Melting point detection method based on automatic parameter correction and high-precision melting point detector

Abstract

The invention discloses a melting point detection method based on automatic parameter correction and a high-precision melting point detector, and belongs to the technical field of instruments and meters and image recognition. According to the invention, the compensation temperature is calculated according to the temperature and the environmental parameters, and then the melting point of the to-be-detected sample is calculated according to the temperature parameter and the compensation temperature of the to-be-detected sample, so that compared with the melting point detected by a temperature probe, the environmental parameters and the temperature parameters related to the properties of the sample are added in the calculation process; the influence of the environment and the properties of the sample on the confidence of the detection result is integrated; besides, since the temperature change in the placing groove can be accurately reflected in the image, compared with a temperature probe, temperature parameters and environment parameters are obtained through image information, errors generated by the temperature probe are avoided, the confidence coefficient of a detection result is improved, and compared with the prior art, through multiple times of detection, the detection frequency is reduced, and the detection efficiency is improved. And compared with the prior art, the test time is shortened, and the test efficiency is improved.

Description

Technical Field

[0001] The invention relates to the fields of instrumentation technology and image recognition technology, and in particular to a melting point detection method based on automatic parameter correction and a high-precision melting point detector. Background Art

[0002] According to the definition of physical chemistry, the melting point of a substance refers to the temperature at which the substance changes from solid to liquid. In the field of organic chemistry, melting point determination is a basic means of identifying the nature of a substance and is also one of the important methods for purity determination. Therefore, the melting point tester plays an important role in the chemical industry and pharmaceutical research, and is an essential instrument for the production of drugs, spices, dyes and other organic crystalline substances.

[0003] During the use of the existing melting point tester, the test accuracy of the entire melting point tester is actually related to multiple factors such as the environment and the properties of the sample to be tested. Any factor will affect the confidence of the test result. In order to ensure the confidence, the existing technology often adopts multiple tests and then eliminates the values ​​with large deviations to ensure the confidence, but this scheme increases the test time and reduces the test efficiency. Summary of the invention

[0004] In order to solve the problems of the prior art, the embodiment of the present invention provides a melting point detection method based on automatic parameter correction and a high-precision melting point detector. The technical solution is as follows: On the one hand, a melting point detection method based on automatic parameter correction is provided, the method is applied to a high-precision melting point instrument, the high-precision melting point instrument includes a placement tank and a temperature probe, the method includes: After starting the melting point test, obtaining image information in the placement tank and the temperature in the placement tank measured by the temperature probe; Identifying a temperature parameter of a sample to be tested in the image information; identifying environmental parameters in the image information; Calculating a compensation temperature according to the temperature and the environmental parameters; The melting point of the sample to be tested is calculated according to the temperature parameter and the compensation temperature.

[0005] Optionally, the identifying a temperature parameter of the sample to be tested in the image information includes: Identifying a sample to be tested in the image information; The temperature parameter is identified according to the characteristic value of the image region where the sample to be tested is located.

[0006] Optionally, after identifying the temperature parameter of the sample to be tested in the image information, the method further includes: A heating parameter of the sample to be tested is calculated according to the temperature parameter, and the heating parameter is used to indicate the heating uniformity of the sample to be tested.

[0007] Optionally, the identifying the environmental parameters in the image information includes: The environmental parameter is identified according to the characteristic value of the image information in the placement slot, and the environmental parameter is used to indicate the environmental temperature in the placement slot.

[0008] Optionally, calculating the compensation temperature according to the temperature and the environmental parameter includes: calculating a correlation parameter between the temperature, the heating parameter and the environmental parameter; Based on the RBF network, a temperature compensation model corresponding to the temperature, the heating parameter and the environmental parameter is constructed; the temperature, the heating parameter and the environmental parameter are input into the corresponding temperature compensation model respectively, and an initial compensation temperature is output; Calculating a secondary compensation temperature according to the correlation parameter and initial compensation temperatures corresponding to the temperature, the heating parameter and the environmental parameter respectively; The secondary compensation temperature is input into a time series prediction model to obtain the compensation temperature.

[0009] Optionally, the calculating the melting point of the sample to be tested according to the temperature parameter and the compensation temperature includes: constructing a melting point determination model corresponding to the temperature parameter and the compensation temperature respectively based on a decision forest algorithm; Input the temperature parameter and the compensation temperature into the corresponding melting point determination model, and output the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature; The melting point is calculated according to the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature.

[0010] Optionally, after acquiring the image information in the placement slot and the temperature measured by the temperature probe, the method further includes: Obtaining reference image information in the placement tank and reference temperature measured by the temperature probe; Correcting the temperature parameter according to the reference image information; The temperature is corrected according to the reference temperature.

[0011] On the other hand, a high-precision melting point detector is provided, the high-precision melting point detector comprising a placement slot, a temperature probe, an image acquisition module and a processing module, wherein: After the melting point test starts, the image acquisition module is used to obtain image information in the placement tank and the temperature in the placement tank measured by the temperature probe; The processing module is used to identify the temperature parameter of the sample to be tested in the image information; The processing module is used to identify environmental parameters in the image information; The processing module is used to calculate the compensation temperature according to the temperature and the environmental parameters; The processing module is used to calculate the melting point of the sample to be tested according to the temperature parameter and the compensation temperature.

[0012] Optionally, the processing module is specifically used for: Identifying a sample to be tested in the image information; The temperature parameter is identified according to the characteristic value of the image region where the sample to be tested is located.

[0013] Optionally, the processing module is specifically used for: A heating parameter of the sample to be tested is calculated according to the temperature parameter, and the heating parameter is used to indicate the heating uniformity of the sample to be tested.

[0014] Optionally, the processing module is specifically used for: The environmental parameter is identified according to the characteristic value of the image information in the placement slot, and the environmental parameter is used to indicate the environmental temperature in the placement slot.

[0015] Optionally, the processing module is specifically used for: calculating a correlation parameter between the temperature, the heating parameter and the environmental parameter; Based on the RBF network, a temperature compensation model corresponding to the temperature, the heating parameter and the environmental parameter is constructed; the temperature, the heating parameter and the environmental parameter are input into the corresponding temperature compensation model respectively, and an initial compensation temperature is output; Calculating a secondary compensation temperature according to the correlation parameter and initial compensation temperatures corresponding to the temperature, the heating parameter and the environmental parameter respectively; The secondary compensation temperature is input into a time series prediction model to obtain the compensation temperature.

[0016] Based on the decision forest algorithm, a melting point determination model corresponding to the temperature parameter and the compensation temperature is constructed; the temperature parameter and the compensation temperature are input into the corresponding melting point determination model, and a predicted value corresponding to the temperature parameter and a predicted value corresponding to the compensation temperature are output; The melting point is calculated according to the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature.

[0017] Optionally, the processing module is specifically used for: Obtaining reference image information in the placement tank and reference temperature measured by the temperature probe; Correcting the temperature parameter according to the reference image information; The temperature is corrected according to the reference temperature.

[0018] Optionally, the high-precision melting point detector also includes a network module, which is used to transmit at least one of the image information, the temperature, the temperature parameter, the environmental parameter, the calculated compensation temperature and the melting point.

[0019] The technical solution provided by the embodiment of the present invention has the following beneficial effects: 1. By calculating the compensation temperature according to the temperature and environmental parameters, and then calculating the melting point of the sample to be tested according to the temperature parameters of the sample to be tested and the compensation temperature, compared with detecting the melting point by a temperature probe, the environmental parameters and the temperature parameters related to the properties of the sample itself are added in the calculation process, and the influence of the environment and the properties of the sample itself on the confidence of the test results is combined, thereby improving the confidence of the test results. Compared with the existing technology, the test time is reduced and the test efficiency is improved; 2. During the testing process, the temperature parameters and environmental parameters of the sample to be tested are obtained by image recognition of the image information in the placement tank. Since the temperature changes in the placement tank will be accurately reflected in the image, the temperature parameters and environmental parameters of the sample to be tested are obtained through image information. Compared with obtaining the above parameters with a temperature probe, the error caused by the temperature probe itself is avoided, thereby improving the confidence of the test results; 3. This method generates a compensation temperature by combining the environmental parameters and temperature parameters in the image information in a single detection, compensates the temperature measured by the temperature probe, and obtains the melting point of the sample to be tested. Compared with the existing technology, multiple tests are performed to eliminate the deviation value of the temperature measured by the temperature probe. While ensuring the confidence level, the number of tests is reduced. Moreover, compared with the existing technology, the test time is reduced and the test efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 is a flow chart of a melting point detection method based on automatic parameter correction provided by an embodiment of the present invention; Figure 2 is a flow chart of a melting point detection method based on automatic parameter correction provided by an embodiment of the present invention; Figure 3 It is a structural schematic diagram of a high-precision melting point detector provided by an embodiment of the present invention; Figure 4 It is a schematic diagram of a melting point detection system based on automatic parameter correction provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0023] Reference Figure 1 As shown, a melting point detection method based on automatic parameter correction is provided, the method is applied to a high-precision melting point instrument, the high-precision melting point instrument includes a placement tank and a temperature probe, the method includes: 101, after starting the melting point test, obtaining image information in the placement tank, and the temperature in the placement tank measured by the temperature probe; 102. Identify the temperature parameter of the sample to be tested in the image information; 103. Identify environmental parameters in image information; 104. Calculate the compensation temperature according to the temperature and environmental parameters; 105. Calculate the melting point of the sample to be tested based on the temperature parameters and compensation temperature.

[0024] The melting point detection method described in the embodiment of the present invention calculates the compensation temperature through the temperature and environmental parameters during the detection process, and then calculates the melting point of the sample to be tested according to the temperature parameters of the sample to be tested and the compensation temperature. Compared with the melting point detection by the temperature probe, the environmental parameters and the temperature parameters related to the properties of the sample itself are added in the calculation process, and the influence of the environment and the properties of the sample itself on the confidence of the detection result is combined, thereby improving the confidence of the detection result; In addition, during the detection process, the temperature parameters and environmental parameters of the sample to be tested are obtained by image recognition of the image information in the placement tank. Since the temperature changes in the placement tank will be accurately reflected in the image, the temperature parameters and environmental parameters of the sample to be tested are obtained through image information. Compared with obtaining the above parameters with a temperature probe, the error caused by the temperature probe itself is avoided, thereby improving the confidence of the detection result; Moreover, this method generates a compensation temperature by combining the environmental parameters and temperature parameters in the image information in a single detection, compensates the temperature measured by the temperature probe, and obtains the melting point of the sample to be tested. Compared with the prior art, which eliminates the deviation value of the temperature measured by the temperature probe through multiple tests, this method reduces the number of tests while ensuring confidence, thereby reducing the test time and improving the test efficiency.

[0025] Reference Figure 2 As shown, in another preferred embodiment, the process of identifying the temperature parameter of the sample to be tested in the image information can be specifically as follows: 201. Build an image recognition model through convolutional neural network; 202. Identify the sample to be tested in the image information through an image recognition model; 203. Identify the temperature parameter according to the characteristic value of the image region where the sample to be tested is located.

[0026] The above process of building an image recognition model can be specifically as follows: A convolutional layer, a sampling layer, and a fully connected layer of a convolutional neural network are set, and the sampling layer described in the embodiment of the present invention adopts downsampling; multiple types of sample images to be tested are used as training sets to train the convolutional neural network, and the embodiment of the present invention does not limit the specific convergence function; After the convolutional neural network training is completed, the image recognition model is obtained.

[0027] Through the above image recognition model, identifying the sample to be tested in the image information includes: Extract the image to be identified from the image captured by the image acquisition module at a preset time interval (e.g., 0.1s); The image to be identified is input into the image recognition model, and the image area where the sample to be tested is located is output; You can also: The image acquisition module is set to capture the image to be identified at a preset time interval, and the image to be identified is input into the image recognition model to output the image area where the sample to be tested is located.

[0028] The characteristic value described in the embodiment of the present invention is the pixel value of the pixel point. According to the characteristic value of the image area where the sample to be tested is located, the color and shape of the heated sample to be tested will change at different temperatures. The reason is that the atoms of the sample to be tested are excited at different high temperatures, and the electron transition generates light of a specific wavelength. With continuous heating, the number of activated metal atoms continues to increase, and the color of the sample to be tested will also change. The sample to be tested will also be converted from solid to liquid with heating, and the above-mentioned changes in color and shape will be reflected in the image of the sample to be tested, that is, the image pixel value of the sample to be tested will change with heating, so the process of identifying the temperature parameter can be: A temperature recognition model is preset, which is also constructed based on a convolutional neural network; its input is set as the feature value of the image area, that is, the pixel value of all pixels in the image area; its output is the temperature value of the image area, and the training process of the temperature recognition model can be specifically as follows: For the same type of sample to be tested, the sample to be tested is continuously heated, and the temperature of the sample to be tested is measured in real time, and images of the sample to be tested at different temperatures are collected; in practical applications, the image of the sample to be tested can be collected every time the temperature changes by 0.1°C from the beginning of heating until the sample to be tested melts; Through the image recognition model, for an image corresponding to any temperature, the image area where the sample to be tested is located in the image is identified, and for the image information corresponding to the next temperature, the image area where the sample to be tested is located in the temperature image is continuously identified, until the image areas where the sample to be tested is located in the images corresponding to all temperatures are identified, and multiple image areas where the sample to be tested is located corresponding to the temperature are output; (in actual use, the multiple image areas where the sample to be tested is located are marked or named according to the temperature); Extracting characteristic values ​​of the image regions where the multiple samples to be tested corresponding to the temperature are located; Set the input value of the convolutional neural network as the feature value of the image area where the sample to be tested is located, and the output value is the predicted temperature; For the characteristic value of the image area where the sample to be tested is located corresponding to any temperature, perform the following operations: The characteristic value of the image area where the sample to be tested is located corresponding to the temperature is input into the convolutional neural network, and the predicted temperature is output; According to the difference between the temperature and the predicted temperature, the loss function of the convolutional neural network is adjusted to obtain an adjusted convolutional neural network; the feature value of the image area where the sample to be tested is located corresponding to the next temperature is continuously input into the adjusted convolutional neural network, the predicted temperature is output, and the loss function is continuously adjusted according to the difference between the temperature and the predicted temperature; When the difference between the temperature and the predicted temperature reaches 0, the training is stopped and the trained convolutional neural network is used as the temperature recognition model.

[0029] The characteristic value of the image area where the sample to be tested is located is input into the temperature recognition model, and the temperature parameter is output.

[0030] In another preferred embodiment, after identifying the temperature parameter of the sample to be tested in the image information, the method further comprises: The heating parameter of the sample to be tested is calculated according to the temperature parameter of the sample to be tested. The heating parameter is used to indicate the heating uniformity of the sample to be tested. The heating parameter is the temperature average of all pixels in the image area where the sample to be tested is located.

[0031] The above calculation process can be specifically as follows: Calculate the average temperature of all pixels in the image area where the sample to be tested is located; this process includes: Calculate the temperature of all pixels in the image area where the sample to be tested is located; this method is the same as the calculation process of the temperature parameter. Since the temperature recognition model can identify the temperature corresponding to the image area containing multiple pixels, then: For any pixel point in the image area where the sample to be tested is located, the pixel point is input into the temperature recognition model as an image area containing only one pixel point, and the temperature of the image area containing only one pixel point is output, that is, the temperature of the pixel point; according to the temperature of all pixel points and the number of pixel points, the average temperature of all pixel points is obtained; Calculate the temperature of all pixels and the difference between them and the mean temperature; If the number of differences between 0 and 0.5 in the plurality of differences is greater than or equal to 80% of the total of the differences, the heating parameter is a1; If the number of differences between 0 and 0.5 in the plurality of differences is less than 80% of the total difference, and the number of differences between 0.5 and 1 in the plurality of differences is less than 80% of the total difference, the heating parameter is a2; If the number of differences between 0.5 and 1 in the plurality of differences is greater than or equal to 80% of the total of the differences, the heating parameter is b1; If the number of differences between 0.5 and 1 in the plurality of differences is less than 80% of the total difference, and the number of differences between 1 and 1.5 in the plurality of differences is less than 80% of the total difference, then the heating parameter is b2; and so on; It should be noted that 0 <a1<a2<<b1<b2<1。

[0032] In another preferred embodiment, identifying environmental parameters in the image information includes: According to the characteristic value of the image information in the placement slot, the environmental parameter is identified, and the environmental parameter is used to indicate the ambient temperature in the placement slot.

[0033] The process of identifying the ambient temperature includes: The feature values ​​of the remaining images in the image information except the image area where the sample to be tested is located are input into the temperature recognition model to calculate the temperature of all pixels in the remaining images; this method is the same as calculating the temperature of all pixels in the image area where the sample to be tested is located, that is: For any pixel point of the remaining images, the pixel point is input into the temperature recognition model according to the image area containing only one pixel point, and the temperature of the image area containing only one pixel point is output, which is the temperature of the pixel point; According to the temperature of all pixels and the number of pixels, the average temperature of all pixels is obtained, and the average temperature is the ambient temperature.

[0034] In another preferred embodiment, calculating the compensation temperature according to the temperature and the environmental parameters comprises: Calculate correlation parameters between temperature, heating parameters and environmental parameters; Based on the RBF network, temperature compensation models corresponding to temperature, heating parameters and environmental parameters are constructed respectively; Input the temperature, heating parameters and environmental parameters into the corresponding temperature compensation model respectively, and output the initial compensation temperature; Calculate the secondary compensation temperature according to the correlation parameter and the initial compensation temperatures corresponding to the temperature, heating parameter and environmental parameter respectively; The quadratic compensated temperature is input into the time series prediction model to obtain the compensated temperature.

[0035] The process of calculating the correlation parameters between the temperature, heating parameters and environmental parameters can be realized by the Pearson correlation coefficient, specifically: By the following formula: Among them, X and Y represent the observed values ​​of two variables, and and represent their means respectively; For temperature, the above formulas are used to obtain the correlation coefficients Tr1 and Tr2 with heating parameters and environmental parameters respectively; For heating parameters, the correlation coefficients Hr1 and Hr2 with temperature and environmental parameters are obtained through the above formulas respectively; For environmental parameters, the correlation coefficients Er1 and Er2 with temperature and heating parameters are obtained through the above formulas respectively; It should be noted that the value of the correlation parameter Tr1 between the temperature and the heating parameter is equal to the value of the correlation parameter Hr1 between the heating parameter and the temperature. Similarly, the values ​​of Tr2 and Er1 are equal, and the values ​​of Hr2 and Er2 are equal.

[0036] The above process of constructing a temperature compensation model based on the RBF network can be specifically as follows: Since the RBF network can predict time series and the temperature increases with time, the temperature at any time can be predicted after the RBF network is trained through the time in the historical data and the temperature corresponding to the time; since the increasing values ​​of temperature, environmental parameters and heating parameters are different over time, the RBF networks corresponding to temperature, environmental parameters and heating parameters are set respectively; the input is time, the output of the RBF network corresponding to temperature is the predicted value of temperature at that time, the output of the RBF network corresponding to environmental parameters is the predicted value of environmental parameters at that time, and the output of the RBF network corresponding to heating parameters is the predicted value of heating parameters at that time; In the RBF network described in the embodiment of the present invention, the activation function of the hidden layer neuron is a Gaussian radial basis function; the Gaussian radial basis function is: Among them, c j represents the center corresponding to the jth hidden layer neuron, x i represents sample data, and ||·|| represents Euclidean distance; wherein, for temperature, the sample data is the value of temperature changing with time, that is, the sample data includes multiple moments from the start of heating to the current moment, and the temperatures corresponding to the multiple moments; similarly, for heating parameters, the sample data includes multiple moments from the start of heating to the current moment, and the heating parameters corresponding to the multiple moments; for environmental parameters, the environmental data includes multiple moments from the start of heating to the current moment, and the environmental parameters corresponding to the multiple moments.

[0037] The RBF network can be: Among them, q is the number of neurons in the hidden layer of the RBF network, c j and w j are the center and weight corresponding to the jth hidden layer neuron, p(θ i , c j ) is a radial basis function, which is a scalar function that is symmetric along the radial direction and is the sample θ i To data center c j A monotonic function of the Euclidean distance between the input vector elements θ through the final RBF training i After that, the real value of the output; set the temperature compensation model to RBF network and the initial compensation temperature calculation formula; The calculation formula of initial compensation temperature is: initial compensation temperature = predicted value at current time - input value at current time; the predicted value at current time is the output value of RBF network; That is: input temperature and time into the temperature compensation model; the output value is the initial compensation temperature corresponding to the temperature, which is specifically: the output value of the RBF network - temperature. Similarly, the initial compensation temperature corresponding to the heating parameters and the initial compensation temperature corresponding to the environmental parameters can be obtained.

[0038] In the above process, according to the correlation parameter and the initial compensation temperature corresponding to the temperature, heating parameter and environmental parameter, the process of calculating the secondary compensation temperature can be specifically as follows: For temperature, the secondary compensation temperature is: Secondary compensation temperature corresponding to temperature = (initial compensation temperature corresponding to heating parameter*Tr1) / secondary compensation temperature corresponding to heating parameter+(initial compensation temperature corresponding to environmental parameter*Tr2) / secondary compensation temperature corresponding to environmental parameter.

[0039] For the heating parameters, the secondary compensation temperature is: The secondary compensation temperature corresponding to the heating parameter = (the initial compensation temperature corresponding to the temperature * Hr1) / the secondary compensation temperature corresponding to the temperature + (the initial compensation temperature corresponding to the environmental parameter * Hr2) / the secondary compensation temperature corresponding to the environmental parameter.

[0040] For environmental parameters, the secondary compensation temperature is: The secondary compensation temperature corresponding to the environmental parameter = (the initial compensation temperature corresponding to the temperature * Er1) / the secondary compensation temperature corresponding to the temperature + (the initial compensation temperature corresponding to the heating parameter * Er2) / the secondary compensation temperature corresponding to the heating parameter.

[0041] The above-mentioned process of inputting the quadratic compensation temperature into the time series prediction model to obtain the compensation temperature can also be: inputting the quadratic compensation temperature corresponding to the temperature into the time series prediction model, setting the time series prediction model as the RBF network + output value calculation formula; its output value calculation formula is: (quadratic compensation temperature prediction value + input value) / 2, wherein the RBF network is trained by time and quadratic compensation temperature, and RBF networks corresponding to temperature, environmental parameters and heating parameters are obtained respectively; the quadratic compensation temperature prediction value is the output value of the RBF network, and the input value is the quadratic compensation temperature; Input the current time into the corresponding RBF network, and output the quadratic compensation temperature prediction value corresponding to the temperature, the quadratic compensation temperature prediction value corresponding to the environmental parameter, and the quadratic compensation temperature prediction value corresponding to the heating parameter respectively; The output value calculation formula can be obtained: Temperature output value = (secondary compensation temperature prediction value corresponding to temperature + secondary compensation temperature corresponding to temperature) / 2; Similarly, the environmental parameter output value and the heating parameter output value are obtained in sequence; the time series prediction model can be a RBF network.

[0042] Compensation temperature = temperature output value*(Hr1*Er1)+environmental parameter output value*(Tr2*Hr2)+heating parameter output value*(Tr1*Er2).

[0043] In addition, the calculation process of the secondary compensation temperature can also be: Secondary compensation temperature = initial compensation temperature corresponding to temperature * (Hr1*Er1) + initial compensation temperature corresponding to environmental parameters * (Tr2*Hr2) + initial compensation temperature corresponding to environmental parameters * (Tr1*Er2); The secondary compensation temperature is input into the time series prediction model to obtain the compensation temperature.

[0044] In another preferred embodiment, according to the temperature and environmental parameters, the compensation temperature can also be calculated as follows: Input the temperature, heating parameters and environmental parameters into the corresponding temperature compensation model respectively, and output the corresponding compensation value; After the temperature compensation model outputs the compensation value corresponding to the temperature, the compensation value corresponding to the heating parameter, and the compensation value corresponding to the environmental parameter respectively, the following formula is used to output the compensation temperature.

[0045] Compensation temperature = compensation value corresponding to temperature * c1 + compensation value corresponding to heating parameter * c2 + compensation value corresponding to environmental parameter * c3; Wherein, c1, c2 and c3 are between 0 and 1, wherein c1 is the correlation parameter between temperature and melting point, c2 is the correlation parameter between heating parameter and melting point, and c3 is the correlation parameter between environmental parameter and melting point; The c1 can be the melting point and temperature in the historical data of melting point detection, calculated by the Pearson correlation coefficient. The melting point is the melting point of the sample to be tested, and the temperature is the temperature in the placement tank measured by the temperature probe.

[0046] Similarly, c2 and c3 can be obtained.

[0047] In another preferred embodiment, calculating the melting point of the sample to be tested according to the temperature parameter and the compensation temperature includes: constructing melting point determination models corresponding to the temperature parameter and the compensation temperature respectively based on a decision forest algorithm; Input the temperature parameter and the compensation temperature into the corresponding melting point determination model, and output the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature; The melting point is calculated based on the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature.

[0048] Among them, for the prediction value with the highest score output by the two decision forests, the melting point is calculated according to the following formula: Melting point = temperature parameter + compensation temperature * d; The value of d is between 0 and 1. In practical applications, the value of d can be set manually.

[0049] In another preferred embodiment, after acquiring the image information in the placement slot and the temperature measured by the temperature probe, the method further includes: Obtaining reference image information in the placement tank and reference temperature measured by the temperature probe; the reference image information is the image information in the placement tank before the melting point test is started; the reference temperature is the temperature in the placement tank measured by the temperature probe before the melting point test is started; correcting the temperature parameters according to the reference image information; Correct the temperature according to the reference temperature; Among them, the process of correcting the temperature parameter according to the reference image information can be: Calculating a first temperature average value in the reference image information. The calculation method of the first temperature average value is the same as the calculation method of the heating parameter, which will not be described in detail here; After the melting point test is started, a second temperature average value in the image information in the placement tank is calculated in real time; Calculate the absolute value of the difference between the first temperature average and the second temperature average; If the difference between the absolute value of the difference and the environmental parameter is outside the interval [0, 1], the calculation time of the second temperature average is recorded, and the image information collected during the calculation time and the temperature measured by the temperature probe are deleted; The above process of correcting the temperature according to the reference temperature can be: The standard temperature of the temperature probe is set before the melting point test is started; if the difference of the standard temperature is different from the reference temperature, step 101 is stopped until the difference of the standard temperature is the same as the reference temperature.

[0050] Optionally, after acquiring the image information in the placement tank and the temperature in the placement tank measured by the temperature probe, the method further includes: Acquire multiple melt images of the sample to be tested; The image recognition model is trained through multiple melt images; After the melting point test is started, steps 101 to 105 are executed in a loop, and the image recognition model determines in real time whether the sample to be tested is melted. If it is melted, steps 101 to 105 are continued and the process ends.

[0051] Reference Figure 3 As shown, a high-precision melting point detector is provided, which includes a placement slot 301, a temperature probe 302, an image acquisition module 303 and a processing module 304, wherein: After the melting point test starts, the image acquisition module 303 is used to obtain the image information in the placement slot 301 and the temperature measured by the temperature probe 302; The processing module 304 is used to identify the temperature parameters of the sample to be tested in the image information; The processing module 304 is used to identify environmental parameters in the image information; The processing module 304 is used to calculate the compensation temperature according to the temperature and environmental parameters; The processing module 304 is used to calculate the melting point of the sample to be tested according to the temperature parameter and the compensation temperature.

[0052] Optionally, the processing module 304 is specifically used for: Identify the sample to be tested in the image information; The temperature parameter is identified according to the characteristic value of the image area where the sample to be tested is located.

[0053] Optionally, the processing module 304 is specifically used for: According to the temperature parameter of the sample to be tested, the heating parameter of the sample to be tested is identified, and the heating parameter is used to indicate the heating uniformity of the sample to be tested.

[0054] Optionally, the processing module 304 is specifically used for: According to the characteristic value of the image information in the placement slot, the environmental parameter is identified, and the environmental parameter is used to indicate the ambient temperature in the placement slot.

[0055] Optionally, the processing module 304 is specifically used for: Based on RBF network, a temperature compensation model is constructed; Input temperature and environmental parameters into the temperature compensation model and output the compensation temperature; Based on the decision forest algorithm, a melting point determination model was constructed; The temperature parameters and compensation temperature are input into the melting point determination model, and the melting point is output.

[0056] Optionally, the processing module 304 is specifically used for: Obtain reference image information in the placement tank and reference temperature measured by the temperature probe.

[0057] Optionally, the high-precision melting point detector further includes a network module 305, which is used to transmit at least one of image information, temperature parameters, environmental parameters, calculated compensation temperature and melting point.

[0058] It should be noted that, in practical applications, in addition to the technical solution implemented by the processing module being completed locally, that is, configuring the processing module on the high-precision melting point detector and configuring the image acquisition module accordingly, the high-precision melting point detector described in the embodiment of the present invention can further be: Reference Figure 4 As shown, multiple node processing devices 401 and a processing server 402 are configured; the high-precision melting point detector 403 includes a network module 4031, an image acquisition module 4032, a placement slot 4033, and a temperature probe 4034; The node processing device 401 is connected to the high-precision melting point detector 403 through the network module 4031; the node processing device 401 is connected to the processing server 402; In order to ensure processing efficiency, the node processing device 401 can execute the steps performed by the processing module, thereby implementing the solution described in the embodiment of the present invention by means of edge processing nodes; In order to improve accuracy, the processing server 402 can execute the steps performed by the processing module. Since the image recognition model, temperature compensation model and melting point determination model configured by the processing server are often more complex and sophisticated than those of the node processing device, and the performance of the processor is also often better than that of the node processing device; based on the configuration of the database and the big data model, the processing server 402 can achieve more refined training and updating of the above models through a large amount of training sample data, thereby improving confidence.

[0059] During the process of the processing server or node processing device 401 executing the steps executed by the processing module, the image acquisition module 4032 can send the initial image and residual data to the processing server 402 or the node processing device 401. The residual data is the pixel value of the pixel point in other images, which is the difference between the pixel value of the same pixel point in the initial image.

[0060] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present invention, which will not be described in detail here.

[0061] It should be noted that: when the high-precision melting point detector provided in the above embodiment performs the melting point detection method with the automatic parameter correction function, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the system and the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the melting point detection method with the automatic parameter correction function provided in the above embodiment and the embodiment of the high-precision melting point detector belong to the same concept. The specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0062] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A melting point detection method based on automatic parameter correction, characterized in that: The method is applied to a high-precision melting point instrument, the high-precision melting point instrument comprises a placement tank and a temperature probe, and the method comprises: After starting the melting point test, obtaining image information in the placement tank and the temperature in the placement tank measured by the temperature probe; Identifying a temperature parameter of a sample to be tested in the image information; identifying environmental parameters in the image information; Calculating a compensation temperature according to the temperature and the environmental parameters; The melting point of the sample to be tested is calculated according to the temperature parameter and the compensation temperature.

2. The method according to claim 1, characterized in that: The identifying the temperature parameter of the sample to be tested in the image information comprises: Identifying a sample to be tested in the image information; Identifying the temperature parameter according to the characteristic value of the image area where the sample to be tested is located; The image area is the image of the sample to be tested in the image information.

3. The method according to claim 2, characterized in that After identifying the temperature parameter of the sample to be tested in the image information, the method further includes: A heating parameter of the sample to be tested is calculated according to the temperature parameter, and the heating parameter is used to indicate the heating uniformity of the sample to be tested.

4. The method according to claim 3, characterized in that The identifying of the environmental parameters in the image information comprises: The environmental parameter is identified according to the feature value of the image information, and the environmental parameter is used to indicate the environmental temperature in the placement tank.

5. The method according to claim 4, characterized in that Calculating the compensation temperature according to the temperature and the environmental parameter comprises: calculating a correlation parameter between the temperature, the heating parameter and the environmental parameter; Based on the RBF network, a temperature compensation model corresponding to the temperature, the heating parameter and the environmental parameter is constructed; respectively inputting the temperature, the heating parameter and the environmental parameter into the corresponding temperature compensation model, and outputting an initial compensation temperature; Calculating a secondary compensation temperature according to the correlation parameter and initial compensation temperatures corresponding to the temperature, the heating parameter and the environmental parameter respectively; The secondary compensation temperature is input into a time series prediction model to obtain the compensation temperature.

6. The method according to claim 5, characterized in that Calculating the melting point of the sample to be tested according to the temperature parameter and the compensation temperature includes: Based on the decision forest algorithm, constructing a melting point determination model corresponding to the temperature parameter and the compensation temperature respectively; Input the temperature parameter and the compensation temperature into the corresponding melting point determination model, and output the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature; The melting point is calculated according to the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature.

7. The method according to claim 6, characterized in that After acquiring the image information in the placement slot and the temperature measured by the temperature probe, the method further includes: Obtaining reference image information in the placement tank and reference temperature measured by the temperature probe; Correcting the temperature parameter according to the reference image information; The temperature is corrected according to the reference temperature.

8. A high-precision melting point detector, characterized in that: The high-precision melting point detector includes a placement tank, a temperature probe, an image acquisition module and a processing module, wherein: After the melting point test starts, the image acquisition module is used to obtain image information in the placement tank and the temperature in the placement tank measured by the temperature probe; The processing module is used to identify the temperature parameter of the sample to be tested in the image information; The processing module is used to identify environmental parameters in the image information; The processing module is used to calculate the compensation temperature according to the temperature and the environmental parameters; The processing module is used to calculate the melting point of the sample to be tested according to the temperature parameter and the compensation temperature.

9. The high-precision melting point detector according to claim 8, characterized in that: The processing module is specifically used for: calculating a correlation parameter between the temperature, the heating parameter and the environmental parameter; Based on the RBF network, a temperature compensation model corresponding to the temperature, the heating parameter and the environmental parameter is constructed; respectively inputting the temperature, the heating parameter and the environmental parameter into the corresponding temperature compensation model, and outputting an initial compensation temperature; Calculating a secondary compensation temperature according to the correlation parameter and initial compensation temperatures corresponding to the temperature, the heating parameter and the environmental parameter respectively; Inputting the secondary compensation temperature into a time series prediction model to obtain the compensation temperature; Based on the decision forest algorithm, constructing a melting point determination model corresponding to the temperature parameter and the compensation temperature respectively; Input the temperature parameter and the compensation temperature into the corresponding melting point determination model, and output the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature; The melting point is calculated according to the predicted value corresponding to the temperature parameter and the predicted value corresponding to the compensation temperature.

10. The high-precision melting point detector according to claim 9, characterized in that: The high-precision melting point detector also includes a network module, which is used to transmit at least one of the image information, the temperature, the temperature parameter, the environmental parameter, the calculated compensation temperature and the melting point.

Images