is located inside the light intensity detection field stop is the light intensity detection area.

[0015] Furthermore, the light intensity detection field stop has a cylindrical or annular structure.

[0016] Furthermore, the image acquisition area covers the light intensity detection area.

[0017] Compared with the prior art, the present invention can achieve the following beneficial effects: (1) The image transmittance correction model created in this invention predicts the transmittance by using an image containing the light-transmitting surface features of the container, and is used to correct the influence of transmittance differences on light intensity detection. The detection process does not require the container to be in a blank state.

[0018] (2) The image transmittance correction model created in this invention has no requirements on the spectral characteristics of the target substance to be tested and has wide applicability.

[0019] (3) The method of the present invention can acquire images of the corresponding area before, during or after light intensity detection, thereby correcting the transmittance difference, which can greatly reduce the requirements for container consistency and effectively reduce the cost of the detection container and the detection cost. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments and descriptions of the invention are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 A flowchart illustrating a container permeability difference correction method according to an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a light-transmitting surface image acquisition device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a light intensity detection device according to an embodiment of the present invention.

[0021] Explanation of reference numerals in the attached figures: 10. First light source; 11. Image acquisition field stop; 12. Image detector; 13. Image acquisition area; 20. Second light source; 21. Light intensity detection field stop; 22. Light intensity detector; 23. Light intensity detection area. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and do not constitute a limitation thereof.

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0024] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0025] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0026] The invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] like Figure 1 As shown, a method for correcting differences in container permeability includes the following steps: S1: Construct a blank sample set, and use a light intensity detection device to detect the incident light intensity I of at least 25,000 blank samples in the blank sample set. 入射 And the transmitted light intensity I of each blank sample 透射 And calculate the transmittance T of the blank sample, using the following formula: T= I 透射 / I 入射 ; S2: Acquire at least 25,000 blank sample transmission images of the light-transmitting surface of the blank sample set using a light-transmitting surface image acquisition device, and perform image normalization processing on all acquired blank sample transmission images to obtain blank sample normalized image information. Image normalization is a preprocessing step that converts the pixel values ​​of image data to a specific range, usually 0 to 1 or -1 to 1. S3: Map the transmittance of at least 20,000 blank samples to the normalized image information of the blank samples to form an image transmittance dataset; S4: Input the image transmittance dataset into the deep learning neural network model for training to obtain the image transmittance correction model, and optimize the image transmittance correction model by cropping, quantizing and compressing the data. Export the optimized model in a format that can be deployed on the data processing hardware platform, that is, deploy the trained image transmittance correction model to the light intensity detection data processing module. The transmittance of at least 5000 blank samples in step S3 and the normalized image information of blank samples are used to verify the image transmittance correction model in step S4. The difference between the predicted transmittance obtained from the input image and the transmittance obtained from the test should be within ±5%. If it exceeds this range, the transmission images of blank samples and the transmittance data of blank samples in the image transmittance dataset are added for further training. After training, the model is verified again until the transmittance difference is within ±5%. The model training is then complete. The deep learning neural network model is any one or a combination of modified models such as CNN, Transformer, or RNN.

[0028] S5: Acquire a transmission image of the test sample's transmission surface using a light-transmitting surface image acquisition device. Detect the transmitted light intensity of the test sample using a light intensity detection device, i.e., the light intensity data of the test sample. Perform image normalization processing on the transmission image of the test sample to obtain the normalized image information of the test sample. S6: Input the normalized image information of the test sample into the image transmittance correction model to obtain the transmittance corresponding to the transmission image of the test sample; S7: Divide the detected transmitted light intensity of the test sample by the transmittance corresponding to the obtained transmission image of the test sample to obtain the light intensity data that theoretically does not include the influence of the transmittance of the test sample container. The deviation of the obtained light intensity data is within ±5%, which is consistent with the correction result. The above light intensity value can be used for subsequent analysis. Since this intensity value is theoretically unaffected by the container transmittance, it can effectively correct the influence of container transmittance differences on light intensity detection, thereby improving detection accuracy, repeatability, and sensitivity.

[0029] Convolutional Neural Networks (CNNs) are a class of deep learning models specifically designed for processing grid-like data such as images. Transformers, based on self-attention mechanisms, can process long sequences of data in parallel and are widely used in Natural Language Processing (NLP) tasks, such as machine translation, text summarization, and generative models (e.g., GPT). Their core components include multi-head self-attention mechanisms, positional encoding, and feedforward neural networks. Recurrent Neural Networks (RNNs) are a class of artificial neural networks with internal loop connections, used for processing sequential data. Their most significant characteristic is the presence of loops within the network, allowing information to circulate within the network, enabling the storage and processing of sequential information.

[0030] Steps S1-S4 are the process of obtaining the image transmittance correction model based on neural network training, and steps S5-S7 are the process of correcting the transmittance differences of the test sample containers.

[0031] Specifically, considering the influence of container transmittance, the actual measured light intensity during the light intensity detection process is: I 检测 =I 真值 *T 容器 ; Where I 检测 To test the transmitted light intensity of the sample, I 真值 To determine the true value of transmitted light intensity of the test sample unaffected by container transmittance, T 容器 To test the transmittance of the sample container; The transmittance of different containers varies due to differences in manufacturing precision. This transmittance difference directly affects the light intensity detection result as an error. If the container transmittance is known, the true value of the light intensity can be calculated. The specific calculation formula is as follows: I 检测 / T 容器 =I 真值 ; This application predicts the transmittance of a container based on an image, and uses the predicted T value for correction. When the prediction accuracy of the T value for a single container is within ±5%, the corrected light intensity range is: I 检测 / (0.95* T) 容器 )~ I 检测 / (1.05*T) 容器 ), that is: (0.952~1.053)I 真值 The limit deviation of the same sample in different containers is about 10%.

[0032] Without correction, the deviation of T between different containers produced by ordinary injection molding process can reach 30% or even more than 50%. This deviation is directly transmitted to the light intensity results. The limit deviation of light intensity data of the same sample from different containers can reach more than 30%, which greatly affects the accuracy and reliability of the test results.

[0033] like Figure 2 As shown, a light-transmitting surface image acquisition device is applied to a method for correcting differences in the transmittance of a container. The device includes a first light source 10, an image acquisition field stop 11, and an image detector 12. The first light source 10 is located at the light-incident end of the sample cup, and the image detector 13 is located at the light-exit end of the sample cup. The image acquisition field stop 11 is disposed between the first light source 10 and the image detector 12. That is, the image acquisition field stop 11 can be disposed between the first light source 10 and the sample cup, or it can be disposed between the sample cup and the image detector 12.

[0034] The image acquisition field stop 11 has a cylindrical or ring-shaped structure, specifically a circular or rectangular ring structure.

[0035] When the sample cup is not filled with solvent, it serves as a blank sample, and the image detector 12 is used to acquire the transmission image of the blank sample on the transparent surface. When the sample cup is filled with solvent, it serves as a test sample, and the image detector 12 is used to acquire the transmission image of the test sample on the transparent surface.

[0036] The area on the sample cup that corresponds to the inside of the image acquisition field stop 11 is the image acquisition area 13.

[0037] In the specific acquisition process, the first light source 10, the sample cup image acquisition field aperture 11, and the image detector 11 are arranged along the same optical path. The first light source 10 emits light and illuminates the sample cup. The light passes through the sample cup and the image acquisition field aperture 11 limits the image acquisition area. Finally, the image detector 12 obtains the transmitted image in the image acquisition area 13.

[0038] like Figure 3 As shown, a light intensity detection device is used in a method for correcting differences in the transmittance of a container. The device includes a second light source 20, a light intensity detection field stop 21, and a light intensity detector 22. The second light source 20 is located at the light-incident end of the sample cup, and the light intensity detector 22 is located at the light-excising end of the sample cup. The light intensity detection field stop 21 is disposed between the second light source 20 and the light intensity detector 22. That is, the light intensity detection field stop 21 can be disposed between the second light source 20 and the sample cup, or between the sample cup and the light intensity detector 22.

[0039] The light intensity detection field stop 21 has a cylindrical or ring-shaped structure. Specifically, the light intensity detection field stop 21 has a circular or rectangular ring structure.

[0040] When the sample cup is not filled with solvent, it serves as a blank sample. The light intensity detector 22 is used to detect the transmitted light intensity of the blank sample on the transparent surface. When the sample cup is filled with solvent, it serves as a test sample. The image detector 12 is used to detect the transmitted light intensity of the test sample on the transparent surface of the blank sample. When the sample cup is not placed, the incident light intensity is detected.

[0041] The area on the sample cup corresponding to the inside of the light intensity detection field stop 21 is the light intensity detection area 23.

[0042] During the specific data collection, the second light source 20, the sample cup, the light intensity detection field stop 21, and the light intensity detector 22 are arranged along the same optical path. The second light source 20 emits light and illuminates the sample cup. The light passes through the sample cup and is limited to the light intensity detection area by the light intensity detection field stop 21. Finally, the transmitted light intensity in the light intensity detection area 23 is obtained by the light intensity detector 22.

[0043] The image acquisition area 13 completely covers the light intensity detection area 23, meaning the area of ​​the image acquisition area 13 is greater than or equal to the light intensity detection area 23. This ensures that each light-transmitting surface with detected light intensity has a corresponding image, facilitating the formation of an effective transmission image and transmittance dataset.

[0044] In actual testing, only the first light source 10, the image acquisition field aperture 11, and the image detector 12 need to be used, as illustrated. Figure 2 The device is used for detection. The sample cup is placed in the optical path, and an image is acquired. The light intensity values ​​of all pixels in the acquired image are superimposed to obtain the transmitted light intensity value of the sample. At this time, the image acquisition area 13 and the light intensity detection area 23 coincide and are equal in size. Alternatively, the light intensity values ​​of some pixels in the acquired image can be superimposed to obtain the transmitted light intensity value of the sample. In this case, the image acquisition area 13 covers the light intensity detection area 23, and the area of ​​the image acquisition area 13 is larger than that of the light intensity detection area 23.

[0045] It should be understood that the various forms of processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this invention disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this invention can be achieved, and this is not limited herein.

[0046] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be covered within the scope of protection of this invention.

Claims

1. A method for correcting differences in container permeability, characterized in that, Includes the following steps: S1: Construct a blank sample set, and detect the incident light intensity I of each blank sample in the blank sample set using a light intensity detection device. 入射 And the transmitted light intensity I of each blank sample 透射 And calculate the transmittance T of the blank sample, using the following formula: T= I 透射 / I 入射 ; S2: Acquire the transmission image of the light-transmitting surface of each blank sample in the blank sample set using the light-transmitting surface image acquisition device, and perform image normalization processing on all acquired blank sample transmission images to obtain blank sample normalized image information. S3: Map the transmittance of blank samples to the normalized image information of blank samples to form an image transmittance dataset. S4: Input the image transmittance dataset into a deep learning neural network model for training to obtain an image transmittance correction model; S5: Acquire the transmission image of the test sample on the light-transmitting surface of the test sample through the light-transmitting surface image acquisition device, detect the transmission light intensity of the test sample through the light intensity detection device, and perform image normalization processing on the transmission image of the test sample to obtain normalized image information of the test sample. S6: Input the normalized image information of the test sample into the image transmittance correction model to obtain the transmittance corresponding to the light-transmitting surface of the test sample; S7: Divide the detected transmitted light intensity of the test sample by the transmittance corresponding to the light-transmitting surface of the test sample to obtain light intensity data that theoretically does not include the influence of the transmittance of the test sample container.

2. A light-transmitting surface image acquisition device, applied to the container transmittance difference correction method described in claim 1, characterized in that: It includes a first light source (10), an image acquisition field stop (11), and an image detector (12). The first light source (10) is located at the light-incident end of the sample cup, the image detector (12) is located at the light-outcident end of the sample cup, and the image acquisition field stop (11) is disposed between the first light source (10) and the image detector (12). When the sample cup is not filled with solvent, it is used as a blank sample, and the image detector (12) is used to acquire the blank sample transmission image; when the sample cup is filled with solvent, it is used as a test sample, and the image detector (13) is used to acquire the test sample transmission image of the light-transmitting surface on the test sample.

3. The light-transmitting surface image acquisition device according to claim 2, characterized in that: The corresponding area on the sample cup and inside the image acquisition field stop (11) is the image acquisition area (13).

4. A light intensity detection device, applied to the container transmittance difference correction method described in claim 1, characterized in that: It includes a second light source (20), a light intensity detection field stop (21), and a light intensity detector (22). The second light source (20) is located at the light-incident end of the sample cup, the light intensity detector (22) is located at the light-outcident end of the sample cup, and the light intensity detection field stop (21) is disposed between the first light source (10) and the light intensity detector (22). When the sample cup is not filled with solvent, it is used as a blank sample. The light intensity detector (22) is used to detect the light intensity transmitted by the blank sample on the light-transmitting surface of the blank sample. When the sample cup is filled with solvent, it is used as a test sample. The image detector (12) is used to detect the light intensity transmitted by the test sample on the light-transmitting surface of the test sample. When the sample cup is not placed, the incident light intensity is detected.

5. The light-transmitting surface image acquisition device according to claim 4, characterized in that: The corresponding area on the sample cup and inside the light intensity detection field stop (21) is the light intensity detection area (23).

Images