Fish freshness prediction method and device and storage medium

Abstract

The invention provides a fish freshness prediction method and device and a storage medium, and belongs to the technical field of quality detection of agricultural and animal products. The fish freshness prediction method comprises the following steps: determining a target fish part, a target constant temperature value and a target duration; respectively inputting the target fish meat part, the target constant temperature value and the target duration into a fish meat freshness prediction model to obtain a freshness index value output by the fish meat freshness prediction model; wherein the fish freshness prediction model is obtained by training a preset sample fish part, a constant sample temperature value, a preset sample duration and a freshness sample index value. The method is suitable for application scenes of large-scale and rapid detection of the freshness of the fish meat.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural and livestock product quality detection, and in particular to a fish meat freshness prediction method, device and storage medium. Background Art

[0002] Fish meat has high water and protein content, which makes it very susceptible to microorganisms and enzymes during storage, and it becomes stale or even spoiled. Fish meat freshness assessment methods generally rely on sensory evaluation, chemical analysis and microbial testing. Although these methods can provide accurate freshness indicators, they often require professional equipment, are time-consuming and costly. my country's annual output of freshwater fish is very high, so how to detect the freshness of fish meat on a large scale and quickly is an important issue that the industry needs to solve urgently. Summary of the invention

[0003] In view of the problems existing in the prior art, the present invention provides a method, device and storage medium for predicting the freshness of fish meat.

[0004] The present invention provides a method for predicting the freshness of fish meat, comprising: determining a target fish meat part, a target constant temperature value and a target duration; inputting the target fish meat part, the target constant temperature value and the target duration into a fish meat freshness prediction model respectively, and obtaining a freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by using a preset sample fish meat part, a constant sample temperature value, a preset sample duration and a freshness sample index value.

[0005] According to a fish meat freshness prediction method provided by the present invention, the fish meat freshness prediction model includes: an input unit, a processing unit and an output unit; the training method of the fish meat freshness prediction model includes: inputting the sample fish meat part, the constant sample temperature value, and the preset sample time length into the input unit, and processing them through the processing unit and the output unit in sequence, and outputting the freshness sample prediction value; comparing the freshness sample prediction value with the freshness sample index value, and adjusting the parameters of the input unit, the processing unit and the output unit according to the comparison result, until the difference between the freshness sample prediction value and the freshness sample index value is less than a threshold value.

[0006] According to a method for predicting the freshness of fish meat provided by the present invention, the freshness sample index value at least includes a total viable bacteria count sample index value, a volatile basic nitrogen content sample index value, a freshness K value sample index value, a sensory evaluation total score sample index value and a sensory evaluation overall acceptability index value; the processing unit includes: a fish meat total viable bacteria count prediction subunit, a fish meat volatile basic nitrogen content prediction subunit, a fish meat freshness K value prediction subunit, a fish meat sensory evaluation total score prediction subunit and a fish meat sensory evaluation overall acceptability prediction subunit; the output unit includes : a fish meat total viable bacteria count output subunit, a fish meat volatile basic nitrogen content output subunit, a fish meat freshness K value output subunit, a fish meat sensory evaluation total score output subunit and a fish meat sensory evaluation overall acceptability output subunit; the sample fish meat part, the constant sample temperature value, and the preset sample time are input into the input unit, and are processed by the processing unit and the output unit in sequence to output a freshness sample prediction value, including: the sample fish meat part, the constant sample temperature value, and the preset sample time are input into the input unit, and are processed by the fish meat total viable bacteria count prediction subunit and The total viable bacteria count output subunit of the fish meat processes and outputs the sample prediction value of the total viable bacteria count; the sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit in sequence, and the sample prediction value of the fish meat volatile basic nitrogen content is output; the sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit in sequence, and the fish meat freshness K value is output. The sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit in sequence, and the sample prediction value of the fish meat sensory evaluation total score is output; the sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit in sequence, and the sample prediction value of the fish meat sensory evaluation overall acceptability is output.

[0007] According to a fish meat freshness prediction method provided by the present invention, the freshness sample prediction value is compared with the freshness sample index value, and the parameters of the input unit, the processing unit and the output unit are adjusted according to the comparison result, including: comparing the total viable bacteria count sample prediction value with the total viable bacteria count sample index value, and adjusting the parameters of the input unit, the fish meat total viable bacteria count prediction subunit and the fish meat total viable bacteria count output subunit according to the comparison result; comparing the volatile basic nitrogen content sample prediction value with the volatile basic nitrogen content sample index value, and adjusting the parameters of the input unit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit according to the comparison result; comparing the freshness K value sample prediction value with the volatile basic nitrogen content sample index value, and adjusting the parameters of the input unit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit according to the comparison result; comparing the freshness K value sample prediction value with the total viable bacteria count sample index value, and adjusting the parameters of the input unit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit according to the comparison result. The predicted value is compared with the sample index value of the freshness K value, and the parameters of the input unit, the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit are adjusted according to the comparison result; the sample predicted value of the total score of the sensory evaluation is compared with the sample index value of the total score of the sensory evaluation, and the parameters of the input unit, the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit are adjusted according to the comparison result; the sample predicted value of the overall acceptability of the sensory evaluation is compared with the sample index value of the overall acceptability of the sensory evaluation, and the parameters of the input unit, the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit are adjusted according to the comparison result.

[0008] According to a fish meat freshness prediction method provided by the present invention, the freshness index value at least includes a total viable bacteria count index value, a volatile basic nitrogen content index value, a freshness K value index value, a sensory evaluation total score index value and a sensory evaluation overall acceptability index value; the target fish meat part, the target constant temperature value and the target time are respectively input into the fish meat freshness prediction model to obtain the freshness index value output by the fish meat freshness prediction model, including: inputting the target fish meat part, the target constant temperature value and the target time into the input unit, and processing them through the fish meat total viable bacteria count prediction subunit and the fish meat total viable bacteria count output subunit to obtain the total viable bacteria count index value; inputting the target fish meat part, the target constant temperature value and the target time into the input unit, and processing them through the fish meat volatile basic nitrogen content prediction subunit The target fish meat part, target constant temperature value and target time are input into the input unit, and processed by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit to obtain the freshness K value; the target fish meat part, target constant temperature value and target time are input into the input unit, and processed by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit to obtain the sensory evaluation total score; the target fish meat part, target constant temperature value and target time are input into the input unit, and processed by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit to obtain the sensory evaluation overall acceptability.

[0009] The present invention also provides a fish meat freshness prediction device, comprising: an input acquisition module, used to determine a target fish meat part, a target constant temperature value and a target duration; a freshness index value output module, used to input the target fish meat part, the target constant temperature value and the target duration into a fish meat freshness prediction model respectively, and obtain the freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by using a preset sample fish meat part, a constant sample temperature value, a preset sample duration and a freshness sample index value.

[0010] The present invention also provides an electronic device, comprising a storage device, a processor, and a computer program stored in the storage device and executable on the processor, wherein when the processor executes the program, any of the above-mentioned methods for predicting the freshness of fish meat is implemented.

[0011] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for predicting the freshness of fish meat as described above is implemented.

[0012] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for predicting the freshness of fish meat.

[0013] The fish meat freshness prediction method, device and storage medium provided by the present invention can directly obtain the freshness index value of the target fish meat part through the target fish meat part, target constant temperature value and target duration, and can detect the freshness of fish meat on a large scale and quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0015] Figure 1 It is a flow chart of the fish meat freshness prediction method provided by an embodiment of the present invention.

[0016] Figure 2 It is a schematic diagram of the operation of the fish freshness prediction model provided by an embodiment of the present invention.

[0017] Figure 3 This is one of the structural schematic diagrams of the fish freshness prediction model provided by the embodiment of the present invention.

[0018] Figure 4 The second structural schematic diagram of the fish freshness prediction model provided in the embodiment of the present invention.

[0019] Figure 5A It is a schematic diagram of the sample index value of the total colony count of grass carp back meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0020] Figure 5B It is a schematic diagram of the sample index value of the total colony count of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0021] Figure 5C It is a schematic diagram of the sample index value of the total colony count of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0022] Figure 5D It is a schematic diagram of the sample index value of the total colony count of grass carp tail meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0023] Fig. 6A It is a schematic diagram of sample index values ​​of total sensory evaluation of grass carp back meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0024] Figure 6BIt is a schematic diagram of sample index values ​​of total sensory evaluation scores of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0025] Figure 6C It is a schematic diagram of sample index values ​​of the total sensory evaluation score of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0026] Fig.6D It is a schematic diagram of sample index values ​​of total sensory evaluation of grass carp tail meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0027] Fig. 7A It is a schematic diagram of the sample index values ​​of the overall acceptability of the sensory evaluation of grass carp back meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0028] Figure 7B It is a schematic diagram of the sample index values ​​of the overall acceptability of the sensory evaluation of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0029] Figure 7C It is a schematic diagram of the sample index values ​​of the overall acceptability of the sensory evaluation of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0030] Fig.7D It is a schematic diagram of the sample index values ​​of the overall acceptability of the sensory evaluation of grass carp tail meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0031] Fig. 8A It is a schematic diagram of the sample index value of the TVN-B content of grass carp back meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0032] Figure 8B It is a schematic diagram of the sample index value of the TVN-B content of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0033] Figure 8C It is a schematic diagram of the sample index value of the TVN-B content of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0034] Fig.8D It is a schematic diagram of the sample index value of the TVN-B content of grass carp tail meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0035] Fig.9AIt is a schematic diagram of the sample index value of the freshness K value of grass carp back meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0036] Fig. 9B It is a schematic diagram of the sample index value of the freshness K value of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0037] Fig. 9C It is a schematic diagram of the sample index value of the freshness K value of grass carp belly meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0038] Fig.9D It is a schematic diagram of the sample index value of the freshness K value of grass carp tail meat after storing a preset sample time at a constant sample temperature value provided by an embodiment of the present invention.

[0039] Fig.10 It is a module schematic diagram of a fish freshness prediction device provided by an embodiment of the present invention.

[0040] Fig.11 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention.

[0041] Reference numerals: 201: input acquisition module; 202: freshness index value output module; 310: processor; 320: communication interface; 330: memory; 340: communication bus. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Combine the following Figure 1-Figure 11 The invention describes a method, device and storage medium for predicting the freshness of fish meat.

[0044] Fish meat is easy to deteriorate because it is rich in protein and water. Eating spoiled fish meat will endanger human health. Therefore, how to detect the freshness of fish meat so that the fish meat is in a good state of freshness when it is sold and consumed is a problem that needs to be solved in the fish meat sales process. The existing methods for detecting the freshness of fish meat generally use chemical analysis, microbial detection, etc., and it is necessary to go through steps such as sampling and experimental operation to obtain the test results of the freshness of fish meat, and the detection efficiency is relatively low. However, my country's fish production is relatively large, and the sales volume of fish meat is relatively large, so it is necessary to improve the detection efficiency of the freshness of fish meat. By storing fish meat from different parts at different constant temperatures, measuring the changes in the freshness sample index values ​​during storage at different constant temperatures for different times, establishing a reliable, high-accuracy, and freshness prediction method suitable for different fish meat parts has become an urgent problem to be solved.

[0045] To this end, an embodiment of the present invention provides a method for predicting the freshness of fish meat. Figure 1 A schematic diagram of a method for predicting fish freshness provided by an embodiment of the present invention is shown in FIG. Figure 1 As shown, the method includes: Step 101, determine the target fish meat part, target constant temperature value and target duration.

[0046] Specifically, the target fish meat part is a small portion product obtained by cutting different parts of the whole fish. Exemplarily, the target fish meat part may include back fish meat, belly fish meat, belly fish meat, and tail fish meat, etc. For example, when grass carp is cut and stored, the belly of the grass carp is the target fish meat part.

[0047] The target constant temperature value refers to keeping the storage temperature of the target fish meat part unchanged. Exemplarily, the target constant temperature value environment can be obtained by refrigeration, or by slight freezing. When the target constant temperature value environment is obtained by refrigeration, the target fish meat part can be placed in a freezer. For example, the target constant temperature value can be 0-12 degrees Celsius.

[0048] The target duration refers to the storage duration of the target fish meat part at the target constant temperature value. Exemplarily, the target duration can be the general duration from obtaining the target fish meat part to selling the target fish meat part. For example, the target duration can be 0-10 days.

[0049] Step 102: input the target fish meat part, target constant temperature value and target time into the fish meat freshness prediction model respectively to obtain the freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by preset sample fish meat part, constant sample temperature value, preset sample time and freshness sample index value.

[0050] Specifically, the freshness index value is used to reflect the quality and freshness of the fish. For example, the freshness grade of the fish can be determined according to the freshness index value, and fish with different freshness grades have different qualities and freshness.

[0051] Before executing step 102, a fish meat freshness prediction model can also be pre-trained. Specifically, the fish meat freshness prediction model can be trained in the following manner: first, a large number of sample fish meat parts, constant sample temperature values, sample durations, and freshness sample index values ​​are collected, and multiple groups of sample fish meat parts, constant sample temperature values, sample durations, and freshness sample index values ​​are obtained through manual annotation. Each group of sample fish meat parts, constant sample temperature values, sample durations, and freshness sample index values ​​have a corresponding relationship. Then, each group of sample fish meat parts, constant sample temperature values, sample durations, and freshness sample index values ​​are input into the initial model for training, thereby obtaining a fish meat freshness prediction model.

[0052] The fish meat freshness prediction method provided by the embodiment of the present invention directly obtains the freshness index value of the target fish meat part through the target fish meat part, the target constant temperature value and the target duration, and can detect the freshness of fish meat on a large scale and quickly.

[0053] Compared with using chemical analysis and microbial testing to obtain the freshness index value of the target fish meat part, using the fish meat freshness prediction model to obtain the freshness index value of the target fish meat part does not require sampling of the target fish meat part, and can obtain the freshness index value of the target fish meat part without loss, without affecting the sales of the target fish meat part.

[0054] There are differences in the content of ingredients in different fish parts. Therefore, when the fish freshness prediction model is obtained by pre-training with sample fish parts of specific parts, there is a large deviation when predicting the freshness of target fish parts of other parts. For example, the fat content of the belly and belly of grass carp is higher than that of the back and tail, while the protein content of the belly and belly of grass carp is lower than that of the back and tail. The protein content will affect the deterioration rate of fish parts. Therefore, when the fish freshness prediction model is obtained by pre-training with the back of grass carp, there will be a large deviation when predicting the freshness index value of the belly of grass carp. Compared with the fish freshness prediction model obtained by pre-training with the target fish parts of specific parts, the fish freshness prediction model obtained by pre-training with a large number of sample fish parts of different parts can reduce the deviation of the fish freshness prediction model in predicting the freshness index values ​​of target fish parts of different parts, and improve the accuracy of predicting the freshness index value of the target fish parts.

[0055] Based on the above embodiment, the fish meat freshness prediction model includes: an input unit, a processing unit and an output unit. The training method of the fish meat freshness prediction model includes: Step 201: input the sample fish meat part, constant sample temperature value, and preset sample duration into an input unit, and process them in sequence through a processing unit and an output unit to output a freshness sample prediction value.

[0056] Specifically, the input unit is the part of the fish freshness prediction model that receives raw data, the processing unit is the part of the fish freshness prediction model that performs data processing and feature extraction, and the output unit is the part of the fish freshness prediction model that generates a predicted freshness index value.

[0057] It is understandable that at least the input sample fish meat part, constant sample temperature value, and preset sample duration need to be processed for data standardization so as to convert the input sample fish meat part, constant sample temperature value, and preset sample duration into the data format required by the processing unit. Figure 2 and Figure 3 As shown, data normalization processing can be performed by a separate data normalization unit, or can be performed in an input unit or a processing unit.

[0058] Step 202: compare the freshness sample prediction value with the freshness sample index value, and adjust the parameters of the input unit, the processing unit, and the output unit according to the comparison result until the difference between the freshness sample prediction value and the freshness sample index value is less than a threshold value.

[0059] The difference between the freshness sample prediction value and the freshness sample index value may be the difference between the freshness sample prediction value and the freshness sample index value, or the ratio between the freshness sample prediction value and the freshness sample index value. The threshold may be a preset tolerance range, and an exemplary threshold may be -20% to +20%.

[0060] In the case where the difference between the freshness sample prediction value and the freshness sample index value is a ratio between the two, the difference between the freshness sample prediction value and the freshness sample index value can be obtained by the following formula as an example.

[0061]

[0062] Among them, RE is the difference between the freshness sample prediction value and the freshness sample index value; PV is the freshness sample prediction value, and EV is the freshness sample index value.

[0063] The fish freshness prediction method provided by the embodiment of the present invention can evaluate the performance of the model in the current state and quantify the prediction deviation of the model in the current state by comparing the freshness sample prediction value with the freshness sample index value. By making the difference between the freshness sample prediction value and the freshness sample index value less than a threshold, the prediction deviation of the trained fish freshness prediction model can be controlled, thereby improving the accuracy of predicting the freshness index value of the target fish meat part.

[0064] After step 102, the method further includes inputting the test sample fish meat part, the test constant sample temperature value, and the test preset sample time into the input unit, and processing them through the processing unit and the output unit in sequence, and outputting the test freshness sample prediction value; comparing the test freshness sample prediction value with the test freshness sample index value, and adjusting the parameters of the input unit, the processing unit, and the output unit according to the comparison result until the difference between the test freshness sample prediction value and the test freshness sample index value is less than a threshold value.

[0065] Among them, at least the test constant sample temperature value does not overlap with the range of the aforementioned constant sample temperature value. In this way, by comparing the test freshness sample prediction value and the test freshness sample index value at a test constant sample temperature value that is at least different from the constant sample temperature value, the performance of the model in the current state can be better evaluated, and the prediction deviation of the model in the current state can be quantified, so as to improve the accuracy of predicting the freshness index value of the target fish meat part.

[0066] like Figure 4 As shown, based on any of the above embodiments, the freshness sample index value includes at least a total viable count (TVC) sample index value, a volatile base nitrogen (TVB-N) content sample index value, a freshness K value sample index value, a sensory evaluation total score sample index value, and a sensory evaluation overall acceptability index value.

[0067] Specifically, regarding the total viable count sample index value, illustratively, 5 grams of sample fish meat can be weighed in a sterile clean bench, the sample fish meat can be placed in a sterile homogenizing bag or a sterile sampling bag, and a sterile saline solution with a capacity of 45 ml and a mass concentration of 0.85% (W / V) can be added to the bag. The sample fish meat in the bag is patted for 15 seconds to fully mix and homogenize the sample fish meat with the sterile saline solution. Obtain the homogenate produced by the sample fish meat, select 2 to 3 appropriate dilution gradients according to the growth of sample microorganisms, and use sterile physiological saline to dilute the homogenate 10 times. Pipette 100 microliters of dilution solution and add it to the surface of the plate count agar and spread it evenly, and spread two plates for each dilution. The coated plates are aerobically cultured in a 30-degree Celsius incubator for 48 hours. Select plates with a colony count between 30 and 300 to obtain the total viable count of the sample fish meat. The colony count unit is log CFU / g, and a beating homogenizer can be used to beat the sample fish meat in the bag.

[0068] Regarding the sample index value of volatile basic nitrogen content, for example, 2.5 grams of sample fish meat can be weighed, the minced sample fish meat is mixed with 25 ml of distilled water in a centrifuge tube, homogenized for 1 minute, oscillated at low speed for 30 minutes, and centrifuged at 4000 rpm for 5 minutes. 5 ml of supernatant is measured and mixed with a magnesium oxide solution with a concentration of 10 grams per liter in equal proportion, the mixed solution is distilled for 5 minutes using a Kjeldahl nitrogen analyzer, and the distillate is absorbed in a conical flask containing 10 ml of a boric acid solution with a concentration of 20 grams per liter and 100 microliters of a methyl red-methylene blue indicator with a concentration of 2 grams per liter. The absorption liquid is titrated with a hydrochloric acid standard solution with a concentration of 0.02 mol per liter until the absorption liquid turns blue-purple, and the amount of hydrochloric acid standard solution is recorded. Distilled water is used instead of the supernatant as a blank control. The sample index value of volatile basic nitrogen content of the sample fish meat part is obtained.

[0069] Regarding the freshness K value sample index value, for example, 1 gram of sample fish meat can be weighed and homogenized with 2 ml of a 10% (W / W) perchloric acid solution, and then centrifuged at 5000 rpm for 3 minutes. The supernatant is collected in a centrifuge tube, 2 ml of a 5% (W / W) PCA solution is added to the precipitate and mixed, and the centrifugation is continued at 5000 rpm for 3 minutes, repeated twice, and the supernatant is combined. The pH of the supernatant is adjusted to 6.4±0.05 with 10 mol / L and mol / L KOH, and the supernatant is collected after centrifugation at 5000 rpm for 3 minutes. The volume is made up to 10 ml with a neutralized PCA solution to obtain an ATP-associated substance extract. The ATP-associated substance extract and the ATP-associated substance standard are filtered with a 0.22 micron water ultrafiltration membrane and sampled and measured on a high performance liquid chromatograph (HPLC). ATP-associated substances in the solution were monitored using HPLC equipped with a COSMOSIL 5C18-PAQ liquid chromatography column (4.6 ID×250 mm×5 μm). The chromatographic conditions were as follows: detection wavelength 254 nm; column temperature 30 degrees Celsius; injection volume 50 μL; mobile phase A was 0.05 mol / L Na2HPO4-NaH2PO4 buffer (pH 6.8); mobile phase B was chromatographically pure methanol; flow rate was milliliters per minute. The content of each associated substance in the sample was calculated based on the standard curve of the peak area and concentration of the standard. The freshness K value can be calculated by the following formula.

[0070]

[0071] Among them, H X Hypoxanthine, H X R is inosine, ATP is adenosine triphosphate (ATP), ADP is adenosine diphosphate (ADP), AMP is adenosine monophosphate (AMP), and IMP is inosine monophosphate (IMP).

[0072] The freshness K value is obtained by measuring the content changes of ATP and its decomposition products in the fish meat parts of the sample. For example, the fish meat parts with a freshness K value of less than 20% are fresh, and the fish meat parts with a freshness K value of more than 60% are not fresh.

[0073] Regarding the total sensory evaluation score sample index value and the overall acceptability index value of the sensory evaluation, for example, 6 women and 6 men who have undergone sensory training can give the total sensory evaluation score sample index value and the overall acceptability index value of the sensory evaluation based on the texture, color, smell and tissue morphology of the sample fish meat parts.

[0074] For example, the sensory evaluation scoring rules for the fish meat parts of the samples can be referred to Table 1.

[0075] Table 1 Sensory evaluation scoring rules for fish meat parts of samples

[0076] The processing unit includes: a fish meat total viable bacteria count prediction subunit, a fish meat volatile basic nitrogen content prediction subunit, a fish meat freshness K value prediction subunit, a fish meat sensory evaluation total score prediction subunit and a fish meat sensory evaluation overall acceptability prediction subunit.

[0077] The output units include: a total live bacteria count output subunit for fish meat, a volatile basic nitrogen content output subunit for fish meat, a fish meat freshness K value output subunit, a fish meat sensory evaluation total score output subunit, and a fish meat sensory evaluation overall acceptability output subunit.

[0078] The sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the processing unit and the output unit in sequence to output the freshness sample prediction value, including: the sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat total live bacteria count prediction subunit and the fish meat total live bacteria count output subunit in sequence to output the total live bacteria count sample prediction value.

[0079] The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed in turn by the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit to output the sample prediction value of the fish meat volatile basic nitrogen content.

[0080] The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed in turn by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit to output the fish meat freshness K value sample prediction value.

[0081] The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed in turn by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit to output the sample prediction value of the fish meat sensory evaluation total score.

[0082] The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed in turn by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit to output the sample prediction value of the fish meat sensory evaluation overall acceptability.

[0083] The sample fish meat part, constant sample temperature value, and preset sample time are the same input values; the total live bacteria count sample index value, volatile basic nitrogen content sample index value, freshness K value sample index value, sensory evaluation total score sample index value and sensory evaluation overall acceptability index value are different freshness sample index values; using the same sample fish meat part, constant sample temperature value, preset sample time value, and different freshness sample index values ​​to train a fish meat freshness prediction model with a single processing unit requires overcoming conflicts in the process of predicting different freshness sample index values, which will increase the difficulty of training the fish meat freshness prediction model and the complexity of the trained fish meat freshness prediction model.

[0084] The fish freshness prediction method provided by the embodiment of the present invention predicts the corresponding freshness sample index values ​​of the sample fish meat parts, constant sample temperature values, and preset sample time through different freshness sample index value prediction sub-units, which can reduce the training difficulty of the fish meat freshness prediction model and the complexity of the trained fish meat freshness prediction model.

[0085] Furthermore, a fish freshness prediction model with a single processing unit trained using the same sample fish part, constant sample temperature value, preset sample duration, and different freshness sample index values ​​may reduce the accuracy of the fish freshness prediction model in predicting a single freshness sample prediction value due to the balance of different freshness sample prediction values, so as to ensure the accuracy of the fish freshness prediction model in comprehensively predicting the prediction value of each freshness sample. The fish freshness prediction method provided by the embodiment of the present invention can avoid the above-mentioned problem.

[0086] like Figure 3 As shown, based on any of the above embodiments, the freshness sample prediction value is compared with the freshness sample index value, and the parameters of the input unit, the processing unit and the output unit are adjusted according to the comparison results, including: comparing the total live bacteria count sample prediction value with the total live bacteria count sample index value, and adjusting the parameters of the input unit, the fish total live bacteria count prediction subunit and the fish total live bacteria count output subunit according to the comparison results.

[0087] The predicted values ​​of the volatile basic nitrogen content samples are compared with the index values ​​of the volatile basic nitrogen content samples, and the parameters of the input unit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit are adjusted according to the comparison results.

[0088] The freshness K value sample prediction value is compared with the freshness K value sample index value, and the parameters of the input unit, the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit are adjusted according to the comparison results.

[0089] The predicted value of the total sensory evaluation score sample is compared with the index value of the total sensory evaluation score sample, and the parameters of the input unit, the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit are adjusted according to the comparison results.

[0090] The sample prediction value of the overall acceptability of sensory evaluation was compared with the sample index value of the overall acceptability of sensory evaluation, and the parameters of the input unit, the prediction subunit of the overall acceptability of fish meat sensory evaluation and the output subunit of the overall acceptability of fish meat sensory evaluation were adjusted according to the comparison results.

[0091] The method of comparing the total live bacteria count sample predicted value with the total live bacteria count sample index value to obtain the difference between the total live bacteria count sample predicted value and the total live bacteria count sample index value is basically the same as the aforementioned method of comparing the freshness sample predicted value with the freshness sample index value to obtain the difference between the freshness sample predicted value and the freshness sample index value, and will not be repeated here.

[0092] The fish meat freshness prediction method provided by the embodiment of the present invention, the total live bacteria count sample prediction value and the total live bacteria count sample index value can evaluate the performance of the fish meat total live bacteria count prediction subunit and the fish meat total live bacteria count output subunit in the current state, quantify the prediction deviation of the fish meat total live bacteria count prediction subunit in the current state, and adjust the parameters of the fish meat total live bacteria count prediction subunit and the fish meat total live bacteria count output subunit through the prediction deviation, so that the difference between the total live bacteria count sample prediction value and the total live bacteria count sample index value is less than a threshold value, thereby improving the accuracy of predicting the total live bacteria count of the target fish meat part.

[0093] The technical effects of the working principles of the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit, the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit, the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit are basically the same as those of the fish meat total viable bacteria count prediction subunit and the fish meat total viable bacteria count output subunit, and will not be repeated here.

[0094] Based on any of the above embodiments, the freshness index value includes at least a total viable bacteria count index value, a volatile basic nitrogen content index value, a freshness K value index value, a sensory evaluation total score index value, and a sensory evaluation overall acceptability index value.

[0095] The target fish meat part, target constant temperature value and target time are respectively input into the fish meat freshness prediction model to obtain the freshness index value output by the fish meat freshness prediction model, including: inputting the target fish meat part, target constant temperature value and target time into the input unit, and processing them through the fish meat total live bacteria count prediction subunit and the fish meat total live bacteria count output subunit to obtain the total live bacteria count index value.

[0096] The target fish meat part, target constant temperature value and target duration are input into the input unit, and processed by the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit to obtain the volatile basic nitrogen content index value.

[0097] The target fish meat part, target constant temperature value and target duration are input into the input unit, and processed by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit to obtain the freshness K value.

[0098] The target fish meat part, target constant temperature value and target duration are input into the input unit, and processed by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit to obtain the sensory evaluation total score.

[0099] The target fish meat part, target constant temperature value and target duration are input into the input unit, and processed by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit to obtain the overall acceptability of the sensory evaluation.

[0100] The fish meat freshness prediction method provided by the embodiment of the present invention directly obtains the various freshness index values ​​of the target fish meat part through the target fish meat part, the target constant temperature value and the target duration, and can detect the freshness of the fish meat on a large scale and quickly.

[0101] Compared to using a fish meat freshness prediction model with a single processing unit to predict different freshness index values, the fish meat freshness prediction provided by the embodiment of the present invention uses a fish meat freshness prediction model with a processing unit with multiple different sub-units. Different sub-units are used to predict corresponding freshness index values ​​respectively, which can obtain better performance and improve the accuracy of predicting each freshness index value of the target fish meat part.

[0102] In order to specifically illustrate the fish meat freshness prediction method provided by this embodiment, a specific example is provided below. Before training the prediction model of freshness of different target fish meat of grass carp, it is necessary to first obtain the sample fish meat parts, constant sample temperature value, preset sample duration and freshness sample prediction value. Among them, the sample fish meat parts are samples corresponding to different parts of grass carp fish meat.

[0103] Based on the sample fish meat part, constant sample temperature value, preset sample duration and freshness sample prediction value, the fish meat freshness prediction model can be trained through support vector regression (SVR) based on the principle of support vector machine (SVM). Among them, support vector regression predicts continuous variables with the smallest possible error by finding an optimal function. At the same time, when the sample size data is small, SVR can still maintain good prediction performance and is suitable for small data sets. In addition, the SVR model has good prediction performance for indicators such as the total sensory evaluation score, total viable count (TVC), volatile basic nitrogen (TVB-N), freshness K value, pH value and overall acceptability index value of sensory evaluation during the storage of aquatic products.

[0104] Specifically, 96 fresh grass carps weighing 2.01-2.07 kg and 53.8-58 cm in length from the aquatic market were transported to the laboratory in oxygenated transport bags. After the fish were slightly stable, the grass carps were stunned by hitting their heads, and then scaled, cleaned, and washed. Subsequently, the grass carps were decapitated and cut into four parts: back, abdomen, belly, and tail, to obtain sample fish meat parts.

[0105] 0 degrees Celsius, 3 degrees Celsius, 6 degrees Celsius, 9 degrees Celsius, and 12 degrees Celsius are set as constant sample temperature values. Exemplarily, the constant temperature environment condition can be obtained by 5 freezers, and during the constant temperature storage process, each sample fish meat part of the grass carp is placed separately in a storage drawer.

[0106] At 0 degrees Celsius, 0 days, 2 days, 4 days, 6 days, 8 days and 10 days can be set as the preset sample duration; at 3 degrees Celsius, 0 days, 1 day, 2 days, 4 days, 6 days and 8 days can be set as the preset sample duration; at 6 degrees Celsius, 0 days, 1 day, 2 days, 3 days, 4 days and 5 days can be set as the preset sample duration; at 9 degrees Celsius, 0 days, 1 day, 2 days, 3 days, 4 days and 5 days can be set as the preset sample duration; at 12 degrees Celsius, 0 days, 0.5 days, 1 day, 1.5 days, 2 days, 2.5 days, 3 days and 4 days can be set as the preset sample duration.

[0107] After storing the sample fish meat parts for a preset sample time at a constant sample temperature value, the total viable count, total sensory evaluation score, overall sensory evaluation acceptability, volatile basic nitrogen content and freshness K value of grass carp back meat, grass carp belly meat, grass carp belly meat and grass carp tail meat are respectively detected to obtain the freshness index data of grass carp meat at different parts at a constant storage temperature, as shown in Figures 5 to 9. The detection methods of the total viable count, total sensory evaluation score, overall sensory evaluation acceptability, volatile basic nitrogen content and freshness K value can be referred to the aforementioned embodiments, which will not be repeated here.

[0108] Specifically, as shown in Figure 5, when stored at 0 to 12 degrees Celsius, the initial TVC of each sample fish part is 4.03 to 5.06 log CFU / g, and with the extension of storage time, it shows a continuous growth trend. Compared with 0 degrees Celsius, it can be clearly seen that the growth rate of microorganisms at 12 degrees Celsius is significantly accelerated. The higher the temperature, the faster the growth rate of microorganisms. When the TVC in the sample fish part exceeds 7.0 log CFU / g, it means that the number of microorganisms has exceeded the standard, and the sample fish part is corrupt and deteriorated and cannot be eaten.

[0109] As shown in Figures 6 and 7, the total initial sensory evaluation score of each fish part of the sample was 89.00-96.70, indicating that the sensory quality of each fish part of the sample was excellent before storage. During storage, the total sensory evaluation score of the fish parts of the sample decreased with the increase of storage temperature, and the higher the storage temperature, the faster the total sensory evaluation score decreased. When the total sensory evaluation score is less than 40, it means that the average scores of the texture, color, smell, tissue morphology and overall acceptability of the fish parts of the sample are low, and the edible acceptability is low; when the total sensory evaluation score is less than 20, the sensory state of the fish parts of the sample is extremely poor, and the edible acceptability is extremely low, that is, the fish is inedible. According to the sensory evaluation table of the fish parts of the sample, when the overall acceptability score is less than 8 points, it is considered that the overall sensory of the fish parts of the sample is poor and the edible acceptability is low; when the overall acceptability score is less than 4 points, it is considered that the overall sensory of the fish parts of the sample is extremely poor and the edible acceptability is extremely low. The overall acceptability of the initial sensory evaluation of each group was 17.30-19.30, indicating that the overall sensory perception of the fish meat parts of the samples was good and the edible acceptability was high before storage began. During storage, the overall sensory acceptability of each group decreased rapidly with the increase of storage temperature.

[0110] As shown in Figure 8, as the storage temperature increases, the TVB-N content increases faster, and the order and severity of the spoilage of the fish meat parts of the samples are different at different temperatures. The belly meat of grass carp is the most severely spoiled when stored at 0 degrees Celsius.

[0111] As shown in Figure 9, the changes in the freshness K value of the sample fish meat parts when stored at 0 to 12 degrees Celsius are shown in Figures 5-6. The initial freshness K value of the sample fish meat parts is between 15.65% and 19.99%, and the freshness is relatively high. With the increase in storage time, the freshness K value of each group increases significantly. After 10 days of storage at 0 degrees Celsius, the freshness K value of grass carp belly meat, grass carp belly meat and grass carp tail meat all exceeded 60%, which were 63.56%, 60.55% and 61.91% respectively. When stored at 3 degrees Celsius, the freshness K value of grass carp belly meat, grass carp belly meat and grass carp tail meat exceeded the limit after 6 days of storage, and after 8 days of storage, the freshness K value of grass carp belly meat, grass carp belly meat and grass carp tail meat reached 66.05%, 68.32% and 65.00% respectively. When stored at 6 degrees Celsius, the freshness K value of grass carp belly meat and grass carp belly meat exceeded the limit after 4 days of storage, and then the freshness K value of grass carp tail meat exceeded the limit after 5 days of storage. When stored at 9 degrees Celsius, the freshness K value of grass carp tail meat and grass carp belly meat exceeded the limit after 4 days of storage, and after 5 days of storage, the freshness K value of grass carp back meat, grass carp belly meat, grass carp belly meat and grass carp tail meat all exceeded the limit, which were 70.55%, 70.94%, 62.84% and 63.94% respectively. When stored at 12 degrees Celsius, the freshness K value of grass carp tail meat exceeded the limit after 2.5 days (60 hours). After 3 days (72 hours), the freshness K values ​​of grass carp back meat, grass carp belly meat and grass carp belly meat all exceeded the limit. On the 4th day (96 hours) of storage, the freshness K values ​​of the sample fish meat parts of grass carp back meat, grass carp belly meat, grass carp belly meat and grass carp tail meat were 70.06%, 71.02%, 68.48% and 69.53% respectively.

[0112] The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and each prediction subunit and output subunit outputs the corresponding freshness index value in turn. The parameters of the input unit, each prediction subunit and output subunit are adjusted according to the comparison results until the difference between each freshness sample prediction value and the corresponding freshness sample index value is less than a threshold value.

[0113] Exemplarily, the constant sample temperature value is 4 degrees Celsius. As shown in Table 2, the REs predicted by TVC are all between -10% and 10%, indicating that the constructed SVR model has a high prediction accuracy for the TVC of the target fish meat part and can accurately predict the total colony count of the target fish meat part during the storage period of 0 to 12 degrees Celsius.

[0114] Table 2 Total number of colonies during storage at a constant sample temperature of 4 degrees Celsius

[0115] As shown in Tables 3 and 4, the REs of the total sensory evaluation score and the overall sensory acceptability prediction were both between -10% and 10%, indicating that the constructed SVR model can accurately predict the total sensory evaluation score and overall acceptability of the target fish meat parts during the storage period of 0 to 12 degrees Celsius.

[0116] Table 3 Total sensory evaluation scores during storage at a constant sample temperature of 4 degrees Celsius

[0117] Table 4 Overall acceptability of sensory evaluation during storage at a constant sample temperature of 4 degrees Celsius

[0118] As shown in Table 5, the REs of TVB-N content prediction were all between -20% and 20%, of which 33.33% of TVB-N content had REs between -10% and 10%, indicating that the constructed SVR model had a high prediction accuracy for the TVB-N content of the target fish meat parts.

[0119] Table 5 TVB-N content of samples during storage at a constant temperature of 4 degrees Celsius

[0120] As shown in Table 6, the REs of the freshness K value predictions are all between -20% and 20%, of which 62.50% of the freshness K values ​​have REs between -10% and 10%, indicating that the constructed SVR model can accurately predict the freshness K value of the target fish meat parts during the storage period of 0 to 12 degrees Celsius.

[0121] Table 6 Freshness K value during storage at a constant sample temperature of 4 degrees Celsius

[0122] In the process of training the prediction model of the freshness of different target parts of grass carp through support vector regression (SVR) based on the principle of support vector machine (SVM), necessary libraries can be imported first, such as pandas for data processing and analysis, the function train_test_split for dividing data into training sets and test sets, and OneHotEncoder for one-hot encoding of categorical variables (such as storage parts). Among them, pandas can facilitate reading and manipulating data, such as loading Excel files including preset sample fish parts, constant sample temperature values, preset sample durations, and freshness sample index values. train_test_split can facilitate the training and evaluation of the model. OneHotEncoder can easily convert the input into a numerical format suitable for processing by machine learning algorithms. For example, when initializing OneHotEncoder, setting the sparse parameter to False can return a dense array. The encoder function encoder.fit_transform can be used to one-hot encode the parts column (storage parts) in all data and convert it to a numerical format. Let parts_encoded = encoder.fit_transform (storage parts); parts_encoded is a one-hot encoded numerical format. For example, if there are three types of sample fish meat, back, abdomen, and tail, they will be converted into three separate columns, each corresponding to one part. Convert the encoded array to a pandas data frame for subsequent processing. The storage part can also be called the type of sample fish meat.

[0123] You can use the pandas.concat function to merge the one-hot encoded part features with other numerical features, such as storage temperature and storage time, to form a complete feature set. For example, features_encoded = pd.concat (storage part, storage temperature, storage time)

[0124] You can also import more function libraries, such as SVR, mean_squared_error, and sqrt. SVR supports vector regression models and can predict continuous values; mean_squared_error can evaluate the mean square error of the model's prediction results; sqrt can calculate the root mean square error (RMSE), and the root mean square error can evaluate the prediction accuracy.

[0125] The support vector regression model can be trained and evaluated using the train_evaluate_svr function, splitting the features and target variables into a training set (80%) and a test set (20%). The SVR model is initialized and trained using the radial basis kernel function (rbf). Predictions are made on the test set, and the root mean square error (RMSE) is calculated, which is the square root of the error between the predicted value and the actual value, which is used to measure the performance of the model.

[0126] The targets list contains multiple freshness indicators, and an empty dictionary models can be initialized to store the trained model and evaluation results. The following operations can be performed on each target variable: filter out rows containing missing target values; extract corresponding features and target variables; call the train_evaluate_svr function for training and evaluation; store the trained model and RMSE in the models dictionary.

[0127] The trained model can be used through the predict_new_data function to predict the target fish meat part, target constant temperature value and target duration, and output the corresponding freshness index value.

[0128] The method for predicting the freshness of grass carp meat at different target parts provided by the embodiment of the present invention supports vector regression training of a prediction model for the freshness of grass carp meat at different target parts, and can process high-dimensional data and nonlinear relationships, so that the trained prediction model for the freshness of grass carp meat at different target parts provides a high-precision freshness prediction result; the deterioration speed and pattern of grass carp meat at different parts under storage conditions are different, and the prediction model for the freshness of grass carp meat at different target parts can perform detailed freshness prediction for meat at different parts of the grass carp body, rather than just for back meat or the whole fish, thereby increasing the accuracy of the prediction results and providing a comprehensive evaluation of the quality of grass carp meat at different parts.

[0129] The fish freshness prediction device provided by the present invention is described below. The fish freshness prediction device described below and the fish freshness prediction method described above can be referenced to each other.

[0130] Fig.10 Schematic diagram of the module of the device for predicting the freshness of fish meat provided by the present invention. Fig.10 The device shown includes an input acquisition module 201, which is used to determine the target fish meat part, the target constant temperature value and the target time; a freshness index value output module 202, which is used to input the target fish meat part, the target constant temperature value and the target time into the fish meat freshness prediction model respectively, and obtain the freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by preset sample fish meat part, constant sample temperature value, preset sample time and freshness sample index value.

[0131] The fish meat freshness prediction device provided by the present invention directly obtains the freshness index value of the target fish meat part through the target fish meat part, the target constant temperature value and the target duration, and can detect the freshness of the fish meat on a large scale and quickly.

[0132] It can be understood that the fish freshness prediction device provided by the present invention corresponds to the fish freshness prediction device provided by the above-mentioned embodiments. The relevant technical features of the fish freshness prediction device provided by the present invention can refer to the relevant technical features of the fish freshness prediction device provided by the above-mentioned embodiments, which will not be repeated here.

[0133] Fig.11 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.11 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the logic instructions in the memory 330 to execute the fish freshness prediction method, which includes: determining the target fish meat part, the target constant temperature value and the target duration; inputting the target fish meat part, the target constant temperature value and the target duration into the fish meat freshness prediction model respectively, and obtaining the freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by the preset sample fish meat part, the constant sample temperature value, the preset sample duration and the freshness sample index value.

[0134] In addition, the logic instructions in the above-mentioned storage 330 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0135] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the fish freshness prediction method provided by the above-mentioned methods, which includes: determining the target fish meat part, target constant temperature value and target time; inputting the target fish meat part, target constant temperature value and target time into a fish meat freshness prediction model respectively, and obtaining the freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by preset sample fish meat part, constant sample temperature value, preset sample time and freshness sample index value.

[0136] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the fish freshness prediction method provided by the above-mentioned methods, the method comprising: determining a target fish meat part, a target constant temperature value and a target duration; inputting the target fish meat part, the target constant temperature value and the target duration into a fish meat freshness prediction model respectively, and obtaining a freshness index value output by the fish meat freshness prediction model; wherein the fish meat freshness prediction model is trained by using a preset sample fish meat part, a constant sample temperature value, a preset sample duration and a freshness sample index value.

[0137] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0138] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting fish freshness, characterized in that: include: Determine the target fish meat part, target constant temperature value and target duration; Inputting the target fish meat part, the target constant temperature value and the target duration into the fish meat freshness prediction model respectively, and obtaining the freshness index value output by the fish meat freshness prediction model; The fish meat freshness prediction model is obtained by training with preset sample fish meat parts, constant sample temperature values, preset sample durations and freshness sample index values.

2. The method for predicting fish freshness according to claim 1, characterized in that: The fish meat freshness prediction model comprises: an input unit, a processing unit and an output unit; The training method of the fish freshness prediction model comprises: The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and processed by the processing unit and the output unit in sequence to output the freshness sample prediction value; The freshness sample prediction value is compared with the freshness sample index value, and the parameters of the input unit, the processing unit and the output unit are adjusted according to the comparison result until the difference between the freshness sample prediction value and the freshness sample index value is less than a threshold value.

3. The method for predicting fish freshness according to claim 2, characterized in that: The freshness sample index value at least includes the total viable bacteria sample index value, the volatile basic nitrogen content sample index value, the freshness K value sample index value, the sensory evaluation total score sample index value and the sensory evaluation overall acceptability index value; The processing unit includes: a fish meat total viable bacteria count prediction subunit, a fish meat volatile basic nitrogen content prediction subunit, a fish meat freshness K value prediction subunit, a fish meat sensory evaluation total score prediction subunit, and a fish meat sensory evaluation overall acceptability prediction subunit; The output unit includes: a total viable bacteria count output subunit of fish meat, a volatile basic nitrogen content output subunit of fish meat, a fish meat freshness K value output subunit, a fish meat sensory evaluation total score output subunit, and a fish meat sensory evaluation overall acceptability output subunit; The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed by the processing unit and the output unit in sequence to output the freshness sample prediction value, including: The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and processed by the fish meat total viable bacteria count prediction subunit and the fish meat total viable bacteria count output subunit in sequence, and the total viable bacteria count sample prediction value is output; The sample fish meat part, constant sample temperature value, and preset sample duration are input into an input unit, and processed by a fish meat volatile basic nitrogen content prediction subunit and a fish meat volatile basic nitrogen content output subunit in sequence, and a sample prediction value of fish meat volatile basic nitrogen content is output; The sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit in sequence, and the fish meat freshness K value sample prediction value is output; the sample fish meat part, constant sample temperature value, and preset sample time are input into the input unit, and are processed by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit in sequence, and the fish meat sensory evaluation total score sample prediction value is output; The sample fish meat part, constant sample temperature value, and preset sample duration are input into the input unit, and are processed in sequence by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit to output a sample prediction value of the fish meat sensory evaluation overall acceptability.

4. The method for predicting fish freshness according to claim 3, characterized in that: The freshness sample prediction value is compared with the freshness sample index value, and the parameters of the input unit, the processing unit and the output unit are adjusted according to the comparison result, including: Compare the total viable bacteria count sample prediction value with the total viable bacteria count sample index value, and adjust the parameters of the input unit, the fish meat total viable bacteria count prediction subunit, and the fish meat total viable bacteria count output subunit according to the comparison result; Comparing the sample prediction value of volatile basic nitrogen content with the sample index value of volatile basic nitrogen content, and adjusting the parameters of the input unit, the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit according to the comparison result; Compare the freshness K value sample prediction value with the freshness K value sample index value, and adjust the parameters of the input unit, the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit according to the comparison result; The sample prediction value of the total score of the sensory evaluation is compared with the sample index value of the total score of the sensory evaluation, and the parameters of the input unit, the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit are adjusted according to the comparison results; the sample prediction value of the overall acceptability of the sensory evaluation is compared with the sample index value of the overall acceptability of the sensory evaluation, and the parameters of the input unit, the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit are adjusted according to the comparison results.

5. The method for predicting fish freshness according to claim 3, characterized in that: The freshness index value at least includes the total viable bacteria index value, the volatile basic nitrogen content index value, the freshness K value index value, the sensory evaluation total score index value and the sensory evaluation overall acceptability index value; Inputting the target fish meat part, the target constant temperature value and the target duration into the fish meat freshness prediction model respectively, and obtaining the freshness index value output by the fish meat freshness prediction model, including: The target fish meat part, the target constant temperature value and the target duration are input into the input unit, and processed by the fish meat total viable bacteria count prediction subunit and the fish meat total viable bacteria count output subunit to obtain the total viable bacteria count index value; The target fish meat part, the target constant temperature value and the target duration are input into the input unit, and processed by the fish meat volatile basic nitrogen content prediction subunit and the fish meat volatile basic nitrogen content output subunit to obtain the volatile basic nitrogen content index value; the target fish meat part, the target constant temperature value and the target duration are input into the input unit, and processed by the fish meat freshness K value prediction subunit and the fish meat freshness K value output subunit to obtain the freshness K value; The target fish meat part, the target constant temperature value and the target duration are input into the input unit, and processed by the fish meat sensory evaluation total score prediction subunit and the fish meat sensory evaluation total score output subunit to obtain the sensory evaluation total score; The target fish meat part, target constant temperature value and target duration are input into the input unit, and processed by the fish meat sensory evaluation overall acceptability prediction subunit and the fish meat sensory evaluation overall acceptability output subunit to obtain the overall acceptability of the sensory evaluation.

6. A device for predicting the freshness of fish, characterized in that: include: An input acquisition module is used to determine a target fish meat part, a target constant temperature value, and a target duration; A freshness index value output module is used to input the target fish meat part, the target constant temperature value and the target duration into the fish meat freshness prediction model to obtain the freshness index value output by the fish meat freshness prediction model; The fish meat freshness prediction model is obtained by training with preset sample fish meat parts, constant sample temperature values, preset sample durations and freshness sample index values.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the fish freshness prediction method according to any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting fish freshness according to any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting fish freshness according to any one of claims 1 to 5 is implemented.

Images