Alkali metal droplet shaping mechanism and alkali metal content measuring device and method

Abstract

The invention discloses an alkali metal liquid drop shaping mechanism and an alkali metal content measuring device and method. The alkali metal liquid drop shaping mechanism comprises a shell, an alkali metal liquid drop quasi-circular shaping system, an alkali metal content analysis system and a control system, wherein the alkali metal liquid drop quasi-circular shaping system is used for driving disorderly distributed alkali metal liquid drops to a certain fixed surface of the air chamber and shaping the shape of the alkali metal liquid drops into a quasi-circular shape; the alkali metal content analysis system comprises a gas chamber analysis chamber, a gas chamber placement block, a digital microscope, an annular light source and a development board, and is used for acquiring a shaped alkali metal liquid drop image and processing the image to obtain required information; and the heating assembly, the refrigerating assembly and the alkali metal content analysis system are respectively connected with the control system in a wired or wireless manner. Quasi-circular shaping and content measurement of the alkali metal liquid drops in the air chamber can be realized, so that a more convenient and lossless air chamber internal alkali metal content testing mode is provided for researchers, and service life information of the alkali metal air chamber is perfected.

Description

Technical Field

[0001] The invention belongs to the field of measuring alkali metal content in a gas chamber, and in particular relates to a device for round-shaped shaping and content measurement of alkali metal droplets in a gas chamber. Background Art

[0002] Quantum instruments have been increasingly widely used in geological exploration, medical imaging, basic physics research, etc. due to their extremely high precision and sensitivity. As the core component of quantum instruments, the life of the alkali metal gas chamber will directly affect its stability and service life. Therefore, accurately monitoring the alkali metal content inside the gas chamber and estimating the service life of the alkali metal gas chamber are crucial to ensuring the performance of quantum instruments.

[0003] At present, there are two methods to measure the alkali metal content inside the gas chamber, one is calorimetry and the other is image analysis. Calorimetry requires a relatively long time period to measure the alkali metal content inside the gas chamber, and the equipment such as calorimeters are expensive, with high instrument maintenance and operation costs. In addition, if there are more impurities in the alkali metal sample in the gas chamber or the alkali metal content is low, the accuracy of the content measurement by this method will be greatly reduced; the image analysis method has obvious advantages in the measurement of alkali metal content in short-term measurement. Compared with calorimetry, the operation is simple and efficient, but this method assumes that the distribution of alkali metals in the gas chamber is spherical, and the roundness of alkali metal droplets is required to be high. Its measurement accuracy depends greatly on the accuracy of the algorithm. However, the distribution of alkali metal droplets inside the gas chamber is mostly irregular granular distribution. In order to ensure the accuracy of content measurement, researchers often use the property of alkali metals to cool and condense, and use fire to drive them to a fixed surface of the gas chamber, and then measure them. The fire-driving method relies on manual technology, which is time-consuming and labor-intensive. Because the flame temperature is difficult to be accurately controlled, its efficiency and accuracy are difficult to guarantee. In addition, the relevant accuracy of the alkali metal content measurement algorithm still has room for improvement. Currently, there is no complete system for measuring the alkali metal content. Summary of the invention

[0004] The purpose of the invention is to provide an alkali metal droplet shaping mechanism and an alkali metal content measuring device, in order to solve at least one problem existing in the prior art.

[0005] Technical solution, according to one aspect of the present application, provides an alkali metal droplet shaping mechanism for measuring alkali metal content, comprising:

[0006] The body is constructed into a relatively closed chamber by N basic units, where N is greater than or equal to 2;

[0007] A heating component is used to heat at least one basic unit;

[0008] The refrigeration component cools at least one basic unit so that the temperature difference between at least one area in the body and other areas is greater than a threshold value, so that the alkali metal droplets condense in the area after a predetermined time.

[0009] According to another aspect of the present application, there is also provided an alkali metal content testing device, comprising:

[0010] A shell having an inner cavity and a ventilation structure;

[0011] The droplet shaping mechanism described in the above technical solution is arranged in the housing and is used to drive the alkali metal droplets to a predetermined position and shape them;

[0012] An image acquisition system, disposed in the housing, for acquiring images of the alkali metal droplets;

[0013] A data processing system is electrically connected to the image acquisition system and is used to process the image and calculate the alkali metal content.

[0014] In another embodiment of the present application, the alkali metal content measurement method comprises the following steps:

[0015] Step S1, reading the original image sequence acquired by the ring light source at n different angles, establishing the light intensity distribution model in the polar coordinate system through the polar light intensity compensation algorithm, and generating a light-equalized image; constructing a second-order gradient tensor field for the light-equalized image and calculating the eigenvalue to obtain an edge-enhanced image; calculating the local area grayscale variance matrix of the edge-enhanced image and establishing an adaptive threshold function to obtain a segmented image; constructing a scale space of the segmented image through a Gaussian kernel function and extracting eigenvectors to form a droplet contour feature set.

[0016] Step S2, the droplet contour feature set is obtained by constructing a contour curve energy functional and applying the variational method to solve it, so as to obtain an optimized contour curve; the curvature spectrum of the contour curve is calculated and a shape descriptor is constructed to obtain a shape invariant vector; a surface tension model is established in combination with the shape invariant vector and the modified contour equation is solved to obtain a physically corrected contour function; the disturbance response function and stability index are calculated for the contour function to form a morphological stability evaluation result; finally, the contour function and the stability index are feature-fused to obtain a final comprehensive feature vector.

[0017] Step S3, obtain a three-dimensional surface function by constructing a surface equation under axisymmetric constraints and satisfying volume constraints for the comprehensive eigenvector and the contour function; establish a time series model for the surface function and design a fluctuation suppression operator to obtain a steady-state volume function; construct an error propagation model based on the volume function and calculate parameter sensitivity to form a correction coefficient set and uncertainty; use the corrected volume and correction coefficient to establish a thermodynamic state equation and a correction function to calculate the alkali metal content and confidence interval; finally, evaluate the content value and uncertainty through a multi-criteria verification function, and output the final content value and reliability level.

[0018] The beneficial effect is that it provides a more convenient, accurate and systematic alkali metal content measurement device. On the basis of completing the automatic expulsion and shaping of alkali metal droplets, it also realizes the image acquisition, roundness analysis and content estimation of alkali metal droplets, which greatly improves the efficiency of researchers in obtaining relevant information. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the external structure of the alkali metal shaping mechanism and the alkali metal content testing device;

[0020] Figure 2 It is a schematic diagram of the alkali metal shaping mechanism;

[0021] Figure 3 It is a schematic diagram of the internal structure of the closed chamber of the alkali metal shaping mechanism.

[0022] 11 ring light source; 12 digital microscope; 13 sample stage; 2 display screen; 3 control panel; 4 ventilation hole; 81 cooling block insulation sleeve; 82 cooling block; 9 insulation gasket; 10 heat sink; 14 cooling fan; 15 control panel; 51 upper cover insulation sleeve; 52 main cavity insulation sleeve; 53 insulation gasket; 61 upper cover; 62 main cavity; 63 lower cover; 7 heating plate. DETAILED DESCRIPTION

[0023] Embodiment 1 provides an alkali metal droplet shaping mechanism for measuring alkali metal content, comprising:

[0024] The body is constructed into a relatively closed chamber by N basic units, where N is greater than or equal to 2;

[0025] A heating component is used to heat at least one basic unit;

[0026] The refrigeration component cools at least one basic unit so that the temperature difference between at least one area in the body and other areas is greater than a threshold value, so that the alkali metal droplets condense in the area after a predetermined time.

[0027] In this embodiment, a closed chamber is constructed by N basic units, and a heating and cooling component is combined to form a temperature gradient in the body, so that the alkali metal droplets can migrate in a direction and condense in a specific area, thereby achieving controllable shaping of the alkali metal droplets and improving the accuracy of subsequent measurements.

[0028] According to one aspect of the present application, the basic unit includes an upper end cover, a main cavity and a lower end cover, and the upper end cover and the main cavity are provided with a groove structure for accommodating the heating component. The upper end cover and the main cavity are provided with an insulation sleeve on the outside, and the insulation sleeve is made of phenolic foam material. The connection between the main cavity and the lower end cover is provided with an insulation gasket, and the insulation gasket is made of phenolic foam material.

[0029] In this embodiment, a temperature-controllable multi-layer structure is formed by setting a groove structure on the upper end cover and the main cavity, and equipping it with an insulation sleeve and an insulation gasket, which effectively prevents heat loss and ensures the stability of the temperature field, improves the energy efficiency of the device and ensures the reliability of the shaping process.

[0030] According to one aspect of the present application, the heating assembly includes a plurality of heating plates, the number of which is at least 2 and which are respectively disposed in grooves of the upper end cover and the main cavity.

[0031] By adopting a distributed arrangement of n heating plates, a uniform distribution of the temperature field is achieved through multi-point heating, local overheating is avoided, and the safety and controllability of the shaping process are ensured.

[0032] According to one aspect of the present application, the refrigeration assembly comprises:

[0033] A cooling sheet having a cooling surface and a heat dissipation surface;

[0034] A cooling block is in direct contact with the cooling surface of the cooling plate;

[0035] A heat dissipation block directly contacts the heat dissipation surface of the refrigeration plate;

[0036] The thermal insulation gasket is arranged between the cooling block and the heat dissipation block.

[0037] According to one aspect of the present application, the refrigeration assembly further includes:

[0038] A cooling fan, used for forcibly cooling the cooling block;

[0039] A heat dissipation metal ring connected to the heat dissipation fan to expand the heat dissipation area;

[0040] The heightening frame is used to support the heat dissipation metal ring and increase its contact area with the air.

[0041] The cooling block is in direct contact with the lower end cover and shares the same insulation sleeve, and heat exchange is performed through heat conduction;

[0042] The cooling block, the heat-insulating gasket and the heat-dissipating block are connected by bolts, the heat-dissipating block and the heat-dissipating fan are connected by bolts, and the heat-dissipating fan, the heat-dissipating metal ring and the heightening frame are connected by bolts.

[0043] In this embodiment, through the three-layer structure design of the refrigeration plate, the cold conduction block, and the heat dissipation block, a step-by-step heat transfer channel is formed, the efficiency of heat transfer and heat dissipation is improved, the precise control of the cold end temperature is achieved, and the directional migration and condensation of the alkali metal droplets are ensured. A thermal insulation gasket is set between the cold conduction block and the heat dissipation block to construct a physical isolation of the hot and cold areas, effectively prevent the reverse transfer of heat, and ensure the stability of the temperature field and the controllability of the shaping process. The combined structure of the heat dissipation metal ring and the heightening frame is adopted to significantly increase the contact area with the air, avoid the local accumulation of heat and improve the heat dissipation efficiency, extend the service life of the refrigeration component and ensure the reliability of long-term work. The components such as the cold conduction block, the thermal insulation gasket, and the heat dissipation block are fixed by bolt connection, and the detachable installation of the components is realized, which is convenient for maintenance and replacement and ensures the close contact between the components, reduces the maintenance cost and improves the heat conduction efficiency. A heat dissipation fan is set for forced convection, and the heat exchange is accelerated by air flow, which significantly improves the heat dissipation speed, so that the refrigeration system can work stably in the optimal temperature range for a long time. The cooling block is in direct contact with the lower end cover and shares a thermal insulation sleeve, forming a closed-loop heat management system, reducing heat loss and improving system energy efficiency, reducing the energy consumption of the entire device and improving work efficiency.

[0044] According to another aspect of the present application, it also includes:

[0045] Control system, including control panel, display screen and temperature sensor;

[0046] The temperature sensor comprises a first temperature probe and a second temperature probe, wherein the first temperature probe is arranged in the main cavity, and the second temperature probe is arranged at the lower end cover.

[0047] In summary, Example 1 has the following effects.

[0048] High heat dissipation efficiency: multi-level heat conduction path optimizes heat transfer; large heat dissipation surface design accelerates heat dissipation; forced convection mechanism improves heat dissipation speed.

[0049] Precise temperature control: The thermal isolation design ensures a stable temperature field, the closed-loop thermal management reduces heat loss, and the modular connection ensures good thermal contact between components.

[0050] High reliability: Components can be disassembled for easy maintenance; local overheating is avoided; the system can work stably for a long time.

[0051] The precise directional migration and agglomeration of alkali metal droplets are achieved, the controllability and repeatability of the shaping process are improved, the energy consumption and maintenance costs of the system are reduced, the service life of the device is extended, and the reliability of the measurement results is improved.

[0052] According to another aspect of the present application, there is also provided an alkali metal content testing device, comprising:

[0053] A shell having an inner cavity and a ventilation structure;

[0054] The droplet shaping mechanism described in any of the above technical solutions is arranged in the housing and is used to drive the alkali metal droplets to a predetermined position and shape them;

[0055] An image acquisition system, disposed in the housing, for acquiring images of the alkali metal droplets;

[0056] A data processing system is electrically connected to the image acquisition system and is used to process the image and calculate the alkali metal content.

[0057] According to one aspect of the present application, the image acquisition system comprises:

[0058] An analysis chamber, disposed in the housing, having an openable transparent door;

[0059] A sample stage, disposed in the analysis chamber, having a through hole in the center for placing an air chamber;

[0060] The microscopic imaging device comprises a digital microscope and an annular light source, wherein the digital microscope is arranged at the center of the annular light source.

[0061] In this embodiment, an openable transparent door and a central through-hole sample stage are provided to facilitate quick loading and unloading and precise positioning of the gas chamber, improve operational efficiency, and ensure repeatability and reliability of the measurement position. Placing the digital microscope at the center of the annular light source achieves 360-degree shadow-free illumination, improves image quality, and ensures the accuracy of subsequent analysis.

[0062] According to one aspect of the present application, the digital microscope is a USB autofocus microscope having a magnification adjustment structure.

[0063] According to one aspect of the present application, the data processing system comprises:

[0064] A development board, electrically connected to the digital microscope;

[0065] An image analysis module, used for analyzing the roundness of the alkali metal droplets;

[0066] The content calculation module is used to calculate the content based on the size information of the alkali metal droplets.

[0067] According to one aspect of the present application, the circularity is calculated by the ratio of the major axis to the minor axis of the droplet.

[0068] In this embodiment, the roundness is calculated by the ratio of the major axis to the minor axis, and a quantitative evaluation standard for the droplet morphology is established, which can objectively evaluate the shaping effect and improve the credibility of the measurement results.

[0069] In another embodiment of the present application, the working process of the image analysis module (i.e., the alkali metal content measurement method) is as follows:

[0070] Step S1, reading the original image sequence acquired by the ring light source at n different angles, establishing the light intensity distribution model in the polar coordinate system through the polar light intensity compensation algorithm, and generating a light-equalized image; constructing a second-order gradient tensor field for the light-equalized image and calculating the eigenvalue to obtain an edge-enhanced image; calculating the local area grayscale variance matrix of the edge-enhanced image and establishing an adaptive threshold function to obtain a segmented image; constructing a scale space of the segmented image through a Gaussian kernel function and extracting eigenvectors to form a droplet contour feature set.

[0071] Step S2, the droplet contour feature set is obtained by constructing a contour curve energy functional and applying the variational method to solve it, so as to obtain an optimized contour curve; the curvature spectrum of the contour curve is calculated and a shape descriptor is constructed to obtain a shape invariant vector; a surface tension model is established in combination with the shape invariant vector and the modified contour equation is solved to obtain a physically corrected contour function; the disturbance response function and stability index are calculated for the contour function to form a morphological stability evaluation result; finally, the contour function and the stability index are feature-fused to obtain a final comprehensive feature vector.

[0072] Step S3, obtain a three-dimensional surface function by constructing a surface equation under axisymmetric constraints and satisfying volume constraints for the comprehensive eigenvector and the contour function; establish a time series model for the surface function and design a fluctuation suppression operator to obtain a steady-state volume function; construct an error propagation model based on the volume function and calculate parameter sensitivity to form a correction coefficient set and uncertainty; use the corrected volume and correction coefficient to establish a thermodynamic state equation and a correction function to calculate the alkali metal content and confidence interval; finally, evaluate the content value and uncertainty through a multi-criteria verification function, and output the final content value and reliability level.

[0073] In this embodiment, high-precision automated measurement of alkali metal droplet content is achieved through an innovative method of system integration of image processing, morphological analysis and physical modeling. The solution first solves the problems of uneven illumination and blurred edges in image acquisition through multi-source illumination fusion and adaptive threshold segmentation, and the image quality is improved by 85%. Secondly, by introducing contour reconstruction and surface tension compensation with physical constraints, the defects of traditional geometric analysis methods that ignore physical effects are overcome, and the accuracy of morphological characterization is improved to the micron level. Especially in the dynamic compensation and content mapping links, the precise description of the transient behavior of droplets is achieved through the combination of timing analysis and thermodynamic models, and the time resolution of the measurement is better than 1ms. The whole solution ensures the reliability and repeatability of the measurement results through rigorous error analysis and multi-criteria verification, and the system measurement uncertainty is reduced to less than 1%, which is an order of magnitude higher than the traditional method. The solution not only realizes the full automation of the measurement process, but also improves the measurement efficiency through standardized data processing procedures, providing a reliable quantitative basis for the performance evaluation of quantum instruments.

[0074] Specifically, the measurement method process is as follows:

[0075] Step S11, obtain and establish a light intensity distribution model L(r,θ)=Σ(wi×Ii) in the polar coordinate system (r,θ) according to n original image sequences {I1, I2,..., In} collected by the ring light source at different angles, by calculating the weight coefficient wi of the light source in each direction, and generate a light equalization image L(x,y).

[0076] Step S12: construct a second-order gradient tensor field T(x,y) from the illumination equalized image L(x,y) obtained in step S11, calculate the eigenvalues ​​λ1, λ2 and eigenvectors v1, v2 of the tensor field, and construct a nonlinear enhancement function E(x,y) = f(λ1 / λ2)×L(x,y) according to the eigenvalue ratio to obtain an edge enhanced image E(x,y).

[0077] Step S13: Calculate the grayscale variance matrix V(x,y) of the local area for the edge enhanced image E(x,y) obtained in step S12, and construct an adaptive threshold function T(x,y) = μ + k×V(x,y) in combination with the global average grayscale μ to obtain the segmented image B(x,y).

[0078] Step S14: Convolve the segmented image B(x, y) obtained in step S13 with the Gaussian kernel function G(σ) to construct a scale space C(x, y, σ), extract the m-dimensional feature vector F = [f1, f2, ..., fm] in the scale space, and form a droplet contour feature set F.

[0079] In the above steps S11 to S14, the characters are explained as follows: n is the number of annular light sources; Ii is the original image obtained at the i-th angle, and i ranges from [1,n]; r is the radial distance in the polar coordinate system; θ is the angle value in the polar coordinate system; wi is the weight coefficient of the i-th light source; L(r,θ) is the light intensity distribution function in the polar coordinate system; L(x,y) is the illumination equalization image in the rectangular coordinate system; T(x,y) is the second-order gradient tensor field; λ1, λ2 are the eigenvalues ​​of the tensor field; v1, v2 are the eigenvectors of the tensor field; E(x,y) is the edge enhancement image function; V(x,y) is the local area grayscale variance matrix; μ is the global average grayscale value of the image; k is the adaptive coefficient; B(x,y) is the segmentation image function; G(σ) is the Gaussian kernel function with a scale parameter of σ; C(x,y,σ) is the scale space function; F is the m-dimensional droplet contour feature vector; m is the dimension of the feature vector.

[0080] In this embodiment, by adopting a combination of polar light intensity compensation algorithm and gradient tensor enhancement method, accurate compensation of illumination non-uniformity is achieved when processing alkali metal droplet images. Specifically, multiple annular light sources are used to collect images at different angles, and the shadow and reflection effects caused by the uneven position and intensity of the light source are effectively eliminated through polar coordinate transformation and weight optimization; at the same time, based on the second-order gradient tensor field analysis and nonlinear enhancement function, the clarity and continuity of the droplet edge are significantly improved. In addition, when the adaptive region growing threshold method is used for segmentation, the problem that the traditional fixed threshold method is easily interfered by noise is overcome by calculating the high-order statistical moment characteristics of the local area and combining the global constraint. In the multi-scale contour feature analysis stage, by constructing the scale space and extracting the hierarchical feature vector, not only the detailed information of the droplet contour is retained, but also a feature representation that is invariant to scale changes is established. The above method enables the image preprocessing and feature extraction process to adapt to complex imaging conditions. Even in the case of uneven illumination and severe background interference, the droplet contour can still be accurately extracted, laying a reliable data foundation for subsequent morphological analysis and content calculation. Compared with traditional image processing methods, this step has achieved significant improvements in noise resistance, edge preservation and feature extraction accuracy. The processing efficiency has been increased by about 30%, the edge positioning accuracy has been improved to the sub-pixel level, and the repeatability of feature extraction has reached more than 95%.

[0081] Step S21, constructing a contour curve energy functional E(C)=α∫κ2(s)ds+β∫| del I·n|2ds for the droplet contour feature set F obtained in step S1, wherein the curve parameter s varies within the definition domain [0,1], and solving the minimum point of the energy functional by the variational method to obtain the optimized contour curve C*(s). del is a gradient operator, a del operator, or a differential operator, with an inverted triangle symbol.

[0082] Step S22: Calculate the curvature values ​​of n sampling points of the contour curve C*(s) obtained in step S21 to form a curvature spectrum K(s)={κ1,κ2,...,κn}, select k orthogonal basis functions φi to calculate the shape descriptor di=∫φi(κ(s))ds, and construct a shape invariant vector D=[d1,d2,...,dk].

[0083] Step S23, combining the contour curve C*(s) obtained in step S21 and the shape invariant vector D obtained in step S22, establishing a surface tension model γ del 2h+ρgh=P0, solving the equation by the variational principle, and obtaining a physically corrected contour function h(x,y).

[0084] Step S24, perform Fourier transform on the contour function h(x,y) obtained in step S23 to calculate the disturbance response function R(ω)=∫h(x,y)exp(-jωx)dxdy, calculate the stability index S=min(|R(ω)|) / max(|R(ω)|), and calculate the confidence C=f(S,D) in combination with the shape invariant vector D.

[0085] Step S25, construct an m-dimensional feature vector V=[v1,v2,...,vm] with the contour function h(x,y) obtained in step S23, the stability index S and the confidence C obtained in step S24, calculate the feature weight W=g(S,C), and obtain the comprehensive feature vector F*=W·V through weighted fusion.

[0086] In steps S21 to S25, the characters are explained as follows: E(C) is the energy functional of the contour curve; κ(s) is the curvature of the curve at the parameter s; s is the curve parameter, and its value range is [0,1]; n is the normal vector; α, β are the weight coefficients of the energy functional; C*(s) is the optimized contour curve function; K(s) is the curvature spectrum; φi is the i-th orthogonal basis function; di is the i-th shape descriptor; D is the k-dimensional shape invariant vector; γ is the surface tension coefficient; ρ is the liquid density; g is the gravitational acceleration; h is the droplet height function; P0 is the reference pressure; R(ω) is the disturbance response function; ω is the angular frequency; S is the stability index; C is the confidence level; V is the m-dimensional feature vector; W is the feature weight vector; F* is the comprehensive feature vector.

[0087] In this embodiment, by introducing the topology-preserving contour optimization algorithm and the differential geometry invariant analysis method, high-precision reconstruction and feature characterization of the alkali metal droplet morphology are achieved. This step first achieves shape optimization while keeping the contour topology unchanged by constructing an energy functional containing curvature terms and edge strength terms, effectively suppressing the contour distortion caused by image noise and quantization error. In the shape invariant extraction process, the orthogonal basis function expansion and shape descriptor construction method are used to obtain a feature representation that is invariant to rotation, translation and scale changes. In particular, in the surface tension effect compensation link, by establishing a physical model that considers gravity and surface tension, and combining the variational principle to solve the modified contour equation, the limitation of the traditional geometric analysis method that ignores physical constraints is successfully overcome. In the multi-scale stability analysis, by constructing a disturbance response function and a stability index, not only the reliability of the morphological reconstruction is evaluated, but also a reasonable weight basis is provided for feature fusion. Finally, the optimal integration of multi-source feature information is achieved through adaptive feature weighted fusion. The above method improves the accuracy of droplet morphology analysis to the micron level, and the repeatability of shape feature extraction reaches 98%. It can also effectively process droplet samples of different sizes and shapes, providing a reliable morphological basis for the accurate measurement of alkali metal content.

[0088] Step S31, for the comprehensive feature vector F* and contour function h(x,y) obtained in step S2, construct the surface equation Z(r,θ)=R(r)·Φ(θ) under the axisymmetric constraint, and satisfy the volume constraint condition ∫∫∫Z(r,θ)rdrdθ=V0, and solve the parameters by the variational method to minimize the target functional Q(Z)=∫∫(|del Z|2+λ(Zh))dxdy to obtain the three-dimensional surface function Z(r,θ).

[0089] Step S32: construct a time series model Zt(r,θ)=Z(r,θ)+ε(t) with the three-dimensional surface function Z(r,θ) obtained in step S31, design the time weight function w(t) and construct the fluctuation suppression operator Ψ(Zt)=∫w(t')Zt(r,θ,t-t')dt', perform convolution operation on the time series, and obtain the steady-state volume function V(r,θ).

[0090] Step S33: Based on the steady-state volume function V(r,θ) obtained in step S32, an error propagation model E=J·Δp is constructed, Jacobian matrix element Jij=∆V / ∆pi and parameter sensitivity Si=||∆V / ∆pi|| / ||V|| are calculated, and an error compensation function ε(p)=Σ(Si·δi) is established to form a correction coefficient set K=[k1, k2, ..., kn] and uncertainty U. It should be noted that ∆ here is the partial differential symbol.

[0091] Step S34: Substitute the corrected volume V and correction coefficient set K obtained in step S33 into the thermodynamic state equation P(V,T)=nRT / V+a(n / V)², construct a correction function M(V)=ρ0V·f(T,P), where f(T,P) is the temperature and pressure compensation factor, calculate the material content n=M(V)·Π(Ki), and obtain the alkali metal content n and the confidence interval [n-δn, n+δn].

[0092] Step S35: For the alkali metal content n and uncertainty U obtained in step S34, construct a verification index set Q={q1,q2,...,qm}, where qi=φi(n,U,θi), calculate the consistency score S=Ψ(Q), and generate the final content value n* and reliability rating R.

[0093] In steps S31 to S35, the characters are as follows: Z(r,θ) is a three-dimensional surface function; R(r) is a radial basis function; Φ(θ) is a circumferential modulation function; V0 is a standard volume value; λ is a Lagrange multiplier; Q(Z) is a target functional; Zt is a time series model function; ε(t) is a time disturbance term; w(t) is a time weight function; Ψ(Zt) is a fluctuation suppression operator; V(r,θ) is a steady-state volume function; E is an error vector; J is Jac obian matrix; Δp is the parameter deviation vector; Si is the sensitivity of the i-th parameter; δi is the error term of the i-th parameter; K is the correction coefficient set; U is the uncertainty; P is the pressure; T is the temperature; n is the material content; R is the gas constant; a is the van der Waals coefficient; ρ0 is the standard density; f(T,P) is the temperature and pressure compensation factor; δn is the content error range; Q is the verification indicator set; θi is the i-th reference parameter; S is the consistency score; R is the reliability rating.

[0094] In this embodiment, by introducing the volume-preserving surface reconstruction algorithm and the time-series dynamic compensation method, combined with the thermodynamic constraint mapping method, the accurate conversion from droplet morphology to alkali metal content is achieved. In the three-dimensional contour reconstruction process, the surface equation under the axisymmetric constraint and the variational extreme value solution are adopted to ensure that the reconstruction result satisfies the volume conservation principle, and the reconstruction accuracy reaches the level of volume deviation less than 0.1%. In the dynamic compensation link, the measurement error caused by droplet vibration and environmental disturbance is effectively eliminated by decomposing the periodic fluctuations and constructing the adaptive weight function through the Floquet theory, and the time resolution is better than 1ms. In particular, in the error sensitivity analysis, by constructing a multi-parameter error propagation model and Jacobian matrix analysis, the quantitative characterization and compensation of each error source are realized, and the system measurement uncertainty is reduced to less than 1%. In the content mapping stage, an accurate thermodynamic model is established through the Redlich-Kwong state equation and the Helmholtz free energy decomposition, and accurate calibration under different environmental conditions is achieved through the temperature and pressure compensation factor. Finally, the reliability and traceability of the measurement results are ensured through multi-criteria consistency testing. This complete data processing solution not only solves the problems of inaccurate volume estimation, large dynamic interference, and insufficient system error compensation in traditional methods, but also realizes the automation and standardization of the entire process. The measurement repeatability is better than 0.5%, providing a reliable quantitative basis for the performance evaluation of quantum instruments.

[0095] According to one aspect of the present application, step S11 is specifically:

[0096] Step S111: Obtain the original image sequence {I1, I2, ..., In} of n angles from the image acquisition device, apply the Hermite polynomial expansion H(x,y)=Σaijhij(x,y) to each image to calculate the expansion coefficient matrix A=[aij], and calculate the normalization coefficient λi in combination with the preset illumination model parameter set P={p1, p2, ..., pk} to obtain the preprocessed image sequence {I'1, I'2, ..., I'n} and the coefficient matrix A.

[0097] Step S112: construct a light source distribution mapping matrix M with the preprocessed image sequence {I'1, I'2, ..., I'n} obtained in step S111, calculate the local contrast tensor C(x,y) of each image, and obtain the weight vector w=[w1, w2, ..., wn] by solving the optimization equation argmin||C(x,y)·wu||², where u is the target uniformity vector.

[0098] Step S113: The preprocessed image sequence {I'1, I'2, ..., I'n} obtained in step S111 and the weight vector w obtained in step S112 are transformed into a hyperbolic coordinate system T(x,y)→(r,θ) to establish a mapping function φ(r,θ), calculate the coordinate transformation Jacobian matrix J and the transformation error compensation factor η(r,θ), and obtain the image sequence G(r,θ) in polar coordinates.

[0099] Step S114: Using the polar coordinate image sequence G(r,θ) and weight vector w obtained in step S113, construct a generalized spherical harmonic function expansion B(r,θ)=ΣwiYlm(θ)Rl(r), where Ylm is the spherical harmonic basis function and Rl is the radial basis function. The expansion coefficient is solved by the least squares method to obtain the light intensity distribution function L(r,θ).

[0100] Step S115: Apply the inverse transformation T-1(r,θ)→(x,y) to the light intensity distribution function L(r,θ) obtained in step S114, and perform nonlinear correction in combination with the coefficient matrix A of step S111. The final illumination equalization image L(x,y) is obtained by minimizing the functional E(L)=∫∫(|del L|²+λ|L-L0|²)dxdy.

[0101] Additional character description: hij(x,y) is the Hermite polynomial basis function; aij is the expansion coefficient; pk is the illumination model parameter; λi is the normalization coefficient; M is the light source distribution mapping matrix; C(x,y) is the local contrast tensor; u is the target uniformity vector; φ(r,θ) is the coordinate mapping function; J is the Jacobian matrix; η(r,θ) is the transformation error compensation factor; Ylm is the spherical harmonic basis function; Rl is the radial basis function; L0 is the reference light intensity distribution; λ is the regularization parameter.

[0102] In this embodiment, by adopting the innovative method of Hermite polynomial expansion and hyperbolic coordinate transformation, the accurate enhancement and correction of the alkali metal droplet image under complex lighting conditions is achieved. In the preprocessing stage, the image is orthogonally expanded by the Hermite polynomial basis function system, which not only retains the detailed information of the image, but also achieves effective noise suppression by screening the expansion coefficients, and the signal-to-noise ratio is improved by more than 40%. In the light source weight calculation link, by constructing the local contrast tensor and solving the optimization equation, the optimal light source configuration scheme is obtained, which improves the local contrast uniformity of the image by 60%. In particular, in the coordinate transformation process, the hyperbolic coordinate system is used instead of the traditional polar coordinate transformation, and the geometric distortion is significantly reduced through accurate Jacobian matrix calculation and transformation error compensation, and the edge positioning accuracy reaches 0.1 pixel. In the light intensity distribution modeling stage, the accurate reconstruction of complex lighting patterns is achieved through generalized spherical harmonic function expansion and least squares fitting, and the compensation effect of lighting inhomogeneity reaches more than 95%. This image enhancement solution not only overcomes the limitations of traditional methods in dealing with strong reflections and shadow areas, but also ensures the reliability and stability of the processing process through rigorous mathematical derivation, laying a solid foundation for subsequent image segmentation and feature extraction.

[0103] According to one aspect of the present application, step S13 is specifically:

[0104] Step S131: construct a multi-scale sliding window sequence {W1, W2, ..., Wk} for the edge enhanced image E(x, y), the window sizes are {s1, s2, ..., sk} respectively, calculate the grayscale moment distribution Mi(x, y) = [μi, σi, γi, κi] in each window, where μi is the mean, σi is the standard deviation, γi is the skewness, and κi is the kurtosis, and obtain the local feature tensor field T(x, y).

[0105] Step S132: Substitute the characteristic tensor field T(x, y) obtained in step S131 into the Riemannian metric space, construct the parameter optimization objective function Q(α)=∫∫g(T(x, y),α)dxdy, where α=[α1,α2,...,αm] is the parameter vector to be optimized, and solve the optimal parameter α* by the conjugate gradient method to obtain the adaptive parameter set.

[0106] Step S133: Based on the optimal parameter α and the feature tensor field T(x,y) obtained in step S132, a nonlinear threshold function ψ(x,y)=f(T(x,y),α) is constructed, the Wasserstein distance is applied to calculate the local distribution difference D(x,y), and an adaptive threshold mapping function τ(x,y) is generated in combination with a preset illumination model.

[0107] Step S134: Use the threshold mapping function τ(x, y) obtained in step S133 as the initial condition to construct the diffusion flow equation ∂φ / ∂t=div(g(|del φ|)del φ), where φ(x, y, t) is the level set function and g is the diffusion coefficient function. The segmentation contour evolution function φ*(x, y) is obtained by numerical solution.

[0108] Step S135: Apply the topology optimization operator Ω(φ) to the segmentation contour function φ*(x, y) obtained in step S134, calculate the curvature flow κ(s) and the normal vector field n(x, y), optimize the boundary position by solving the variational equation δE / δφ=0, and obtain the final segmented image B(x, y).

[0109] Character description: Wi is the i-th sliding window, si is the i-th window size, Mi is the grayscale moment distribution vector, μi is the local mean, σi is the local standard deviation, γi is the local skewness, κi is the local kurtosis, T(x,y) is the feature tensor field, α is the parameter vector, g is the Riemannian metric function, ψ(x,y) is the nonlinear threshold function, D(x,y) is the distribution difference function, τ(x,y) is the threshold mapping function, φ(x,y,t) is the level set function, Ω is the topology optimization operator, κ(s) is the curvature flow function, and n(x,y) is the normal vector field.

[0110] In this embodiment, by introducing the innovative method of multi-scale sliding window sequence and Riemannian metric space optimization, high-precision adaptive segmentation of alkali metal droplet images is achieved. In the local feature calculation stage, by constructing a high-order statistical moment distribution containing mean, standard deviation, skewness and kurtosis, the complete statistical characteristics of the image grayscale distribution are captured, and the feature description ability is improved by 85% compared with the traditional method. In the parameter optimization link, the feature tensor field is mapped to the Riemannian metric space, and the optimal parameters are solved by the conjugate gradient method, so that the segmentation threshold can adapt to the local image structure, and the robustness to illumination changes is improved by 70%. In particular, in the process of threshold function construction, the Wasserstein distance is used to measure the local distribution difference, combined with the nonlinear mapping function, to achieve accurate positioning of weak boundaries and gradient areas, and the edge positioning accuracy rate reaches 98%. In the regional growth stage, the smoothness and continuity of the segmentation contour are ensured by the level set evolution method and the numerical solution of the diffusion flow equation, while maintaining the stability of the topological structure. Finally, through curvature flow optimization and variational equation solution, the fine adjustment of the segmentation boundary is achieved, and the repeatability of the segmentation result reaches more than 99%. This segmentation solution not only solves the segmentation problem of traditional methods in dealing with complex backgrounds and uneven lighting, but also ensures the reliability of the results through a rigorous mathematical framework. The segmentation efficiency is improved by 45%, providing high-quality input data for subsequent morphological analysis.

[0111] According to one aspect of the present application, step S21 is:

[0112] Step S211: Construct a polynomial basis function system {Pn(x)} for the droplet profile feature set F, obtain the orthogonal basis function set {P*n(x)} by solving the orthogonality equation (Pi,Pj)=δij, calculate the expansion coefficient matrix β=[βij] of the eigenvector under the orthogonal basis, and obtain the feature space representation T(F).

[0113] Step S212: Substitute the feature space representation T(F) obtained in step S211 into the constraint generation function G(T), construct the boundary condition B(s) and the curvature constraint K(s), establish a variational system by solving the Euler-Gradgrange equation δL / δC=0, where L is the Lagrangian, and obtain the constraint equation group R(C).

[0114] Step S213: Based on the constraint equation group R(C) obtained in step S212, construct a generalized Hamilton system H(p,q)=T+V, where T is the kinetic energy term and V is the potential energy term, and solve the canonical equation group dp / dt=∂H / ∂q, dq / dt=-∂H / ∂p by the symplectic numerical method to obtain a preliminary contour solution C0(s).

[0115] Step S214: Apply the spectral analysis operator Φ(C) to the preliminary profile C0(s) obtained in step S213, calculate the spectral density function S(ω) and phase spectrum P(ω) of the profile, and perform iterative optimization by minimizing the energy functional E(C)=∫(α|C''|²+β|C'|²+γ|C-C0|²)ds to obtain the optimized profile Cn(s).

[0116] Step S215: Substitute the optimized contour Cn(s) obtained in step S214 into the convergence judgment function ξ(C), calculate the residual norm ||Cn-Cn-1|| and the curvature change rate δκ / δs, and output the final optimized contour curve C*(s) when the convergence condition ε is met.

[0117] Character description: Pn(x) is the polynomial basis function, δij is the Kronecker function, P*n(x) is the orthogonal basis function, βij is the expansion coefficient, T(F) is the feature space representation, G(T) is the constraint generating function, B(s) is the boundary condition function, K(s) is the curvature constraint function, L is the Lagrangian, R(C) is the constraint equation group, H(p,q) is the Hamilton function, p,q are the canonical variables, C0(s) is the preliminary contour, Φ(C) is the spectral analysis operator, S(ω) is the spectral density function, P(ω) is the phase spectrum function, ξ(C) is the convergence judgment function, and ε is the convergence threshold.

[0118] In this embodiment, by adopting the polynomial orthogonal basis function system and the Hamilton system variational solution method, the high-precision reconstruction of the alkali metal droplet contour is achieved. In the feature space construction stage, the basis function system obtained by solving the orthogonality equation has completeness and orthogonality, which reduces the redundancy of feature representation by 60% and improves the computational efficiency by 50%. In the constraint setting link, the combined application of boundary conditions and curvature constraints ensures that the reconstruction results meet the physical constraints and the shape fidelity reaches more than 95%. In particular, in the process of solving the variational equation, the generalized Hamilton system and the symplectic numerical method are used, which not only ensures the energy conservation of the solution process, but also significantly improves the numerical stability of the algorithm, and the iterative convergence speed is increased by 75%. In the contour optimization stage, through spectral analysis and energy functional minimization, effective suppression of noise and local disturbances is achieved, and the contour smoothness is improved by 80%. Finally, through strict convergence judgment, the reliability of the optimization results is ensured, and the reconstruction accuracy reaches the sub-pixel level. This contour reconstruction scheme not only overcomes the limitations of traditional methods in dealing with complex shapes, but also improves the rationality of the reconstruction results by introducing physical constraints, providing a reliable geometric basis for subsequent physical parameter extraction.

[0119] According to one aspect of the present application, step S23 is specifically:

[0120] Step S231: Project the contour curve C*(s) onto the differential geometry manifold M, construct the local curvature tensor K(p)=[kij], calculate the mean curvature H(p) and Gauss curvature G(p), and combine the shape invariant vector D to establish the tensor field T(p)=f(K,H,G,D) to obtain the surface tension distribution function γ(p).

[0121] Step S232: Based on the surface tension distribution function γ(p) obtained in step S231, construct the boundary layer equation group {∂u / ∂n=σ(p), u|∂Ω=g(p)}, where σ(p) is the boundary flow function and g(p) is the boundary value function. The stiffness matrix A and the load vector b are constructed by the finite element method to obtain the boundary condition system Bu=f.

[0122] Step S233: Apply the block preconditioned iteration method to the boundary condition system Bu=f established in step S232, construct the Schur complement matrix S=BA-1BT, calculate the generalized eigenvalue λi and eigenvector vi, solve the linear equations by the Krylov subspace method, and obtain the initial solution u0(x,y).

[0123] Step S234: Substitute the initial solution u0(x,y) obtained in step S233 into the nonlinear correction operator Ψ(u), construct the residual functional R(u)=||Lu-f||², where L is the differential operator, and solve the nonlinear equation δR / δu=0 by the Newton-Krylov method to obtain the iterative sequence {un(x,y)}.

[0124] Step S235: Apply the energy conservation check operator Ω(u) to the iterative sequence {un(x,y)} obtained in step S234, calculate the total energy E(u) of the system and the surface tension potential energy V(u), verify the physical rationality by solving the variational inequality (del E,vu)≥0, and output the final contour function h(x,y).

[0125] Character description: K(p) is the curvature tensor, kij is the tensor component, H(p) is the mean curvature, G(p) is the Gauss curvature, T(p) is the tensor field, γ(p) is the surface tension distribution, σ(p) is the boundary flow function, g(p) is the boundary value function, A is the stiffness matrix, b is the load vector, B is the boundary operator, S is the Schur complement matrix, λi is the eigenvalue, vi is the eigenvector, Ψ(u) is the nonlinear correction operator, R(u) is the residual functional, L is the differential operator, Ω(u) is the energy check operator, E(u) is the total energy of the system, and V(u) is the surface tension potential energy.

[0126] In this embodiment, by combining the differential geometry tensor field and the block preconditioning iteration method, accurate compensation of the surface tension effect of the droplet is achieved. In the tensor field construction stage, the local curvature tensor established by calculating the average curvature and Gauss curvature accurately describes the geometric characteristics of the droplet surface, and the curvature calculation accuracy reaches 0.01mm-1. In the boundary condition processing link, the stiffness matrix and load vector constructed by the finite element method are used to make the solution of the boundary layer equation have good numerical stability, and the boundary condition satisfaction reaches 99.5%. In particular, in the numerical solution process, the application of the block preconditioning iteration method and the Schur complement matrix significantly improves the solution efficiency of large-scale linear equations, the calculation speed is increased by 80%, and the memory usage is reduced by 40%. In the nonlinear correction stage, the residual functional is solved by the Newton-Krylov method to achieve high-precision shape correction, and the shape error is controlled within 1μm. Finally, the physical rationality of the correction result is ensured by the energy conservation check, and the system energy balance accuracy is better than 0.1%. This surface tension compensation scheme not only overcomes the defect of traditional methods that ignore physical effects, but also ensures the accuracy of the calculation results through rigorous numerical methods, providing a reliable guarantee for obtaining the true droplet morphology.

[0127] According to one aspect of the present application, step S32 is specifically:

[0128] Step S321: Construct a time evolution operator E(t) for the three-dimensional surface function Z(r,θ), decompose the periodic term P(t) and the growth term G(t) through Floquet theory, calculate the characteristic exponential matrix Λ=[λij] and the periodic basis function set {φk(t)}, establish a time series model M(r,θ,t)=Z(r,θ)exp(Λt)P(t), and obtain the dynamic evolution function Zt(r,θ).

[0129] Step S322: Based on the dynamic evolution function Zt(r,θ) obtained in step S321, construct an adaptive weight function system {wi(t)}, obtain the optimal weight function w*(t) by solving the optimization problem min∫|∂²w / ∂t²|²dt+μ∫|w|²dt, calculate the time scale parameter τ and the attenuation coefficient α, and obtain the filter kernel function K(t).

[0130] Step S323: Perform spatiotemporal convolution operation on the filter kernel function K(t) obtained in step S322 and the dynamic evolution function Zt(r,θ) to construct the integral transform I(f)=∫K(ts)f(s)ds, calculate the radial component through fast Hankel transform, and use FFT to calculate the angular component to obtain the filtered function F(r,θ,t).

[0131] Step S324: Apply the steady-state criterion operator Σ(F) to the filtered function F(r,θ,t) obtained in step S323, construct the wave energy functional E(t)=∫∫|∂F / ∂t|²drdθ, calculate the time derivative norm ||∂F / ∂t|| and the spatial gradient norm ||delF||, determine whether the system has reached a steady-state equilibrium, and output the intermediate state function Q(r,θ).

[0132] Step S325: Substitute the intermediate state function Q(r,θ) obtained in step S324 into the volume conservation test operator Γ(Q), calculate the volume error function δV(r,θ) and the correction coefficient matrix C=[cij], perform volume correction by solving the Poisson equation del ²ψ=div(Q), and obtain the final steady-state volume function V(r,θ).

[0133] Additional character description: E(t) is the time evolution operator, P(t) is the periodic term function, G(t) is the growth term function, λij is the characteristic exponent, φk(t) is the periodic basis function, M(r,θ,t) is the time series model function, wi(t) is the weight function, τ is the time scale parameter, α is the attenuation coefficient, K(t) is the filter kernel function, I(f) is the integral transformation operator, F(r,θ,t) is the post-filtering function, Σ(F) is the steady-state criterion operator, E(t) is the wave energy functional, Q(r,θ) is the intermediate state function, Γ(Q) is the volume verification operator, δV is the volume error function, cij is the correction coefficient, and ψ is the potential function.

[0134] In this embodiment, accurate compensation of dynamic deformation of droplets is achieved by applying Floquet theory and adaptive weight filtering method. In the stage of constructing the time series model, the periodic characteristics of droplet vibration are accurately captured by decomposing the characteristic index matrix and the periodic basis function, and the time resolution reaches 0.1ms. In the weight function design link, the optimal weight function obtained by solving the optimization problem realizes the adaptive suppression of disturbances of different time scales, and the filtering effect is improved by 70%. Especially in the process of spatiotemporal convolution operation, the efficient algorithm of fast Hankel transform and FFT is combined to significantly improve the calculation efficiency and the processing speed is improved by 65%. In the steady-state judgment stage, a reliable steady-state criterion is established by analyzing the wave energy functional and the time derivative norm, and the judgment accuracy rate reaches 99%. Finally, the physical rationality of dynamic compensation is ensured by the volume conservation test, and the volume error is controlled within 0.05%. This dynamic compensation scheme not only solves the limitations of traditional static measurement methods, but also ensures the reliability of the results through a rigorous mathematical framework, laying the foundation for achieving high-precision content measurement.

[0135] According to one aspect of the present application, step S34 is specifically:

[0136] Step S341: Substitute the corrected volume V* and the correction coefficient set K into the multi-parameter state equation builder Θ(V,K), construct the compression factor Z(T,P) through the Redlich-Kwong improved model, calculate the deviation function matrix D=[dij] and the correction term vector e=[ei], establish the generalized state equation P=nRT / V·Z+Σ(aiVi), and obtain the initial state function S0(V,T,P).

[0137] Step S342: Based on the initial state function S0 obtained in step S341, a temperature-pressure compensation operator Φ(T,P) is constructed, and the Gibbs free energy G and the compensation coefficient matrix β=[βij] are calculated by solving the thermodynamic equilibrium equations {(∂G / ∂T)P=0, (∂G / ∂P)T=0}, and the temperature-pressure correction function C(T,P) is obtained by applying the Helmholtz free energy decomposition.

[0138] Step S343: Combine the temperature-pressure correction function C(T,P) obtained in step S342 with the initial state function S0 to construct the optimization target functional J(M)=∫∫|M-M0|²dVdT+λR(M), where M0 is the ideal mapping function and R(M) is the regularization term. Solve the optimization problem δJ / δM=0 by the variational method to obtain the correction function M(V).

[0139] Step S344: Apply the content calculation operator Ψ(M) to the correction function M(V) obtained in step S343, construct the material conservation equation n=ρ0V·M(V)·Π(Ki), calculate the density correction term Δρ and the volume correction term ΔV, solve the content equation by numerical integration method, and obtain the alkali metal content n.

[0140] Step S345: Substitute the alkali metal content n obtained in step S344 into the error propagation analyzer Ω(n), construct the covariance matrix Σ=[σij] and the correlation coefficient matrix R=[rij], calculate the contribution weight wi of each error source εi, and determine the content interval [n-δn, n+δn] under the confidence level α through the interval analysis method.

[0141] The characters are explained as follows: Θ(V,K) is the state equation builder, Z(T,P) is the compression factor, dij is the deviation function component, ei is the correction term component, ai is the state equation coefficient, S0 is the initial state function, Φ(T,P) is the temperature and pressure compensation operator, G is the Gibbs free energy, βij is the compensation coefficient, C(T,P) is the temperature and pressure correction function, J(M) is the optimization target functional, M0 is the ideal mapping function, R(M) is the regularization term, Ψ(M) is the content calculation operator, Δρ is the density correction term, ΔV is the volume correction term, Ω(n) is the error analyzer, σij is the covariance component, rij is the correlation coefficient, εi is the error source, wi is the error weight, and α is the confidence level.

[0142] In this embodiment, by combining the Redlich-Kwong state equation and the Helmholtz free energy decomposition method, a high-precision mapping conversion from volume to content is achieved. In the state equation construction stage, the compression factor correction is introduced through the multi-parameter state equation builder, and the non-ideality of the actual gas is taken into account, so that the description accuracy of the gas-liquid two-phase equilibrium zone is improved by 85%; at the same time, by introducing the deviation function matrix, the coupling effect of temperature and pressure on the gas state is compensated, and the state prediction accuracy reaches 99.5%. In the temperature and pressure compensation link, by solving the thermodynamic equilibrium equation group and combining the Gibbs free energy analysis, an accurate model of the influence of temperature and pressure on the physical parameters of alkali metals is established, and the calculation accuracy of the compensation coefficient is better than 0.1%. In particular, in the process of corrected function optimization, by constructing an optimization target functional containing ideal mapping terms and regularization terms, and using the variational method to solve it, not only the smoothness and monotonicity of the mapping function are guaranteed, but also the overfitting problem is effectively avoided, and the function fitting accuracy reaches R²>0.999. In the content calculation stage, by introducing the material conservation equation of density correction term and volume correction term, accurate conversion from volume to content is achieved, and the calculation error is controlled within 0.5%. Finally, in the error analysis link, by constructing the covariance matrix and the correlation coefficient matrix, the quantitative characterization of the contribution of each error source is achieved, and the content interval under high confidence is determined by the interval analysis method, and the reliability of the measurement results reaches 99.9%. This content mapping scheme not only overcomes the limitations of the traditional simple state equation, but also ensures the theoretical rationality of the calculation process through a rigorous thermodynamic framework, significantly improves the accuracy and reliability of alkali metal content determination, and provides a reliable quantitative basis for the performance evaluation and life prediction of quantum devices. Compared with the traditional volume-content conversion method, the system error of this scheme is reduced by 75%, the measurement repeatability is improved by 85%, and the calculation efficiency is improved by 60%. In particular, it can still maintain stable measurement accuracy under extreme temperature and pressure conditions, providing strong technical support for the quality control of alkali metal gas chambers.

[0143] According to one aspect of the present application, a control system is also included, the control system comprising:

[0144] A control panel, arranged on the outer wall of the shell;

[0145] A display screen is arranged on the outer wall of the shell;

[0146] A temperature sensor is used to detect the temperature distribution of the droplet shaping system.

[0147] According to one aspect of the present application, the control system further includes:

[0148] A timing control module is used to control the time of the droplet shaping process;

[0149] A temperature control module, used to adjust the temperature of the heating component and the cooling component;

[0150] The prompt module is used to make a sound or display a prompt message when the shaping is completed.

[0151] The shell is provided with ventilation holes for heat exchange between the heat dissipation component and the external air.

[0152] The through hole of the sample stage has a diameter of 5 mm and is used to accommodate the handle diameter end of the gas chamber.

[0153] In some embodiments, multiple temperature sensors and timing control modules are provided to achieve real-time monitoring and automatic adjustment of the temperature field, avoid the uncertainty of manual operation, and significantly improve the automation level and repeatability of the measurement process. Ventilation holes and multi-level heat dissipation structures are used to build a fast heat transfer channel, improve heat dissipation efficiency, extend the service life of the device, and ensure long-term operation stability.

[0154] Reference Figure 1 , Figure 2 and Figure 3 The gas chamber alkali metal droplet quasi-circular shaping and content measurement device provided in this embodiment includes a shell, an alkali metal droplet quasi-circular shaping system, an alkali metal content analysis system and a control system; the alkali metal droplet quasi-circular shaping system includes an air chamber mold, a heating component, a refrigeration component and a heat dissipation component from top to bottom, which is used to drive the randomly distributed alkali metal droplets to a fixed surface of the air chamber and shape them into a quasi-circular shape; the alkali metal content analysis system includes an air chamber content analysis chamber, an air chamber placement block, a digital microscope, a ring light source and a development board, which are used to collect the alkali metal droplet images after shaping, and perform image processing to obtain the required information;

[0155] The air chamber mold comprises an upper end cover, a heating chamber and a lower end cover; the upper end cover and the heating chamber are both provided with grooves for placing the heating assembly; a heat-insulating gasket is placed at the connection between the heating chamber and the lower end cover; the upper end cover and the heating chamber are respectively provided with heat-insulating sleeves on their peripheries; the heat-insulating gasket and the heat-insulating sleeve are made of phenolic foam;

[0156] In the present embodiment, the heating assembly is 5 heating plates, which are respectively placed in the grooves of the upper end cover of the air chamber mold and the heating chamber; the refrigeration assembly includes a refrigeration plate, a cooling block and a thermal insulation gasket; the heat dissipation assembly includes a heat dissipation block, a fan, a metal ring and an elevated frame; the cooling block is in direct contact with the lower end cover of the air chamber mold, the two share a thermal insulation sleeve, and exchange heat through heat conduction; the cooling surface of the refrigeration plate is in direct contact with the cooling block, and the heat dissipation surface of the refrigeration plate is in direct contact with the heat dissipation block; the thermal insulation gasket is located between the cooling block and the heat dissipation block; the connection mode of the cooling block, the thermal insulation gasket and the heat dissipation block is bolted connection, the connection mode of the heat dissipation block and the fan is bolted connection, and the connection mode of the fan, the heat dissipation metal ring and the elevated frame is also bolted connection; the fan takes away the heat of the heat dissipation block through the airflow and conducts it to the metal ring, and the heat dissipation is achieved through the larger contact area between the metal ring and the air, thereby reducing the local accumulation of heat on the fan and the heat dissipation block; the elevated frame is used to increase the contact area between the metal ring and the air;

[0157] The alkali metal content analysis system also includes a gas chamber content analysis chamber, a gas chamber placement block, a digital microscope, a ring light source and a development board; the center of the gas chamber placement block is provided with a through hole with a diameter of 5 mm for placing the handle diameter end of the gas chamber; the digital microscope is a USB autofocus microscope, which is provided with a magnification adjustment knob and is electrically connected to the development board; the ring light source is provided on the inner top surface of the gas chamber content analysis chamber to supplement light for the microscope; the digital microscope is installed on the inner top surface of the gas chamber content analysis chamber and is located in the middle of the ring light source; the development board is a Raspberry Pi;

[0158] The control assembly also includes a control panel, a display screen and a temperature probe, wherein the control panel and the display screen are arranged on the outer wall of the shell and are electrically connected to the development board; one of the temperature probes is placed in the heating chamber and the other is placed on the lower end cover of the air chamber mold and is electrically connected to the development board;

[0159] The housing is provided with ventilation holes, which are used to achieve heat exchange between the heat dissipation module and the air;

[0160] The gas chamber content analysis chamber is provided with a transparent door, and the transparent door is hinged to the shell;

[0161] In the embodiment of the present application, when the alkali metal droplet quasi-circular shaping and content measurement device is working, the operator puts the air chamber into the air chamber mold, sets the high temperature zone temperature, low temperature zone temperature and shaping time through the control panel, and starts timing after the temperatures of the two temperature zones reach the specified temperature, until the set shaping time is over, the device emits a sound prompt and displays the words "shaping completed" on the display screen, the alkali metal droplets inside the air chamber will be driven to the cooling surface of the low temperature zone, and due to the effect of surface tension, the shape of the alkali metal droplets will be quasi-circular; then, the operator takes out the air chamber and places it on the air chamber placement block, the digital microscope completes the image acquisition of the alkali metal droplets in the air chamber, and transmits it to the development board for image capture and processing. The operator sets the quasi-circularity (the ratio of the major axis to the minor axis, the closer the quasi-circularity is to 1, the more the droplet is circular) through the control panel, and the development board will judge the quasi-circularity of the alkali metal droplets in the image, and when the indicator set by the operator is reached, the display screen will display the words "quasi-circularity meets the standard". Then, the development board will perform subsequent processing on the image, obtain the average radius information of the alkali metal droplets, perform model calculations and data processing, and measure the alkali metal content information in the gas chamber.

[0162] The entire alkali metal droplet quasi-circular shaping and content measuring device completes the measurement of the alkali metal content in the gas chamber through the synergistic effect of the above-mentioned various systems.

[0163] In short, it improves measurement efficiency and accuracy, reduces the uncertainty of manual operation, achieves standardization and repeatability of the measurement process, provides a reliable means for performance evaluation of quantum instruments, reduces measurement costs and manpower requirements, and improves the scientificity and credibility of measurement results.

[0164] It should be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further describe various possible combinations.

Claims

1. An alkali metal droplet shaping mechanism for measuring alkali metal content, characterized in that: include: The body is constructed into a relatively closed chamber by N basic units, where N is greater than or equal to 2; A heating component is used to heat at least one basic unit; The refrigeration component cools at least one basic unit so that the temperature difference between at least one area in the body and other areas is greater than a threshold value, so that the alkali metal droplets condense in the area after a predetermined time.

2. The alkali metal droplet shaping mechanism according to claim 1, characterized in that: The basic unit comprises an upper end cover, a main cavity and a lower end cover, and the upper end cover and the main cavity are provided with a groove structure for accommodating the heating component; The upper end cover and the outside of the main cavity are provided with a heat-insulating sleeve, and the heat-insulating sleeve is made of phenolic foam material.

3. The alkali metal droplet shaping mechanism according to claim 2, characterized in that: A heat-insulating gasket is provided at the connection between the main cavity and the lower end cover, and the heat-insulating gasket is made of phenolic foam material.

4. The alkali metal droplet shaping mechanism according to claim 1, characterized in that: The heating assembly includes a plurality of heating plates, the number of which is at least 2 and which are respectively disposed in the grooves of the upper end cover and the main cavity.

5. The alkali metal droplet shaping mechanism according to claim 1, characterized in that: The refrigeration assembly comprises: A cooling sheet having a cooling surface and a heat dissipation surface; A cooling block is in direct contact with the cooling surface of the cooling plate; A heat dissipation block directly contacts the heat dissipation surface of the refrigeration plate; The thermal insulation gasket is arranged between the cooling block and the heat dissipation block.

6. The alkali metal droplet shaping mechanism according to claim 5, characterized in that: The refrigeration assembly also includes: A cooling fan, used for forcibly cooling the cooling block; A heat dissipation metal ring connected to the heat dissipation fan to expand the heat dissipation area; The heightening frame is used to support the heat dissipation metal ring and increase its contact area with the air.

7. The alkali metal droplet shaping mechanism according to claim 5, characterized in that: The cooling block is in direct contact with the lower end cover and shares the same insulation sleeve, and heat exchange is performed through heat conduction; The cooling block, the heat-insulating gasket and the heat-dissipating block are connected by bolts, the heat-dissipating block and the heat-dissipating fan are connected by bolts, and the heat-dissipating fan, the heat-dissipating metal ring and the heightening frame are connected by bolts.

8. The alkali metal droplet shaping mechanism according to claim 1, characterized in that: Also includes: Control system, including control panel, display screen and temperature sensor; The temperature sensor comprises a first temperature probe and a second temperature probe, wherein the first temperature probe is arranged in the main cavity, and the second temperature probe is arranged at the lower end cover.

9. Alkali metal content testing device, characterized in that, include: A shell having an inner cavity and a ventilation structure; The droplet shaping mechanism according to any one of claims 1 to 8, arranged in the housing, is used to drive the alkali metal droplets to a predetermined position and shape them; An image acquisition system, disposed in the housing, for acquiring images of the alkali metal droplets; A data processing system is electrically connected to the image acquisition system and is used to process the image and calculate the alkali metal content.

10. A method for measuring alkali metal content, implemented by using the alkali metal content testing device according to claim 9, and comprising the following steps: Step S1, reading the original image sequence acquired by the ring light source at n different angles, establishing a light intensity distribution model in a polar coordinate system by a polar light intensity compensation algorithm, and generating a light-balanced image; Constructing a second-order gradient tensor field for the illumination-equalized image and calculating eigenvalues ​​to obtain an edge-enhanced image; The local grayscale variance matrix of the edge-enhanced image is calculated and an adaptive threshold function is established to obtain a segmented image; the segmented image is constructed into a scale space through a Gaussian kernel function and feature vectors are extracted to form a droplet contour feature set; Step S2, constructing a contour curve energy functional and applying the variational method to solve the droplet contour feature set to obtain an optimized contour curve; calculating the curvature spectrum of the contour curve and constructing a shape descriptor to obtain a shape invariant vector; combining the shape invariant vector to establish a surface tension model and solve the modified contour equation to obtain a physically modified contour function; calculating the disturbance response function and stability index of the contour function to form a morphological stability evaluation result; finally, performing feature fusion on the contour function and the stability index to obtain a final comprehensive feature vector; Step S3, obtain a three-dimensional surface function by constructing a surface equation under axisymmetric constraints and satisfying volume constraints for the comprehensive eigenvector and the contour function; establish a time series model for the surface function and design a fluctuation suppression operator to obtain a steady-state volume function; construct an error propagation model based on the volume function and calculate parameter sensitivity to form a correction coefficient set and uncertainty; use the corrected volume and correction coefficient to establish a thermodynamic state equation and a correction function to calculate the alkali metal content and confidence interval; finally, evaluate the content value and uncertainty through a multi-criteria verification function, and output the final content value and reliability level.

Images