Poultry pre-cooling water intelligent purification method based on AI

Abstract

The invention discloses an AI-based poultry pre-cooling water intelligent purification method. The method comprises the following steps: S1, determining a specified slaughtering line as an information data collection site in the poultry slaughtering industry; s2, preprocessing the collected data, and extracting feature information after preprocessing is completed; and S3, according to the extracted feature data and through an AI learning algorithm, establishing an AI deep learning model, processing the extracted image data through a convolutional neural network, and then carrying out multiple times of test training. The cameras are additionally arranged at multiple positions of the pre-cooling water pool and serve as a main means for information input, the change value of each independent digital display sensor is recognized in real time through images, and the situation that measurement is inaccurate due to the fact that a traditional sensor is prone to being polluted by oil stains is avoided. The change conditions of different scenes can be intelligently identified, for example, the grating is blocked by sundries and needs to be cleaned; when water is turbid, the power of a water pump of the filter needs to be increased, corresponding operation is automatically performed, and an intelligent water purification and cooling system is formed, so that water and energy are saved.

Description

Technical Field

[0001] The present invention relates to the technical field of poultry slaughtering, and in particular to an AI-based intelligent purification method for poultry precooling water. Background Art

[0002] The slaughtering industry is a big consumer of energy and water. Water is essential as a medium for cleaning and cooling. Data show that a processing plant with a slaughter capacity of 10,000 chickens per hour consumes more than 1,000 tons of water every day. In order to reduce bacteria and protect the quality of the slaughtered carcasses, the water temperature must be maintained between 0 and 4 degrees. For this reason, the daily power consumption is more than 10,000 degrees. In addition, in order to prevent the blood from being too thick and causing the carcasses to be colored and affect sales, the blood must be discharged after a period of time, which wastes a lot of energy. What's more troublesome is that the blood is difficult to handle. It will produce a foul odor after being left for a few days, which will seriously pollute the environment. Therefore, People naturally think of using circulating water to reduce losses. The easiest way to think of it is to use various filter materials. However, the transmittance and durability of the filter materials in the past have limited its application. There are many impurities, grease, organic matter and microorganisms in the water, and the clogging problem of the filter materials has been difficult to solve. In addition, the sensor is often contaminated, resulting in inaccurate measurements and failures. The sensor needs to use 485 or Bluetooth to connect and transmit, which also has many inconveniences. The cleaning of the coarse filter grid accumulation still needs to be managed by people. The operation of the refrigerator cannot achieve multi-information evidence linkage, and the effect of water saving and energy saving is not ideal. Summary of the invention

[0003] The purpose of the present invention is to provide an AI-based intelligent purification method for poultry pre-cooling water to solve the problems raised in the above-mentioned background technology.

[0004] To achieve the above purpose, the present invention provides the following technical solution: an AI-based intelligent purification method for poultry precooling water, comprising the following contents:

[0005] S1: For the poultry slaughtering industry, designated slaughtering lines are identified as sites for information data collection;

[0006] S2: preprocess the collected data and extract feature information after preprocessing;

[0007] S3: Based on the extracted feature data and AI learning algorithm, an AI deep learning model is established, and the extracted image data is processed through a convolutional neural network, followed by multiple test trainings;

[0008] S4: Install low-temperature solidified ceramic filter materials on the slaughtering line site, filter the used water through the low-temperature solidified ceramic filter materials, obtain the filtered image of the ceramic filter materials through the camera, and then AI artificial intelligence will regularly feedback information after recognition and specify a cleaning plan;

[0009] S5: Install special green LED lights and use the camera to distinguish the number of lights to estimate the turbidity of blood water.

[0010] S6: Conduct coordination tests through the above multiple filtering, cleaning and identification functions, and then collect and process the functional test results;

[0011] S7: After determining the usage results, formulate AI continuous learning training based on the results and upgrade and optimize the functions.

[0012] Preferably, the S1 comprises the following steps:

[0013] S11: Determine the test site selection requirements for all slaughter lines in the poultry slaughter industry;

[0014] S12: Select the test site for information data collection according to the requirements.

[0015] Preferably, S2 comprises the following steps:

[0016] S21: After completing the collection of the filtration and cleaning information data of the pre-cooling water used in the daily production of the test site, pre-process the data;

[0017] S22: Clean, denoise and mark key points of the collected data;

[0018] S23: Data cleaning can remove duplicate, missing, and erroneous data to ensure data integrity;

[0019] S24: Denoising data is the process of eliminating noise and outliers in data by using algorithms and techniques to improve data accuracy;

[0020] S25: By annotating the data, the key points of the collected data can be labeled and classified to facilitate clear identification by subsequent models.

[0021] Preferably, S3 comprises the following steps:

[0022] S31: After completing the corresponding data preprocessing work, establish an AI deep learning model through the AI ​​learning algorithm;

[0023] S32: Processing the extracted image data through a convolutional neural network;

[0024] S33: Mapping the image data features extracted by the convolutional neural network to the recurrent neural network to extract appropriate feature data;

[0025] S34: Then, by extracting appropriate feature data, reduce redundant information, enhance the generalization ability of the model, and then use the AI ​​deep learning model to perform multiple test training tasks;

[0026] S35: Through the coordination of algorithms and models, multiple test training tasks are carried out, and internal parameters are continuously adjusted to reduce prediction errors.

[0027] Preferably, S4 comprises the following steps:

[0028] S41: Install low-temperature solidified ceramic filter material at the pre-cooling tank of the poultry slaughtering test site, and filter the blood water used by the low-temperature solidified ceramic filter material;

[0029] S42: installing a matching camera at a location where the low-temperature-cured ceramic filter material is installed, and photographing the filtering effect of the low-temperature-cured ceramic filter material through the camera;

[0030] S43: Feedback the image video captured by the camera to the AI ​​artificial intelligence and the established AI deep learning model, identify the fed-back image data information through the AI ​​artificial intelligence and the established AI deep learning model, and formulate a reasonable cleaning plan based on the historical cleaning plan as a reference.

[0031] Preferably, S5 comprises the following steps:

[0032] S51: In addition to the designated functions, green LED row lights are installed at the sewage discharge point after being filtered by low-temperature solidified ceramic filter materials at the location where the camera is installed;

[0033] S52: The LED lights are then photographed through a camera, and the amount of hemoglobin in the water is estimated based on the light transmittance of the water. The information can then be fed back to the AI ​​artificial intelligence and the established AI deep learning model.

[0034] Preferably, S6 comprises the following steps:

[0035] S61: After completing the feedback collection of data information through the above identification work, the specific information of the pollutants is fed back through the image data obtained by the camera to determine the pollutant information;

[0036] S62: Afterwards, cleaning instructions are formulated through AI artificial intelligence, and the dirt in the low-temperature solidified ceramic filter material is cleaned through the cleaning structure;

[0037] S63: After completing the cleanup work, collect and organize the overall work data.

[0038] Preferably, the S7 comprises the following steps:

[0039] S71: Determine the filtering and cleaning effect based on the data information obtained above;

[0040] S72: Determine the filtering and cleaning defects manually and by AI artificial intelligence according to the determination results, then perform learning and training work through AI artificial intelligence and AI deep learning models according to the filtering and cleaning defects, and upgrade and optimize the internal system functions;

[0041] S73: After completing the upgrade and optimization, conduct multiple test runs.

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows: a new type of low-temperature solidified ceramic filter material is installed in the poultry slaughter pre-cooling pool. The new type of low-temperature solidified ceramic filter material has a more accurate and consistent pore size, is not easily adhered by other substances, is more convenient to clean the inner hole, has a lower cost, is easy to be made into various shapes, has good mechanical strength, and is an ideal filter material. Its filtration rate and self-cleaning rate are higher than those of previous filter materials, and is suitable for long-term circulation in pre-cooled blood water. A camera is installed at the position where the low-temperature solidified ceramic filter material is installed. The camera can obtain the filtering dirt situation of the low-temperature solidified ceramic filter material, and then identify and feedback the information that needs to be cleaned, and then formulate a cleaning plan in time. After continuous learning, optimization and upgrading through AI artificial intelligence and AI deep learning models, the pre-cooling work of the slaughter line is guaranteed to prevent affecting the normal operation of the slaughter line. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the process of the present invention;

[0044] Figure 2 Schematic diagram of the structure of the system of the present invention.

[0045] Among them, 1. water guide trough 2. water flow direction 3. water flow rate display 4. water turbidity sensor 5. temperature digital display 6. grille filter 7. camera 8. low-temperature ceramic water purifier 9. cold water exchanger 10. refrigerator 11. refrigeration pipe 12. water baffle 13. underwater conveyor belt 14. poultry carcass 15. low-temperature cleaning water 16. conveyor belt direction 17. water guide trough DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0047] See also Figure 1The present invention provides a technical solution: an AI-based intelligent purification method for poultry precooling water, comprising the following contents:

[0048] S1: For the poultry slaughtering industry, designated slaughtering lines are identified as sites for information data collection;

[0049] S2: preprocess the collected data and extract feature information after preprocessing;

[0050] S3: Based on the extracted feature data and AI learning algorithm, an AI deep learning model is established, and the extracted image data is processed through a convolutional neural network, followed by multiple test trainings;

[0051] S4: Install low-temperature solidified ceramic filter materials on the slaughtering line site, filter the used water through the low-temperature solidified ceramic filter materials, obtain the filtered image of the ceramic filter materials through the camera, and then AI artificial intelligence will regularly feedback information after recognition and specify a cleaning plan;

[0052] S5: Install special green LED lights and use the camera to distinguish the number of lights to estimate the turbidity of blood in the water;

[0053] S6: Conduct coordination tests through the above multiple filtering, cleaning and identification functions, and then collect and process the functional test results;

[0054] S7: After determining the usage results, formulate AI continuous learning training based on the results and upgrade and optimize the functions.

[0055] Specifically, clean water at room temperature is input from the water inlet pipe, and is cold exchanged with the output cold wastewater through the cold exchanger, so that the temperature of the input clean water drops to a certain extent, so as to reduce the water temperature and achieve the effect of low energy consumption. After that, the water passes through the refrigerator and the refrigeration pipe to make the water temperature reach the range of 0 to 4 degrees, and it is required to ensure that the number of microorganisms in the water body and the hemoglobin concentration in the water do not exceed the requirements of the production technical indicators, otherwise it will affect the quality of the product and the appearance of sales. In order to meet the technical requirements of the product and reduce the power consumption cost as much as possible, this system uses an advanced low-temperature curing ceramic filter, which can effectively reduce bacteria and reduce the hemoglobin concentration in the water, reduce the amount of water used, and the filter has a self-cleaning function. When the ceramic micropores are blocked to a certain extent, the spare components can be enabled to work, and the function of the filter can be restored by backwashing and other methods.

[0056] The biggest feature of this system is that it uses artificial intelligence for supervision and uses cameras as the main means of obtaining information. The camera is used to collect real-time images and analyze data in a specific area. For example, parameters such as temperature, water flow, and pressure are directly displayed on digital instruments. The camera is used to obtain data on the instrument, eliminating the wiring trouble of the sensor using 485 transmission in the past, or Bluetooth transmission interference and other problems.

[0057] The hemoglobin concentration information in the water is obtained by a row of green LED lights inserted into the water, with the camera reading the number of distinguishable lights. This not only avoids the problem of previous sensors being contaminated by oil, but also greatly reduces the production and maintenance costs.

[0058] The water pressure, flow rate, etc. in the water purifier 8 also use cameras to collect data on the digital display instrument, and then the AI ​​analysis and decision-making control system works. The grille filter 6 also uses cameras to obtain image information. Artificial intelligence can analyze and identify the accumulation of pollutants, and automatically start the cleaning system when there are too many.

[0059] Wherein, the S1 comprises the following steps:

[0060] S11: Determine the test site selection requirements for all slaughter lines in the poultry slaughter industry;

[0061] S12: Select the test site for information data collection according to the requirements.

[0062] Wherein, the S2 comprises the following steps:

[0063] S21: After completing the collection of the filtration and cleaning information data of the pre-cooling water used in the daily production of the test site, pre-process the data;

[0064] S22: Clean, denoise and mark key points of the collected data;

[0065] S23: Data cleaning can remove duplicate, missing, and erroneous data to ensure data integrity;

[0066] S24: Denoising data is the process of eliminating noise and outliers in data by using algorithms and techniques to improve data accuracy;

[0067] S25: By annotating the data, the key points of the collected data can be labeled and classified to facilitate clear identification by subsequent models.

[0068] Wherein, the S3 comprises the following steps:

[0069] S31: After completing the corresponding data preprocessing work, establish an AI deep learning model through the AI ​​learning algorithm;

[0070] S32: Processing the extracted image data through a convolutional neural network;

[0071] S33: Mapping the image data features extracted by the convolutional neural network to the recurrent neural network to extract appropriate feature data;

[0072] S34: Then, by extracting appropriate feature data, reduce redundant information, enhance the generalization ability of the model, and then use the AI ​​deep learning model to perform multiple test training tasks;

[0073] S35: Through the coordination of algorithms and models, multiple test training tasks are carried out, and internal parameters are continuously adjusted to reduce prediction errors.

[0074] Wherein, the S4 comprises the following steps:

[0075] S41: Install low-temperature solidified ceramic filter material at the pre-cooling tank of the poultry slaughtering test site, and filter the blood water used by the low-temperature solidified ceramic filter material;

[0076] S42: installing a matching camera at a location where the low-temperature-cured ceramic filter material is installed, and photographing the filtering effect of the low-temperature-cured ceramic filter material through the camera;

[0077] S43: Feedback the image video captured by the camera to the AI ​​artificial intelligence and the established AI deep learning model, identify the fed-back image data information through the AI ​​artificial intelligence and the established AI deep learning model, and formulate a reasonable cleaning plan based on the historical cleaning plan as a reference.

[0078] Wherein, the S5 comprises the following steps:

[0079] S51: In addition to the designated functions, green LED row lights are installed at the sewage discharge point after being filtered by low-temperature solidified ceramic filter materials at the location where the camera is installed;

[0080] S52: Then, the LED lights are photographed through a camera, the number of lights is identified according to the light transmittance of the water, and the amount of hemoglobin in the water is estimated to roughly estimate the light transmittance of the water. The information can then be fed back to the AI ​​artificial intelligence and the established AI deep learning model.

[0081] Wherein, the S6 comprises the following steps:

[0082] S61: After completing the feedback collection of data information through the above identification work, the specific information of the pollutants is fed back through the image data obtained by the camera to determine the pollutant information;

[0083] S62: Afterwards, cleaning instructions are formulated through AI artificial intelligence, and the dirt in the low-temperature solidified ceramic filter material is cleaned through the cleaning structure;

[0084] S63: After completing the cleanup work, collect and organize the overall work data.

[0085] Wherein, the S7 comprises the following steps:

[0086] S71: Determine the filtering and cleaning effect based on the data information obtained above;

[0087] S72: Determine the filtering and cleaning defects manually and by AI artificial intelligence according to the determination results, then perform learning and training work through AI artificial intelligence and AI deep learning models according to the filtering and cleaning defects, and upgrade and optimize the internal system functions;

[0088] S73: After completing the upgrade and optimization, conduct multiple test runs.

[0089] Specifically, when the present invention is used, the test site selection requirements are determined for all slaughtering lines in the poultry slaughtering industry, and the information data collection test site is selected according to the requirements. After the information data of the filtration and cleaning work of pre-cooling water used in the daily production of the test site is collected, the data is pre-processed, and the collected data is cleaned, denoised, and key-marked. By cleaning the data, duplicate, accurate, and erroneous data can be removed to ensure the integrity of the data. By denoising the data, the noise and outliers in the data are eliminated by using algorithms and technologies to improve the accuracy of the data. By marking the data, the key points of the collected data can be labeled and classified to facilitate clear identification by subsequent models. After preprocessing the data, an AI deep learning model is established through the AI ​​learning algorithm. The image data is processed through a convolutional neural network, and the image data features extracted through the convolutional neural network are mapped to the recurrent neural network to extract appropriate feature data. Then, through the extracted appropriate feature data, redundant information is reduced and the generalization ability of the model is enhanced. Then, the AI ​​deep learning model is used for multiple test training. Multiple test training is carried out through the coordination of the algorithm and the model, and the internal parameters are continuously adjusted to reduce prediction errors. Low-temperature cured ceramic filter materials are installed in the pre-cooling pool of the slaughtering industry test site, and the discharged sewage is filtered through the low-temperature cured ceramic filter materials. Filtration treatment, and install a matching camera at the location where the low-temperature solidified ceramic filter material is installed, and use the camera to shoot the filtration treatment effect of the low-temperature solidified ceramic filter material, and feed back the image video taken by the camera to AI artificial intelligence and the established AI deep learning model, and use AI artificial intelligence and the established AI deep learning model to identify the feedback image data information, and then formulate a reasonable cleaning plan based on the historical cleaning plan as a reference. In addition to the above filtering and cleaning plan formulation function, special color LED lights, such as green and other colors, are installed at the sewage discharged from the sewage after filtration by the low-temperature solidified ceramic filter material at the installed camera location, and then the special color LED lights are photographed by the camera, and The lamp tubes of LED lights with special colors are identified, and the light intensity is read by the camera to roughly estimate the light transmittance of the water. The information can then be fed back to the AI ​​artificial intelligence and the established AI deep learning model. After the feedback collection of data information is completed through the above identification work, the specific information of the pollutants is fed back through the image data obtained by the camera to determine the pollutant information. After that, the cleaning instructions are formulated through the AI ​​artificial intelligence, and the dirt in the low-temperature solidified ceramic filter material is cleaned through the cleaning structure. After the cleaning work is completed, the overall work data is collected and sorted, and the filtering and cleaning effect is determined based on the data information obtained above. According to the judgment results, the filtering and cleaning defects are determined manually and by AI artificial intelligence.Then, according to the filtering and cleaning defects, AI artificial intelligence and AI deep learning models are used for learning and training, and the internal system functions are upgraded and optimized. After the upgrade and optimization are completed, multiple experimental tests are carried out.

[0090] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An AI-based intelligent purification method for poultry precooling water, characterized in that: Includes the following: S1: For the poultry slaughtering industry, designated slaughtering lines are identified as sites for information data collection; S2: preprocess the collected data and extract feature information after preprocessing; S3: Based on the extracted feature data and AI learning algorithm, an AI deep learning model is established, and the extracted image data is processed through a convolutional neural network, followed by multiple test trainings; S4: Install low-temperature solidified ceramic filter materials on the slaughtering line site, filter the used water through the low-temperature solidified ceramic filter materials, obtain the filtered image of the ceramic filter materials through the camera, and then AI artificial intelligence will regularly feedback information after recognition and specify a cleaning plan; S5: Install special green LED lights and use the camera to distinguish the number of lights to estimate the turbidity of blood in the water; S6: Conduct coordination tests through the above multiple filtering, cleaning and identification functions, and then collect and process the functional test results; S7: After determining the usage results, formulate AI continuous learning training based on the results and upgrade and optimize the functions.

2. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S1 comprises the following steps: S11: Determine the test site selection requirements for all slaughter lines in the poultry slaughter industry; S12: Select the test site for information data collection according to the requirements.

3. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S2 comprises the following steps: S21: After completing the collection of the filtration and cleaning information data of the pre-cooling water used in the daily production of the test site, pre-process the data; S22: Clean, denoise and mark key points of the collected data; S23: Data cleaning can remove duplicate, missing, and erroneous data to ensure data integrity; S24: Denoising data is the process of eliminating noise and outliers in data by using algorithms and techniques to improve data accuracy; S25: By annotating the data, the key points of the collected data can be labeled and classified to facilitate clear identification by subsequent models.

4. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S3 comprises the following steps: S31: After completing the corresponding data preprocessing work, establish an AI deep learning model through the AI ​​learning algorithm; S32: Processing the extracted image data through a convolutional neural network; S33: Mapping the image data features extracted by the convolutional neural network to the recurrent neural network to extract appropriate feature data; S34: Then, by extracting appropriate feature data, reduce redundant information, enhance the generalization ability of the model, and then use the AI ​​deep learning model to perform multiple test training tasks; S35: Through the coordination of algorithms and models, multiple test training tasks are carried out, and internal parameters are continuously adjusted to reduce prediction errors.

5. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S4 comprises the following steps: S41: Install low-temperature solidified ceramic filter materials at the pre-cooling tank of the poultry slaughtering test site, and use the low-temperature solidified ceramic filter materials to filter the discharged wastewater; S42: installing a matching camera at a location where the low-temperature-cured ceramic filter material is installed, and photographing the filtering effect of the low-temperature-cured ceramic filter material through the camera; S43: Feedback the image video captured by the camera to the AI ​​artificial intelligence and the established AI deep learning model, identify the fed-back image data information through the AI ​​artificial intelligence and the established AI deep learning model, and formulate a reasonable cleaning plan based on the historical cleaning plan as a reference.

6. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S5 comprises the following steps: S51: In addition to the above filtering and cleaning plan formulation function, green LED row lights are installed at the sewage discharge point after the sewage is filtered by the low-temperature solidified ceramic filter material at the location where the camera is installed; S52: Then, the LED lights are photographed by a camera, the number of lights is identified according to the light transmittance of water, and the amount of hemoglobin in the water is estimated.

7. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S6 comprises the following steps: S61: After completing the feedback collection of data information through the above identification work, the specific information of the pollutants is fed back through the image data obtained by the camera to determine the pollutant information; S62: Afterwards, cleaning instructions are formulated through AI artificial intelligence, and the dirt in the low-temperature solidified ceramic filter material is cleaned through the cleaning structure; S63: After completing the cleanup work, collect and organize the overall work data.

8. The AI-based intelligent purification method for poultry precooling water according to claim 1, characterized in that: The S7 comprises the following steps: S71: Determine the filtering and cleaning effect based on the data information obtained above; S72: Determine the filtering and cleaning defects manually and by AI artificial intelligence according to the determination results, then perform learning and training work through AI artificial intelligence and AI deep learning models according to the filtering and cleaning defects, and upgrade and optimize the internal system functions; S73: After completing the upgrade and optimization, conduct multiple test runs.

Images