Optical filter rapid detection system and method thereof

Abstract

The invention provides an optical filter rapid detection system and a method thereof. The system comprises a xenon lamp; the filtering wheel bracket is used for supporting and fixing the filtering wheel; the filtering wheel controller is used for receiving the instruction and controlling the rotation of the filtering wheel; the motor is used for providing power for the filtering wheel; a plurality of optical filters are arranged on the filtering wheel; the lens is used for focusing and imaging and projecting the light passing through the optical filter to the camera; the camera is used for capturing the light passing through the lens and the optical filter and converting the light into a digital image; the image work station is used for processing the image data transmitted by the camera and carrying out image analysis and processing operation; and the camera obscura is used for accommodating the xenon lamp, the filtering wheel bracket, the filtering wheel, the filtering wheel controller, the motor, the filtering wheel, the lens and the camera. According to the invention, the spectrum information of the installed optical filter can be rapidly detected, and whether the installation is correct can be rapidly judged, so that the detection efficiency is improved, and the requirement of mass detection is met.

Description

Technical Field

[0001] The present invention relates to the technical field of optical filter detection, and in particular to an optical filter rapid detection system and an optical filter rapid detection method using the system. Background Art

[0002] Small animal live optical imaging is the most common imaging method for small animal live experiments, including bioluminescence imaging and fluorescent labeling imaging. Bioluminescence imaging technology refers to the use of reporter gene-luciferase gene expression in small mammals to produce luciferase protein and its small molecule substrate luciferin in the presence of oxygen and Mg2+ ions to consume ATP and undergo oxidation reaction, converting part of the chemical energy into visible light energy. Then, sensitive CCD equipment is used in vitro to form images. Bioluminescence is essentially a chemical fluorescence. Firefly luciferase can release visible light photons with a wide range of wavelengths in the process of oxidizing its unique substrate luciferin, with an average wavelength of 560nm (460-630nm), including important red light components with wavelengths exceeding 600nm. In mammals, hemoglobin is the main component that absorbs visible light and can absorb most of the visible light in the mid-blue-green band; water and lipids mainly absorb infrared light, but both have poor absorption capacity for red light to near-infrared light with a wavelength of 590-800nm. Therefore, although red light with a wavelength of more than 600nm is partially scattered and consumed, most of it can penetrate mammalian tissues and be detected by highly sensitive CCDs. Excitation fluorescence imaging is to excite fluorescent groups to a high energy state through excitation light, and then generate emission light. The principle of excitation fluorescence imaging can be described as follows: when exogenous light is irradiated on biological tissues with fluorescent groups, the fluorescent groups absorb light energy and cause electrons to transition to an excited state. In the process of electrons returning from the excited state to the ground state, fluorescence is released. The fluorescence moves toward the red end compared to the absorbed light, that is, the emitted fluorescence has lower energy than the absorbed exogenous light. The fluorescence propagates in the tissue and part of it reaches the body surface. The fluorescence emitted from the body surface is received by the detector, thus forming a fluorescence image.

[0003] Small animal live imaging equipment is mainly composed of a dark box, an excitation light source, a filter, a lens, a highly sensitive camera and an image processing workstation. Among them, the filter is a key component in the small animal live imaging equipment. It is generally installed in front of the optical lens or between the optical lens and the camera. It can filter out unnecessary optical signals in the imaging, making the imaging effect more accurate. The small animal live imaging equipment uses a set of bandpass filters with a wavelength range of 400nm-900nm. During the imaging process, if the wrong filter is used, it will cause imaging failure or output an incorrect image. In order to ensure the accuracy of the image and data, it is necessary to ensure that each filter is installed correctly after installation.

[0004] According to the principle of three primary colors, any light visible to the human eye can be made up of three monochromatic colors, red (R), green (G), and blue (B), mixed in appropriate proportions. When using a color camera for imaging, the visible light collected by each sensor can be represented by RGB values. The proportions of the three primary colors of red, green, and blue are different for visible light of different bands, and the corresponding RGB values ​​obtained by the color camera are also different. Based on the principles of spectroscopy and colorimetry, based on the 1931 CIE standard colorimetric system and BP neural network, a model can be established by mapping the relationship between RGB values ​​and light wavelengths, and the light wavelength can be obtained through RGB values. Therefore, the RGB values ​​collected by the camera can be input into the relationship model between RGB and spectral wavelength to obtain the visible light band collected by the camera through the filter.

[0005] After the filter is installed, it is impossible to check whether the filter is installed correctly by appearance, but the spectral information transmitted by the filter can be detected to match the filter model, thereby judging whether the filter is installed correctly. The current method of detecting the spectral information of the filter is to use a spectrometer to detect the spectral range and spectral characteristics of the filter. Although this method can accurately obtain the spectral range of the filter, it is cumbersome to operate, slow, and inconvenient to install and test. Especially after mass production, it cannot meet the needs of rapid detection. Summary of the invention

[0006] In order to solve various problems existing in the method of detecting filters in the prior art, the purpose of the present invention is to provide a filter rapid detection system and method, which can quickly detect the spectral information of installed filters and quickly determine whether they are installed correctly, thereby improving the detection efficiency and meeting the requirements of large-scale detection.

[0007] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0008] A filter rapid detection system, comprising:

[0009] A xenon lamp as a light source is used to provide illumination for the detection system;

[0010] A filter wheel bracket is used to support and fix the filter wheel to ensure that the filter wheel can rotate smoothly and select a suitable filter;

[0011] A filter wheel controller, for receiving instructions and controlling the rotation of the filter wheel to select a correct filter;

[0012] A motor, used for providing power to the filter wheel so that it can rotate;

[0013] A filter wheel, on which a plurality of filters are mounted, for selectively transmitting or blocking light of a specific wavelength to achieve automatic switching of the filters;

[0014] The lens is used for focusing and imaging, projecting the light through the filter onto the camera;

[0015] A camera, which captures light passing through lenses and filters and converts it into a digital image;

[0016] Image workstation, used to process image data transmitted from the camera and perform image analysis and processing operations;

[0017] The dark box is used to accommodate the xenon lamp, filter wheel bracket, filter wheel, filter wheel controller, motor, filter wheel, lens and camera, and provides a light-free environment to avoid interference of external light on the detection system.

[0018] A method for rapid detection of optical filters, which uses the above-mentioned rapid detection system for optical filters to perform detection, comprises the following steps:

[0019] The image workstation starts and is ready to receive data. The camera is initialized and switched to RGB mode, ready to capture images.

[0020] Turn on the xenon lamp as the light source, position the filter wheel to the initial position, and send the filter position information to the imaging workstation;

[0021] The light passing through the filter enters the lens, and the camera collects and outputs RGB image data to the image workstation, which receives and saves the image data and associates it with the location information.

[0022] After completing the first filter image acquisition, the image workstation communicates with the filter wheel controller through serial communication, controls the filter wheel to move to the next filter position, and repeats the acquisition process until the image data of all filters are acquired;

[0023] After the data collection is completed, the image workstation processes the image data collected by each filter in turn, and reconstructs the collected RGB values ​​into corresponding spectral information through the mapping relationship model between RGB and spectral wavelength, compares the obtained spectral information with the spectral information in the database, screens out filters that meet the requirements, performs statistical analysis on the types and proportions of the RGB values ​​of the filters that meet the requirements, and fits the statistical analysis results with the spectral characteristic curve of the filter to confirm the model of the filter;

[0024] After all filter models are confirmed, compare the test results with the installation information and print the test results.

[0025] According to a method for rapid detection of optical filters provided by the present invention, a mapping relationship model corresponding to RGB and spectral wavelength is established, including:

[0026] On the basis of the RGB system, three ideal primary colors are selected to replace the actual three primary colors, thereby converting the spectral tristimulus values ​​in the CIE-RGB system into and the chromaticity coordinates r, g, and b all become positive;

[0027] To find the color tristimulus values ​​X, Y, Z of a known wavelength, the sum Σ is used instead of the integral ∫ in actual calculation, which is expressed as the following formula:

[0028]

[0029] Where K is the adjustment factor, S(λ) is the relative spectral power distribution of a light source to be measured are the spectral tristimulus values.

[0030] According to a fast filter detection method provided by the present invention, the CIE-XYZ spectral chromaticity coordinates x, y, z are calculated and expressed as the following formula:

[0031]

[0032] Among them, an artificial neural network is used to establish a mapping relationship model between r, g, b values ​​and the color tristimulus values ​​(X, Y, Z) defined by the CIE standard chromaticity system.

[0033] According to a method for rapid detection of optical filters provided by the present invention, when an artificial neural network is used to establish a mapping relationship model, the method specifically includes:

[0034] Collect r, g, b values ​​and corresponding X, Y, Z three stimulus value data as training samples;

[0035] Design an artificial neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer accepts r, g, b values ​​as input, and the output layer outputs three stimulus values ​​of X, Y, and Z;

[0036] The artificial neural network is trained using training samples, and the network output is made close to the real X, Y, and Z three stimulus values ​​by adjusting the weights and bias parameters in the network;

[0037] The trained artificial neural network is tested using validation samples to evaluate its performance; if the test results meet the requirements, the network is used in actual color conversion applications.

[0038] According to a method for rapid detection of optical filters provided by the present invention, when collecting spectral characteristic curves of different optical filters, the method comprises:

[0039] The RGB values ​​of all filters are collected 10 times for each filter, and the spectral characteristic curve of the filter is fitted based on the data. Thus, each filter has a specific spectral characteristic curve;

[0040] Install the existing filter to be tested on the filter wheel and irradiate it with a xenon lamp. The spectral wavelength range of the xenon lamp is 400nm-1000nm, including the continuous visible light band and the near-infrared band.

[0041] After the light from the xenon lamp hits the filter, since the filter is a bandpass filter, only light of a specific wavelength can pass through the filter, and the light of the specific wavelength passes through the lens and is then collected by the camera;

[0042] The camera transmits the collected image data to the image workstation, where the camera has two channels, one of which is the RGB channel for collecting spectral information of visible light, and the other is the IR channel for collecting spectral information in the near-infrared region.

[0043] According to a fast filter detection method provided by the present invention, the RGB value and the proportion of each RGB value in each image data are counted by an image workstation, and all spectral bands and the proportion of each band are reconstructed according to the RGB value;

[0044] The 10 sets of image data collected are recorded, and the spectral characteristic curve of the light passing through the filter is fitted according to the 10 sets of data and stored in the database.

[0045] According to a method for rapid detection of optical filters provided by the present invention, the obtained spectral information is compared with the spectral information in a database to screen out optical filters that meet the requirements, including:

[0046] The obtained spectral information and the spectral information in the database are standardized to ensure that the two are compared on the same scale;

[0047] Use a registration algorithm or other alignment method to achieve data alignment so that the peaks and valleys of different spectral data can be aligned;

[0048] The correlation coefficient between the calculated spectral information and the spectral information in the database is set to a threshold. When the correlation coefficient is greater than the threshold, the filter is considered to meet the requirements. The closer the correlation coefficient is to 1, the higher the similarity between the two spectral information.

[0049] According to a method for rapid detection of optical filters provided by the present invention, the statistical analysis of the types and proportions of RGB values ​​of optical filters that meet the requirements includes:

[0050] For each selected filter, extract its corresponding RGB value, and perform deduplication processing on the extracted RGB value;

[0051] Classify RGB values ​​into different color ranges or categories as needed;

[0052] Count the number of RGB values ​​in each color range or category. For each color range or category, calculate the proportion of its RGB value in the total data set, where the proportion is calculated by dividing the number of RGB values ​​in a color range or category by the total number of RGB values.

[0053] Use charts to visualize the types of RGB values ​​and their proportions.

[0054] According to a method for rapid detection of optical filters provided by the present invention, when fitting the statistical analysis results with the spectral characteristic curve of the optical filter, the fitting effect is evaluated by calculating the goodness of fit;

[0055] The fitted spectral characteristic curve is compared with the spectral characteristic curve of the known filter model, which includes whether the reflectivity / transmittance at the key wavelength is consistent and the similarity of the overall spectral shape.

[0056] It can be seen that compared with the prior art, the present invention has the following beneficial effects:

[0057] 1. The system of the present invention is composed of a xenon lamp, a filter wheel bracket, a lens, a camera and a computer. It uses RGB values ​​to measure the visible light band passing through the filter, matches the existing filter data model, and thus confirms the filter model, and then compares it with the installation information to confirm whether the installation is correct.

[0058] 2. The method provided by the present invention can improve the detection efficiency of filters used in small animal live imaging equipment, and can quickly detect whether the installed filters are correctly installed to meet the requirements of batch detection; it reduces the cost of detection equipment and improves the automation of detection equipment.

[0059] 3. The present invention uses the mapping relationship model between RGB and spectral wavelength to reconstruct the collected RGB values ​​into corresponding spectral information, thereby improving the accuracy of spectral information reconstruction. By comparing with the spectral information in the database, filters that meet the requirements can be accurately screened out.

[0060] 4. The present invention can detect all filters installed on the filter wheel to ensure that none are missed, and detailed image data collection and processing are performed on each filter to ensure the comprehensiveness of the detection.

[0061] 5. The automated detection of the present invention reduces labor costs and time costs, improves detection efficiency, and accurate detection results help reduce experimental failures or equipment failures caused by poor filter performance, thereby reducing overall costs.

[0062] 6. The present invention compares the test results with the installation information and prints the test results, making the test results more intuitive and easy to understand, allowing users to quickly understand the performance and model of the filter, and providing convenience for subsequent use and maintenance.

[0063] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a structural schematic diagram of an embodiment of a filter rapid detection system of the present invention.

[0065] Figure 2 The figure is a schematic diagram of the principle of an embodiment of a filter rapid detection system of the present invention.

[0066] Figure 3 It is a structural schematic diagram of a filter wheel in an embodiment of a filter rapid detection system of the present invention.

[0067] Figure 4 It is a structural schematic diagram of a filter wheel bracket in an embodiment of a filter rapid detection system of the present invention.

[0068] Figure 5 It is a schematic diagram of the spectral characteristics of visible light in an embodiment of a filter rapid detection method of the present invention.

[0069] Figure 6 It is a CIE1931 XY chromaticity diagram in an embodiment of a method for rapid detection of optical filters of the present invention.

[0070] Figure 7 The present invention is a flowchart of a method for rapid detection of optical filters. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0073] According to the principle of three primary colors, any light visible to the human eye can be made up of three monochromatic colors, red, green and blue, mixed in appropriate proportions. Figure 5 As shown, each band of visible light can be mixed by two or three of the three primary colors, and the proportions of the three primary colors in each band of visible light are different.

[0074] The RGB color model is a color standard that obtains a variety of colors by changing the three color channels of red, green, and blue and superimposing them on each other. When using a color camera for imaging, the visible light collected by each sensor can be represented by RGB values.

[0075] A filter rapid detection system embodiment

[0076] See also Figures 1 to 4 , this embodiment provides a filter rapid detection system, including:

[0077] A xenon lamp 1 as a light source, used to provide illumination for the detection system;

[0078] The filter wheel bracket 2 is used to support and fix the filter wheel to ensure that the filter wheel can rotate smoothly and select a suitable filter;

[0079] A filter wheel controller 3, used to receive instructions and control the rotation of the filter wheel to select the correct filter;

[0080] The motor 4 is used to provide power to the filter wheel so that it can rotate;

[0081] A filter wheel 5, on which a plurality of filters are mounted for selectively transmitting or blocking light of a specific wavelength to achieve automatic switching of the filters;

[0082] Lens 6, used for focusing and imaging, projects the light through the filter onto the camera;

[0083] A camera 7 for capturing light passing through the lens 6 and the filter and converting it into a digital image;

[0084] The image workstation 8 is used to process the image data transmitted by the camera 7 and perform image analysis and processing operations;

[0085] The dark box 9 is used to accommodate the xenon lamp 1, the filter wheel bracket 2, the filter wheel 5, the filter wheel controller 3, the motor 4, the filter wheel 5, the lens 6 and the camera 7, and provides a light-free environment to avoid interference of external light on the detection system.

[0086] Specifically, the filter rapid detection system is composed of a xenon lamp 11, a filter wheel bracket 2, a filter wheel controller 3, a motor 4, a filter wheel 5, a lens 6, a camera 7, an image workstation 8, a dark box 9, a camera data cable and a serial communication cable. After the filter is installed, install the filter wheel 5 on the Figure 4 On the filter wheel bracket 2 shown, the filter wheel 5 used by the small animal in vivo imaging device has a total of 12 installation positions, and a maximum of 12 filters can be installed. Each filter installation position is marked with a serial number from 0 to 11. Among them, the installation position 0 is the initial position, and the device will be in position 0 after initialization. After the filter is installed on the filter wheel 5, the filter information of each position will be recorded.

[0087] A method for rapid detection of optical filters

[0088] See also Figures 5 to 7 The present invention provides a method for rapid detection of optical filters, which uses the above-mentioned rapid detection system for optical filters to perform detection, and comprises the following steps:

[0089] The image workstation 8 starts and is ready to receive data, the camera 7 is initialized and switched to RGB mode, ready to capture images;

[0090] Turn on the xenon lamp 1 as a light source, position the filter wheel 5 to the initial position, and send the filter position information to the image workstation 8;

[0091] The light passing through the filter enters the lens 6, is collected by the camera 7 and outputs the RGB image data to the image workstation 8, which receives and saves the image data and associates it with the position information;

[0092] After completing the first filter image acquisition, the image workstation 8 communicates with the filter wheel controller 3 via serial communication, controls the filter wheel 5 to move to the next filter position, and repeats the acquisition process until the image data of all filters are acquired;

[0093] After the data collection is completed, the image workstation 8 processes the image data collected by each filter in turn, and reconstructs the collected RGB values ​​into corresponding spectral information through the mapping relationship model corresponding to RGB and spectral wavelength, compares the obtained spectral information with the spectral information in the database, screens out filters that meet the requirements, performs statistical analysis on the types and proportions of the RGB values ​​of the filters that meet the requirements, fits the statistical analysis results with the spectral characteristic curve of the filter, and thus confirms the model of the filter;

[0094] After all filter models are confirmed, compare the test results with the installation information and print the test results.

[0095] In this embodiment, a mapping relationship model between RGB and spectral wavelength is established, including:

[0096] like Figure 6 As shown in the figure, the CIE1931-XYZ system is based on the RGB system, and three ideal primary colors are selected to replace the actual three primary colors, thereby converting the spectral three stimulus values ​​in the CIE-RGB system into and the chromaticity coordinates r, g, and b all become positive;

[0097] To find the color tristimulus values ​​X, Y, Z of a known wavelength, the sum Σ is used instead of the integral ∫ in actual calculation, which is expressed as the following formula:

[0098]

[0099] Where K is the adjustment factor, S(λ) is the relative spectral power distribution of a light source to be measured are the spectral tristimulus values.

[0100] In this embodiment, the 1931 CIE-XYZ spectral chromaticity coordinates x, y, z are calculated and expressed as the following formula:

[0101]

[0102] Among them, the relationship between r, g, b values ​​and the color tristimulus values ​​(X, Y, Z) defined by the 1931 CIE standard colorimetry system is generally nonlinear, so it is necessary to use certain nonlinear conversion calculations to establish a model for conversion. This embodiment uses an artificial neural network to establish a mapping relationship model between r, g, b values ​​and the color tristimulus values ​​(X, Y, Z) defined by the CIE standard colorimetry system.

[0103] In this embodiment, when an artificial neural network is used to establish a mapping relationship model, it specifically includes:

[0104] Collect r, g, b values ​​and corresponding X, Y, Z three stimulus value data as training samples;

[0105] Design an artificial neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer accepts r, g, b values ​​as input, and the output layer outputs three stimulus values ​​of X, Y, and Z;

[0106] The artificial neural network is trained using training samples, and the network output is made close to the real X, Y, and Z three stimulus values ​​by adjusting the weights and bias parameters in the network;

[0107] The trained artificial neural network is tested using validation samples to evaluate its performance; if the test results meet the requirements, the network is used in actual color conversion applications.

[0108] In this embodiment, when collecting the spectral characteristic curves of different filters, it includes:

[0109] The RGB values ​​of all filters are collected 10 times for each filter, and the spectral characteristic curve of the filter is fitted based on the data. Thus, each filter has a specific spectral characteristic curve;

[0110] The existing filter to be tested is mounted on the filter wheel 5, and irradiated with a xenon lamp 1, the spectral wavelength range of the xenon lamp 1 is 400nm-1000nm, including a continuous visible light band and a near infrared band;

[0111] After the light of the xenon lamp 1 hits the filter, since the filter is a bandpass filter, only light of a specific wavelength can pass through the filter, and the light of the specific wavelength passes through the lens 6 and is then collected by the camera 7;

[0112] The camera 7 transmits the collected image data to the image workstation 8, wherein the camera 7 has two channels, one of which is an RGB channel for collecting spectral information of visible light, and the other is an IR channel for collecting spectral information of the near-infrared region.

[0113] The image workstation 8 counts the RGB values ​​and the proportion of each RGB value in each image data, and reconstructs all spectral bands and the proportion of each band according to the RGB values;

[0114] The 10 sets of image data collected are recorded, and the spectral characteristic curve of the light passing through the filter is fitted according to the 10 sets of data and stored in the database.

[0115] In this embodiment, after constructing the mapping relationship model between RGB values ​​and spectral information, for new RGB value input, the trained model is used to perform spectral reconstruction, and complete spectral information is output, including the spectral reflectance or transmittance of each band. According to the reconstructed spectral information, the proportion of each band is calculated, which can be achieved by dividing the spectral reflectance or transmittance of each band by the sum of the spectral reflectance or transmittance of all bands.

[0116] In this embodiment, the obtained spectral information is compared with the spectral information in the database to screen out filters that meet the requirements, including:

[0117] The obtained spectral information and the spectral information in the database are standardized to ensure that the two are compared on the same scale. The standardization process includes data normalization, denoising, smoothing and other processing steps to improve the accuracy and stability of the comparison.

[0118] Use a registration algorithm or other alignment method to align the data so that the peaks and valleys of different spectral data can be aligned. Since the spectral data may be offset or different, it is necessary to align the data to ensure the accuracy of the comparison.

[0119] The correlation coefficient between the calculated spectral information and the spectral information in the database is set to a threshold. When the correlation coefficient is greater than the threshold, the filter is considered to meet the requirements. The closer the correlation coefficient is to 1, the higher the similarity between the two spectral information.

[0120] In this embodiment, regarding database update: the spectral information in the database is updated regularly to ensure the accuracy and reliability of the comparison. When spectral information is added or deleted, standardization processing and data alignment should be re-performed.

[0121] In this embodiment, the statistical analysis of the types and proportions of the RGB values ​​of the filters that meet the requirements includes:

[0122] For each selected filter, extract its corresponding RGB value, and perform deduplication processing on the extracted RGB value;

[0123] Classify RGB values ​​into different color ranges or categories (such as red, green, blue, etc.) as needed;

[0124] Count the number of RGB values ​​in each color range or category. For each color range or category, calculate the proportion of its RGB value in the total data set, where the proportion is calculated by dividing the number of RGB values ​​in a color range or category by the total number of RGB values.

[0125] Use charts to visualize the types of RGB values ​​and their proportions.

[0126] In this embodiment, when the statistical analysis results are fitted with the spectral characteristic curve of the filter, the fitting effect is evaluated by calculating the goodness of fit (such as R2 value, mean square error MSE, etc.). Among them, according to the characteristics of the data and the needs of fitting, a suitable fitting method is selected. Commonly used fitting methods include least squares method, nonlinear regression, etc.

[0127] The fitted spectral characteristic curve is compared with the spectral characteristic curve of the known filter model, which includes whether the reflectivity / transmittance at the key wavelength is consistent and the similarity of the overall spectral shape.

[0128] In addition to the spectral characteristics, other factors such as the filter size, material, and transmittance range also need to be considered.

[0129] Specifically, the imaging band of the small animal live imaging device is 400nm-900nm, of which the visible light band is 400nm-780nm and the near infrared band is 780nm-900nm. The filters used in the device are between 400nm-900nm. First, the RGB values ​​of all filters need to be collected, and each filter is collected 10 times. Based on this data, the spectral characteristic curve of the filter is fitted. Therefore, each filter has a specific spectral characteristic curve. Figure 1 and 2 As shown, the xenon lamp 1, the filter wheel 5, the lens 6 and the camera 7 need to be placed in a dark box 9 to avoid interference from external light. The existing filter to be detected is installed on the filter wheel 5, and the xenon lamp 1 is used for irradiation. The spectral wavelength range of the xenon lamp 1 is 400nm-1000nm, including a continuous visible light band and a near-infrared zone 1 band. After the light of the xenon lamp 1 irradiates the filter, since the filter is a bandpass filter, only light of a specific wavelength can pass through the filter. The light of a specific wavelength passes through the lens 6 and is then collected by the camera 7. The camera 7 has two channels, one of which is an RGB channel, which can collect spectral information of visible light, and the other is an IR channel, which can collect spectral information of the near-infrared zone 1. The camera 7 transmits the collected image data to the image workstation 8. The image workstation 8 counts the RGB values ​​and the proportion of each RGB value in each image data, and reconstructs all spectral bands and the proportion of each band based on the RGB values. The 10 sets of image data collected are recorded, and the spectral characteristic curve of the light passing through the filter is fitted based on the 10 sets of data and stored in the database.

[0130] like Figure 3 As shown, Figure 3The filter wheel 5 used in the small animal in vivo imaging device has 12 positions and an aperture of 50±2 mm. It can be installed with 12 filters. Each filter installation position has a number from 0 to 11. The filters are installed at the positions with designated numbers according to production requirements, and the installation information is input into the image workstation 8.

[0131] During the inspection, the filter wheel 5 is mounted on the filter wheel bracket 2. The filter wheel 5 is controlled by the filter wheel controller 3 to move. First, it is initialized to the 0 position. The xenon lamp 1 is used to illuminate the filter. After the light wave passes through the filter, it reaches the lens 6 and the camera 7. The image workstation 8 controls the camera 7 to collect the image. The camera 7 then transmits the data to the image workstation 8. The image workstation 8 saves the received image data and controls the filter wheel 5 to move to the next filter installation position at the same time. The above acquisition is repeated until all filters are collected. After all filter images are collected, the image workstation 8 analyzes the image data obtained from each filter. First, the spectral information corresponding to the RGB value is reconstructed, and the obtained spectral information is compared with the spectral information in the database to obtain the filter that meets the requirements; then, the types and proportions of the RGB values ​​are statistically analyzed, and fitted with the spectral characteristic curve of the filter that meets the requirements, so as to confirm the model of the filter. After all filter models are confirmed, they are compared with the installation information. If the data matches, the message "Installation is correct!" is output; if the data does not match, the message "Installation is wrong!" is output, and the information that the wrong filter is installed is printed.

[0132] In summary, the system of the present invention is composed of a xenon lamp 1, a filter wheel bracket 2, a lens 6, a camera 7 and a computer, and uses RGB values ​​to measure the visible light band of the filter, matches the existing filter data model, thereby confirming the filter model, and then comparing it with the installation information to confirm whether the installation is correct. The method provided by the present invention can improve the efficiency of filter detection used in small animal live imaging equipment, can quickly detect whether the installed filter is correctly installed, and meet the requirements of batch detection; reduce the cost of detection equipment, and improve the automation of detection equipment.

[0133] Furthermore, the present invention utilizes the mapping relationship model between RGB and spectral wavelength to reconstruct the collected RGB values ​​into corresponding spectral information, thereby improving the accuracy of spectral information reconstruction. By comparing with the spectral information in the database, filters that meet the requirements can be accurately screened out.

[0134] Furthermore, the present invention can detect all filters installed on the filter wheel 5 to ensure that none are missed, and detailed image data collection and processing are performed on each filter, thereby ensuring the comprehensiveness of the detection.

[0135] Furthermore, the automated detection of the present invention reduces labor costs and time costs, improves detection efficiency, and accurate detection results help reduce experimental failures or equipment failures caused by poor filter performance, thereby reducing overall costs.

[0136] Furthermore, the present invention compares the test results with the installation information and prints the test results, making the test results more intuitive and easy to understand, which is convenient for users to quickly understand the performance and model of the filter and provides convenience for subsequent use and maintenance.

[0137] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0138] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A filter rapid detection system, characterized in that: include: A xenon lamp as a light source is used to provide illumination for the detection system; A filter wheel bracket is used to support and fix the filter wheel to ensure that the filter wheel can rotate smoothly and select a suitable filter; A filter wheel controller, for receiving instructions and controlling the rotation of the filter wheel to select a correct filter; A motor, used for providing power to the filter wheel so that it can rotate; A filter wheel, on which a plurality of filters are mounted, for selectively transmitting or blocking light of a specific wavelength, so as to realize automatic switching of the filters; The lens is used for focusing and imaging, projecting the light through the filter onto the camera; A camera, which captures light passing through lenses and filters and converts it into a digital image; Image workstation, used to process image data transmitted from the camera and perform image analysis and processing operations; The dark box is used to accommodate the xenon lamp, filter wheel bracket, filter wheel, filter wheel controller, motor, filter wheel, lens and camera, and provides a light-free environment to avoid interference of external light on the detection system.

2. A method for rapid detection of optical filters, characterized in that: The method uses a filter rapid detection system as claimed in claim 1 to perform detection, comprising the following steps: The image workstation starts and is ready to receive data. The camera is initialized and switched to RGB mode, ready to capture images. Turn on the xenon lamp as the light source, position the filter wheel to the initial position, and send the filter position information to the imaging workstation; The light passing through the filter enters the lens, and the camera collects and outputs RGB image data to the image workstation, which receives and saves the image data and associates it with the location information. After completing the first filter image acquisition, the image workstation communicates with the filter wheel controller through serial communication, controls the filter wheel to move to the next filter position, and repeats the acquisition process until the image data of all filters are acquired; After the data collection is completed, the image workstation processes the image data collected by each filter in turn, and reconstructs the collected RGB values ​​into corresponding spectral information through the mapping relationship model between RGB and spectral wavelength, compares the obtained spectral information with the spectral information in the database, screens out filters that meet the requirements, performs statistical analysis on the types and proportions of the RGB values ​​of the filters that meet the requirements, and fits the statistical analysis results with the spectral characteristic curve of the filter to confirm the model of the filter; After all filter models are confirmed, compare the test results with the installation information and print the test results.

3. The method according to claim 2, characterized in that Establish a mapping relationship model between RGB and spectral wavelength, including: On the basis of the RGB system, three ideal primary colors are selected to replace the actual three primary colors, thereby converting the spectral tristimulus values ​​in the CIE-RGB system into and the chromaticity coordinates r, g, and b all become positive; To find the color tristimulus values ​​X, Y, Z of a known wavelength, the sum Σ is used instead of the integral ∫ in actual calculation, which is expressed as the following formula: Where K is the adjustment factor, S(λ) is the relative spectral power distribution of a light source to be measured are the spectral tristimulus values.

4. The method according to claim 3, characterized in that: Calculate the CIE-XYZ spectral chromaticity coordinates x, y, z, expressed as the following formula: Among them, an artificial neural network is used to establish a mapping relationship model between r, g, b values ​​and the color tristimulus values ​​(X, Y, Z) defined by the CIE standard chromaticity system.

5. The method according to claim 4, characterized in that: When an artificial neural network is used to establish a mapping relationship model, it specifically includes: Collect r, g, b values ​​and corresponding X, Y, Z three stimulus value data as training samples; Design an artificial neural network structure, including an input layer, a hidden layer and an output layer, wherein the input layer accepts r, g, b values ​​as input, and the output layer outputs three stimulus values ​​of X, Y, and Z; The artificial neural network is trained using training samples, and the network output is made close to the real X, Y, and Z three stimulus values ​​by adjusting the weights and bias parameters in the network; The trained artificial neural network is tested using validation samples to evaluate its performance; if the test results meet the requirements, the network is used in actual color conversion applications.

6. The method according to claim 2, characterized in that: When collecting spectral characteristic curves of different filters, including: The RGB values ​​of all filters are collected 10 times for each filter, and the spectral characteristic curve of the filter is fitted based on the data. Thus, each filter has a specific spectral characteristic curve; Install the existing filter to be tested on the filter wheel and irradiate it with a xenon lamp. The spectral wavelength range of the xenon lamp is 400nm-1000nm, including the continuous visible light band and the near-infrared band. After the light from the xenon lamp hits the filter, since the filter is a bandpass filter, only light of a specific wavelength can pass through the filter, and the light of the specific wavelength passes through the lens and is then collected by the camera; The camera transmits the collected image data to the image workstation, where the camera has two channels, one of which is the RGB channel for collecting spectral information of visible light, and the other is the IR channel for collecting spectral information in the near-infrared region.

7. The method according to claim 6, characterized in that: The image workstation counts the RGB values ​​and the proportion of each RGB value in each image data, and reconstructs all spectral bands and the proportion of each band based on the RGB values; The 10 sets of image data collected are recorded, and the spectral characteristic curve of the light passing through the filter is fitted according to the 10 sets of data and stored in the database.

8. The method according to claim 2, characterized in that: The obtained spectral information is compared with the spectral information in the database to select the filter that meets the requirements, including: The obtained spectral information and the spectral information in the database are standardized to ensure that the two are compared on the same scale; Use a registration algorithm or other alignment method to achieve data alignment so that the peaks and valleys of different spectral data can be aligned; The correlation coefficient between the calculated spectral information and the spectral information in the database is set to a threshold. When the correlation coefficient is greater than the threshold, the filter is considered to meet the requirements. The closer the correlation coefficient is to 1, the higher the similarity between the two spectral information.

9. The method according to claim 2, characterized in that: The statistical analysis of the types and proportions of the RGB values ​​of the filters that meet the requirements includes: For each selected filter, extract its corresponding RGB value, and perform deduplication processing on the extracted RGB value; Classify RGB values ​​into different color ranges or categories as needed; Count the number of RGB values ​​in each color range or category. For each color range or category, calculate the proportion of its RGB value in the total data set, where the proportion is calculated by dividing the number of RGB values ​​in a color range or category by the total number of RGB values. Use charts to visualize the types of RGB values ​​and their proportions.

10. The method according to claim 2, characterized in that: When fitting the statistical analysis results with the spectral characteristic curve of the filter, the fitting effect is evaluated by calculating the goodness of fit; The fitted spectral characteristic curve is compared with the spectral characteristic curve of the known filter model, which includes whether the reflectivity / transmittance at the key wavelength is consistent and the similarity of the overall spectral shape.

Images