Rice flour processing quality monitoring method and system based on artificial intelligence

Abstract

The invention discloses a rice flour processing quality monitoring method and system based on artificial intelligence, and relates to the technical field of artificial intelligence, and the method comprises the steps: building a three-dimensional model of a rice flour processing factory, obtaining historical processing records of processing equipment, and obtaining a target record in the processing records; acquiring a historical monitoring picture of a factory, obtaining a local area and a rice flour area in the historical monitoring picture, analyzing the local area and the rice flour area, and extracting all feature time periods in the monitoring picture; obtaining an equipment area corresponding to the processing equipment, and obtaining a first management and control degree and a second management and control degree corresponding to the monitoring picture according to the characteristic time period; and obtaining the total management and control degree of each monitoring picture, obtaining an early warning coefficient according to the picture display duration of the monitoring picture, and judging whether to carry out early warning prompt on the picture display duration or not. According to the invention, the historical monitoring picture and the processing equipment in the factory are analyzed, so that consumers can comprehensively understand the processing process, and the quality monitoring reliability and credibility of rice noodle processing are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and specifically to a method and system for monitoring the processing quality of rice noodles based on artificial intelligence. Background Technique

[0002] As a special snack in southern China, rice noodles are widely favored by many diners. In order to ensure the purity and safety of rice noodles and enhance the brand reputation, it is crucial to strictly monitor the entire process of rice noodle processing; currently, rice noodle processing factories generally conduct visual traceability of the entire production process of making rice noodles. There are usually multiple monitoring devices in the factory. Consumers can scan the traceability code to watch the entire process of the processing process in the rice noodle processing factory in real time.

[0003] Currently, when consumers view the real-time rice noodle processing process, they usually display each monitoring screen in turn according to the display duration of each monitoring device preset manually. However, manual setting is subjective, and the display duration is not reasonably set according to the historical processing information in each monitoring screen, which will lead to insufficient exposure of key processes. Consumers cannot fully understand the processing process, resulting in a decrease in the trust in the product. Therefore, intelligent analysis is needed to give early warning prompts for unreasonable display durations in a timely manner. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for monitoring the processing quality of rice noodles based on artificial intelligence to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] A method for monitoring the processing quality of rice noodles based on artificial intelligence includes the following steps:

[0007] Step S100: Establish a three-dimensional model of the rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the processing time corresponding to the processing records, and analyze the operating state parameters of the processing equipment at each moment during the processing time to obtain the target records in the processing records;

[0008] Step S200: Obtain the historical monitoring screens of the factory, mark the local areas of moving objects in the monitoring screens, and take the areas passed by the rice noodles during the processing as the rice noodle areas. By analyzing the contact situation between the local areas and the rice noodle areas in the monitoring screens, all characteristic time periods in the monitoring screens are extracted;

[0009] Step S300: Obtain the equipment area corresponding to the processing equipment in the monitoring screen. According to the target records corresponding to the processing equipment, obtain the target degree corresponding to each equipment area, and obtain the first control degree corresponding to the monitoring screen based on the target degree; obtain the second control degree corresponding to the monitoring screen according to the characteristic time period corresponding to the monitoring screen.

[0010] Step S400: Obtain the total control degree of each monitoring screen according to the first control degree and the second control degree; according to the current traceability system, the display duration of each monitoring screen shown in turn, and the total control degree, obtain the warning coefficient of the current display duration of the screen, and determine whether to give a warning prompt for the current display duration of the screen according to the warning coefficient.

[0011] Further, step S100 includes:

[0012] Step S110: Obtain a number of processing records corresponding to each processing equipment in history. The processing records are records with a processing duration of D during the normal processing process after the processing equipment is started up; obtain the processing time T and the processing equipment M corresponding to a certain processing record C, and extract the operating state parameters of the processing equipment M at each moment within the processing time T. The operating state parameters include temperature value, feed rate, and discharge rate. Among them, the temperature value is obtained by a temperature sensor deployed at a certain local position of the processing equipment M, and the feed rate and discharge rate are obtained by weight measuring equipment deployed at the material inlet and outlet of the processing equipment M.

[0013] Step S120: According to the temperature value, feed rate, and discharge rate at each moment within the processing time T, calculate the variance corresponding to all temperature values as the first target value, calculate the variance corresponding to all feed rates as the second target value, and calculate the variance corresponding to all discharge rates as the third target value; respectively set the weight values of the temperature value, feed rate, and discharge rate, and obtain the total target value of the processing record C according to the first, second, and third target values. If the total target value is less than the preset target threshold, regard the processing record C as the target information, and then extract all target records in the processing record.

[0014] It should be noted that after the equipment is started up and running, there will be a stage where the operating parameters gradually reach a stable state. Generally, this stage is called the "stable operation stage". However, since the time required for the stable operation state of each equipment is different, it is necessary to analyze the historical processing information of each equipment to obtain the moment when each equipment reaches stability; and obtaining the moment when the equipment reaches stability provides accurate and reliable data for the calculation of the following warning coefficient, making the following data calculation results more reliable and convincing.

[0015] Further, step S200 includes:

[0016] Step S210: Obtain the historical monitoring images obtained by the monitoring devices in the factory. Through target detection technology, obtain the local area R1 corresponding to a certain moving object in the monitoring image, and a certain rice noodle area R2 that the rice noodles pass through during the processing. Among them, the local area R1 is a moving rectangular area, and the rice noodle area R2 is a fixed area. If there is a moment T1 when the intersection area between the local area R1 and the rice noodle area R2 in the monitoring image is not 0, and the intersection area between the local area R1 and the rice noodle area R2 at the previous moment of moment T1 is 0, then take moment T1 as the just-contact moment;

[0017] And extract the monitoring segment F1 with a duration of D1 starting from moment T1. And take the moment corresponding to the maximum intersection area between the local area R1 and the rice noodle area R2 in the monitoring segment F1 as T2. Take the area extended by a length of L outward from the edge of the local area R1 as R3. And take the area where the area R2 intersects with the area R3 at moment T2 as R4. The area R4 is a fixed area;

[0018] Step S220: Perform edge detection on the area R4 in the monitoring image at moments T1 and T2 respectively to obtain the edge images P1 and P2. Randomly extract several edge points in the edge image P1 as reference points, and mark the corresponding positions of each reference point in the image P2. Obtain the edge point in the image P2 that is closest to a certain reference point n, and calculate the distance value between the two points as the target distance of the reference point n. Then, based on the target distances of all reference points, calculate the variance as the deformation degree of the area R4. If the deformation degree is greater than the preset degree threshold, then take the time period between moment T1 and moment T3 as the characteristic time period, and then obtain all the characteristic time periods in the monitoring image.

[0019] In this solution, the moving area is the staff, and the local area is the hand area. To ensure the production quality, it is necessary to strictly supervise the behavior of each staff member in the factory. One of the main things is that the staff's hands come into contact with the rice noodles. For the quality and hygiene of rice noodle processing, the behavior of the staff during direct contact with food should be supervised. Then, direct contact means that the rice noodles have deformed. And in this solution, the corresponding judgment is made based on edge detection, which is a reasonable and reliable judgment basis.

[0020] Furthermore, step S300 includes:

[0021] Step S310: Obtain all target records corresponding to a certain processing equipment H, extract the processing time T corresponding to a certain target record C, match the time record X covering the processing time T according to all start and stop time records of the processing equipment H, obtain the equipment startup time ST of the time record X, obtain the duration between each moment in the processing time T and the equipment startup time ST, and assign points to the duration; establish a score function in which the score changes with the duration from the equipment startup time, obtain the final score corresponding to each duration according to the processing time corresponding to all target records, and take the duration with the largest final score as the target duration of the processing equipment H;

[0022] Taking the start time of a certain equipment of processing equipment H as the starting point, when the target duration is reached, edible pigment is added to the material inlet, and the monitoring equipment is used to detect the time when the edible pigment leaves the feed inlet and enters the machine, and the time when it leaves the machine and enters the discharge port. The time between the two moments is taken as the material processing time d of processing equipment H, and then the target degree of the equipment area corresponding to processing equipment H is obtained: y=e -k×d , k is the degree correlation coefficient, and then the target degree of each equipment area is obtained;

[0023] Step S320: Add the target levels corresponding to all equipment areas involved in a monitoring screen to obtain the total target level Y of the monitoring screen, and then obtain the first control level Z1=1-e of the monitoring screen -Y , as the first control level corresponding to a certain monitoring screen;

[0024] A monitoring segment of the history of a monitoring screen is intercepted to obtain the total duration of the characteristic time period in the monitoring segment, the total duration is divided by the duration corresponding to the monitoring segment to obtain a characteristic ratio, and the characteristic ratio is used as the second control level of the monitoring screen.

[0025] Further, step S400 includes: obtaining a first control degree and a second control degree of the monitoring screen, and setting corresponding weights respectively, thereby obtaining a total control degree of the monitoring screen, and obtaining a screen control sequence according to the order in which the monitoring screens are displayed in turn in the current traceability system;

[0026] According to the current display time of each monitoring screen, the display time sequence is obtained, and the cosine similarity between the screen control sequence and the display time sequence is calculated as the warning coefficient. If the warning coefficient is less than the preset numerical threshold, a warning prompt is given for the current screen display time.

[0027] A rice noodle processing quality monitoring system based on artificial intelligence, comprising a target record extraction module, a feature time period extraction module, a control degree calculation module and an early warning prompt module;

[0028] Target record extraction module: used to establish a 3D model of a rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the corresponding processing times of the processing records, analyze the operating status parameters of the processing equipment at each moment during the processing time, and obtain the target records in the processing records;

[0029] Feature time period extraction module: used to obtain the historical monitoring images of the factory, mark the local areas of moving objects in the monitoring images, and use the areas through which the rice noodles pass during the processing as the rice noodle areas. By analyzing the contact situation between the local areas and the rice noodle areas in the monitoring images, extract all the feature time periods in the monitoring images;

[0030] Control level calculation module: used to obtain the equipment areas corresponding to the processing equipment in the monitoring images, obtain the target levels corresponding to each equipment area according to the target records corresponding to the processing equipment, and obtain the first control level corresponding to the monitoring images according to the target levels; obtain the second control level corresponding to the monitoring images according to the feature time periods corresponding to the monitoring images;

[0031] Early warning prompt module: used to obtain the total control level of each monitoring image according to the first control level and the second control level; obtain the warning coefficient of the current display duration of the monitoring images according to the display duration of each monitoring image displayed in turn in the current traceability system and the total control level, and judge whether to give an early warning prompt for the current display duration of the monitoring images according to the warning coefficient.

[0032] Furthermore, the target record extraction module includes a processing record acquisition unit and a target record extraction unit;

[0033] Processing record acquisition unit: used to obtain the processing records corresponding to the historical processing equipment, obtain the corresponding processing times and processing equipment of the processing records, and extract the operating status parameters of the processing equipment at each moment during the processing time;

[0034] Target record extraction unit: used to obtain the first, second, and third target values according to the operating status parameters at each moment, and then obtain the total target value of the processing record, and then extract all the target records in the processing record.

[0035] Furthermore, the control level calculation module includes a target level calculation unit and a control level calculation unit;

[0036] Target level calculation unit: used to obtain all the target records corresponding to the processing equipment, obtain the target duration of the processing equipment according to the processing time corresponding to the target records, and obtain the target level of the equipment area corresponding to the processing equipment;

[0037] Control degree calculation unit: used to obtain the total target degree of the monitoring screen, and then obtain the first control degree of the monitoring screen; intercept the historical monitoring segments of a certain monitoring screen, obtain the total duration of the characteristic time periods in the monitoring segments, and obtain the second control degree of the monitoring screen.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention provides an artificial intelligence-based method and system for monitoring the quality of rice noodle processing, including: establishing a three-dimensional model of a rice noodle processing factory, obtaining the historical processing records of processing equipment, and obtaining the target records in the processing records; obtaining the historical monitoring screens of the factory, obtaining the local areas and rice noodle areas therein and analyzing them, and extracting all characteristic time periods in the monitoring screens; obtaining the equipment areas corresponding to the processing equipment, and according to the characteristic time periods, obtaining the first and second control degrees corresponding to the monitoring screen; obtaining the total control degree of each monitoring screen, obtaining the warning coefficient according to the display duration of the monitoring screen, and determining whether to give a warning prompt for the display duration of the screen. By analyzing the historical monitoring screens and the processing equipment in the factory, the present invention helps consumers comprehensively understand the processing process and improves the reliability and trust of the quality monitoring of rice noodle processing. Brief Description of the Drawings

[0039] Figure 1 It is a schematic flowchart of an artificial intelligence-based method for monitoring the quality of rice noodle processing according to the present invention;

[0040] Figure 2 It is a structural diagram of an artificial intelligence-based system for monitoring the quality of rice noodle processing according to the present invention. Detailed Embodiments

[0041] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment: As Figure 1 shown, the present invention provides a technical solution for an artificial intelligence-based method for monitoring the quality of rice noodle processing, including the following steps:

[0043] Step S100: Establish a three-dimensional model of a rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the processing times corresponding to the processing records, analyze the operating state parameters of the processing equipment at each moment during the processing time, and obtain the target records in the processing records;

[0044] Step S110: Obtain a number of processing records corresponding to each processing device in history. The processing records are records with a processing duration of D during the normal processing of the processing device after it is powered on. Obtain the processing time T and the processing device M corresponding to a certain processing record C, and extract the operating state parameters of the processing device M at each moment within the processing time T. The operating state parameters include temperature value, feed rate, and discharge rate. Among them, the temperature value is obtained by a temperature sensor deployed at a certain local position of the processing device M, and the feed rate and discharge rate are obtained by weight measuring devices deployed at the material inlet and outlet of the processing device M.

[0045] Step S120: According to the temperature value, feed rate, and discharge rate at each moment within the processing time T, calculate the variance corresponding to all temperature values as the first target value, calculate the variance corresponding to all feed rates as the second target value, and calculate the variance corresponding to all discharge rates as the third target value. Set the weights of the temperature value, feed rate, and discharge rate respectively, and obtain the total target value of the processing record C based on the first, second, and third target values. If the total target value is less than the preset target threshold, regard the processing record C as the target information, and then extract all target records in the processing record.

[0046] In this embodiment, the sum of the weights of the temperature value, feed rate, and discharge rate is set to 1. Since the first, second, and third target values are the variances corresponding to the temperature value, feed rate, and discharge rate respectively, then multiply the corresponding target value by the corresponding weight and add them up to obtain the total target value. The specific process will not be elaborated in detail.

[0047] Step S200: Obtain the historical monitoring images of the factory, mark the local area of the moving object in the monitoring images, and regard the area through which the rice noodles pass during the processing as the rice noodle area. By analyzing the contact situation between the local area and the rice noodle area in the monitoring images, extract all characteristic time periods in the monitoring images.

[0048] Step S210: Obtain the historical monitoring images obtained by the monitoring devices in the factory. Through target detection technology, obtain a local area R1 corresponding to a certain moving object in the monitoring images and a certain rice noodle area R2 through which the rice noodles pass during the processing. Among them, the local area R1 is a moving rectangular area, and the rice noodle area R2 is a fixed area. If there is a certain moment T1 when the intersection area between the local area R1 and the rice noodle area R2 in the monitoring images is not 0, and the intersection area between the local area R1 and the rice noodle area R2 at the previous moment of the moment T1 is 0, then regard the moment T1 as the just-contact moment.

[0049] Extract the monitoring segment F1 starting from time T1 with a duration of D1 backward. Take the time corresponding to the maximum intersection area between the local area R1 and the rice noodle area R2 in the monitoring segment F1 as T2. Take the area extending a length of L outward from the edge of the local area R1 as R3. Take the area where area R2 intersects with area R3 at time T2 as R4, and area R4 is a fixed area;

[0050] Step S220: Perform edge detection on area R4 in the monitoring screen at times T1 and T2 respectively to obtain edge images P1 and P2. Randomly extract several edge points in edge image P1 as reference points, and mark the corresponding positions of each reference point in image P2. Obtain the edge point in image P2 that is closest to a certain reference point n, and calculate the distance value between the two points as the target distance of reference point n. Then, based on the target distances of all reference points, calculate the variance as the deformation degree of area R4. If the deformation degree is greater than the preset degree threshold, take the time period between time T1 and time T3 as the characteristic time period, and then obtain all the characteristic time periods in the monitoring screen.

[0051] Step S300: Obtain the equipment area corresponding to the processing equipment in the monitoring screen. According to the target records corresponding to the processing equipment, obtain the target degree corresponding to each equipment area, and obtain the first control degree corresponding to the monitoring screen according to the target degree. Obtain the second control degree corresponding to the monitoring screen according to the characteristic time period corresponding to the monitoring screen;

[0052] Step S310: Obtain all the target records corresponding to a certain processing equipment H, extract the processing time T corresponding to a certain target record C among them. According to all the start and stop time records of processing equipment H, match the time record X that covers the processing time T, obtain the equipment start time ST of time record X, obtain the duration between each moment within the processing time T and the equipment start time ST, and assign scores to the durations. Establish a score function where the score changes with the duration from the equipment start time. According to the processing times corresponding to all the target records, obtain the final score corresponding to each duration, and take the duration with the maximum final score as the target duration of processing equipment H;

[0053] Given a calculation example: Assume that at a certain moment, the device startup time ST corresponding to X recorded by processing device H is 8 o'clock, and the processing time T corresponding to a certain target record C is from 10 o'clock to 11 o'clock. Then, the duration between each moment within the processing time T and the device startup time ST is between 2 hours and 3 hours. Since the score function has the score as the ordinate and the duration from the device startup time as the abscissa, here, let the score for each duration between 2 hours and 3 hours from the device startup time ST be 1. Then, if the processing time of another target record is from 9 o'clock to 10 o'clock and its corresponding device startup time is 7:30, then let the score for each duration between 1.5 hours and 2 hours, 2.5 hours and 3 hours from the device startup time ST be 1, and the score for each duration between 2 hours and 2.5 hours be 2, and so on, to obtain the final score corresponding to each duration.

[0054] Starting from a certain device startup time of processing device H, when the target duration is reached, an edible pigment is put into the material inlet. The moment when the edible pigment leaves the feed inlet and enters the machine, and the moment when it leaves the machine and enters the discharge outlet are detected by the monitoring device. The duration between the two moments is taken as the material processing duration d of processing device H, and then the target degree of the device area corresponding to processing device H is obtained as: y = e -k×d , where k is the degree-related coefficient, and then the target degree of each device area is obtained;

[0055] Rice noodle processing includes multiple processes, such as rice washing and soaking, grinding, gelatinization, cooling, and drying, etc. No matter which process it is, it takes a certain amount of time. During the actual observation by consumers, if the conversion time of the rice noodles in a certain link is relatively long, for example, the soaking link takes 2 hours, while the gelatinization link only takes 2 minutes. Since the state of the rice during the soaking link does not show immediately, compared with the rapidity of the gelatinization link, consumers are more concerned about the gelatinization link with a shorter time. And because the gelatinization link is faster, during the manufacturing process, the powder-making link has a greater impact on the final finished product effect of the rice noodles. Therefore, the first control degree can be obtained according to the speed of each process time in the rice noodle processing process. In this embodiment, k > 0. Since d > 0, then the value of y ranges from 0 to 1, and y decreases as d increases, which exactly conforms to the above principle.

[0056] Step S320: Add up the target degrees corresponding to all device areas involved in a certain monitoring screen to obtain the total target degree Y of a certain monitoring screen, and then obtain the first control degree Z1 = 1 - e of a certain monitoring screen -Y , as the first control degree corresponding to a certain monitoring screen;

[0057] Intercept the historical monitoring segments of a certain monitoring screen, obtain the total duration of the characteristic time period in the monitoring segment, divide the total duration by the duration corresponding to the monitoring segment to obtain a characteristic ratio, and use the characteristic ratio as the second control level of a certain monitoring screen.

[0058] Function Z1 = 1 - e -Y In this case, Z1 increases as Y increases, and since the value of Y is greater than 0, the value of the first control level Z1 ranges from 0 to 1. In this solution, the characteristic ratio obtained by dividing the total duration by the duration corresponding to the monitoring segment ranges from 0 to 1, so the value of the second control level also ranges from 0 to 1. Then, in step S400, let the first control level be Z1 and the second control level be Z2, and set the sum of the weights of the first control level and the second control level to 1. Then, the value of the total control level obtained also ranges from 0 to 1.

[0059] Step S400: Obtain the total control level of each monitoring screen according to the first control level and the second control level; according to the currently traced system, obtain the warning coefficient of the current display duration of the monitoring screen based on the display duration of each monitoring screen shown in turn and the total control level, and determine whether to give a warning prompt for the current display duration according to the warning coefficient.

[0060] Obtain the first control level and the second control level of the monitoring screen, and set the corresponding weights respectively, so as to obtain the total control level of the monitoring screen, and obtain the screen control sequence according to the order in which each monitoring screen is shown in turn in the currently traced system;

[0061] Obtain the display duration sequence according to the current display duration of each monitoring screen, and calculate the cosine similarity between the screen control sequence and the display duration sequence as the warning coefficient. If the warning coefficient is less than the preset numerical threshold, give a warning prompt for the current display duration.

[0062] The present invention also provides an artificial intelligence-based rice noodle processing quality monitoring system, as Figure 2 shown, including:

[0063] Target record extraction module: used to establish a three-dimensional model of the rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the corresponding processing time of the processing records, and analyze the operating state parameters of the processing equipment at each moment during the processing time to obtain the target records in the processing records;

[0064] Characteristic time period extraction module: used to obtain the historical monitoring screen of the factory, mark the local area of the moving object in the monitoring screen, and use the area where the rice noodles pass during the processing as the rice noodle area. By analyzing the contact situation between the local area and the rice noodle area in the monitoring screen, extract all the characteristic time periods in the monitoring screen;

[0065] The control level calculation module: used to obtain the equipment area corresponding to the processing equipment in the monitoring screen, obtain the target level corresponding to each equipment area according to the target record corresponding to the processing equipment, and obtain the first control level corresponding to the monitoring screen according to the target level; obtain the second control level corresponding to the monitoring screen according to the characteristic time period corresponding to the monitoring screen.

[0066] The warning prompt module: used to obtain the total control level of each monitoring screen according to the first control level and the second control level; obtain the warning coefficient of the current screen display duration according to the screen display duration of each monitoring screen displayed in turn in the current traceability system and the total control level, and determine whether to give a warning prompt for the current screen display duration according to the warning coefficient.

[0067] The target record extraction module includes a processing record acquisition unit and a target record extraction unit;

[0068] The processing record acquisition unit: used to obtain the processing records corresponding to the historical processing equipment, obtain the processing time and processing equipment corresponding to the processing records, and extract the operating state parameters of the processing equipment at each moment during the processing time.

[0069] The target record extraction unit: used to obtain the first, second, and third target values according to the operating state parameters at each moment, and then obtain the total target value of the processing record, and then extract all the target records in the processing record.

[0070] The control level calculation module includes a target level calculation unit and a control level calculation unit;

[0071] The target level calculation unit: used to obtain all the target records corresponding to the processing equipment, obtain the target duration of the processing equipment according to the processing time corresponding to the target record, and obtain the target level of the equipment area corresponding to the processing equipment.

[0072] The control level calculation unit: used to obtain the total target level of the monitoring screen, and then obtain the first control level of the monitoring screen; intercept the monitoring segment of a certain monitoring screen history, obtain the total duration of the characteristic time period in the monitoring segment, and obtain the second control level of the monitoring screen.

[0073] It is obvious to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

Claims

1. A method for monitoring the processing quality of rice noodles based on artificial intelligence, characterized in that, It includes the following steps: Step S100: Establish a three-dimensional model of the rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the corresponding processing times of the processing records, analyze the operating state parameters of the processing equipment at each moment during the processing time, and obtain the target records in the processing records; Step S200: Obtain the historical monitoring images of the factory, mark the local areas of the moving objects in the monitoring images, and use the areas passed by the rice noodles during the processing as the rice noodle areas. By analyzing the contact situation between the local areas and the rice noodle areas in the monitoring images, extract all characteristic time periods in the monitoring images; Step S300: Obtain the equipment areas corresponding to the processing equipment in the monitoring images. According to the target records corresponding to the processing equipment, obtain the target degree corresponding to each equipment area, and obtain the first control degree corresponding to the monitoring image based on the target degree; obtain the second control degree corresponding to the monitoring image according to the characteristic time periods corresponding to the monitoring image; Step S400: Obtain the total control degree of each monitoring image according to the first control degree and the second control degree; According to the current traceability system, rotate the display duration of each monitoring image and the total control degree, obtain the warning coefficient of the current display duration of the image, and determine whether to give a warning prompt for the current display duration of the image according to the warning coefficient.

2. The method for monitoring the processing quality of rice noodles based on artificial intelligence according to claim 1, wherein Step S100 includes: Step S110: Obtain a number of historical processing records corresponding to each processing equipment. The processing records are records with a processing duration of D during the normal processing of the processing equipment after startup; obtain the processing time T and the processing equipment M corresponding to a certain processing record C, and extract the operating state parameters of the processing equipment M at each moment during the processing time T. The operating state parameters include temperature values, feed rates, and discharge rates. Among them, the temperature values are obtained by deploying temperature sensors at a certain local position of the processing equipment M, and the feed rates and discharge rates are obtained by deploying weight measuring equipment at the material inlet and outlet of the processing equipment M; Step S120: According to the temperature values, feed rates, and discharge rates at each moment during the processing time T, calculate the variance corresponding to all temperature values as the first target value, calculate the variance corresponding to all feed rates as the second target value, and calculate the variance corresponding to all discharge rates as the third target value; respectively set the weight values of the temperature values, feed rates, and discharge rates, and obtain the total target value of the processing record C according to the first, second, and third target values. If the total target value is less than the preset target threshold, use the processing record C as the target information, and then extract all the target records in the processing record.

3. The method for monitoring the processing quality of rice noodles based on artificial intelligence according to claim 1, characterized in that, Step S200 includes: Step S210: Obtain the historical monitoring images obtained by the monitoring devices in the factory. Through target detection technology, obtain the local area R1 corresponding to a certain moving object in the monitoring image, and a certain rice noodle area R2 that the rice noodles pass through during the processing. Among them, the local area R1 is a moving rectangular area, and the rice noodle area R2 is a fixed area. If there is a moment T1 when the intersection area between the local area R1 and the rice noodle area R2 in the monitoring image is not 0, and the intersection area between the local area R1 and the rice noodle area R2 at the previous moment of the moment T1 is 0, then take the moment T1 as the just-contact moment; And extract the monitoring segment F1 with a duration of D1 starting from the moment T1. And take the moment corresponding to the maximum intersection area between the local area R1 and the rice noodle area R2 in the monitoring segment F1 as T2. Take the area extended by a length of L outward from the edge of the local area R1 as R3. And take the area where the area R2 intersects with the area R3 at the moment T2 as R4. The area R4 is a fixed area; Step S220: Perform edge detection on the area R4 in the monitoring image at the moments T1 and T2 respectively to obtain the edge images P1 and P2. Randomly extract several edge points in the edge image P1 as reference points, and mark the corresponding positions of each reference point in the image P2. Obtain the edge point in the image P2 that is closest to a certain reference point n, and calculate the distance value between the two points as the target distance of the reference point n. Then, based on the target distances of all reference points, calculate the variance as the deformation degree of the area R4. If the deformation degree is greater than the preset degree threshold, then take the time period between the moment T1 and the moment T3 as the characteristic time period, and thus obtain all the characteristic time periods in the monitoring image.

4. A method for monitoring the processing quality of rice noodles based on artificial intelligence according to claim 1, characterized in that Step S300 includes: Step S310: Obtain all the target records corresponding to a certain processing device H, extract the processing time T corresponding to a certain target record C among them. According to all the start-stop time records of the processing device H, match the time record X that covers the processing time T, obtain the device start time ST of the time record X, obtain the duration between each moment within the processing time T and the device start time ST, and assign scores to the durations. Establish a score function where the score changes with the duration from the device start time. According to the processing times corresponding to all the target records, obtain the final score corresponding to each duration, and take the duration with the maximum final score as the target duration of the processing device H; Starting from a certain equipment startup moment of processing equipment H, when the target duration is reached, an edible pigment is input into the material inlet. The moment when the edible pigment leaves the feed inlet and enters the machine, and the moment when it leaves the machine and enters the discharge outlet are detected by the monitoring equipment. The duration between the two moments is taken as the material processing duration d of processing equipment H. Furthermore, the target degree of the equipment area corresponding to processing equipment H is: y = e -k×d , where k is a degree-related coefficient, and then the target degree of each equipment area is obtained; Step S320: Add up the target degrees corresponding to all device areas involved in a certain monitoring screen to obtain the total target degree Y of the certain monitoring screen, and further obtain the first control degree Z1 = 1 - e -Y , as the first control degree corresponding to the certain monitoring screen; Intercept the monitoring segment of a certain historical monitoring image, obtain the total duration of the characteristic time periods in the monitoring segment, divide the total duration by the duration corresponding to the monitoring segment to obtain the characteristic ratio, and take the characteristic ratio as the second control degree of the certain monitoring image.

5. A method for monitoring the processing quality of rice noodles based on artificial intelligence according to claim 1, characterized in that, Step S400 includes: Obtain the first control degree and the second control degree of the monitoring image, and set the corresponding weights respectively, and then obtain the total control degree of the monitoring image. And according to the order in which each monitoring image is displayed in turn in the current traceability system, obtain the image control sequence; Based on the display duration of each current monitoring screen, obtain a display duration sequence, and calculate the cosine similarity between the screen control sequence and the display duration sequence as the warning coefficient. If the warning coefficient is less than the preset numerical threshold, give a warning prompt for the current screen display duration.

6. A rice noodle processing quality monitoring system for implementing the rice noodle processing quality monitoring method based on artificial intelligence according to any one of claims 1-5, characterized in that, The system includes a target record extraction module, a characteristic period extraction module, a control level calculation module, and a warning prompt module; Target record extraction module: used to establish a three-dimensional model of the rice noodle processing factory and mark the locations of all processing equipment in the factory; obtain the historical processing records of the processing equipment, extract the processing times corresponding to the processing records, analyze the operating state parameters of the processing equipment at each moment during the processing time, and obtain the target records in the processing records; Characteristic period extraction module: used to obtain the historical monitoring screens of the factory, mark the local areas of moving objects in the monitoring screens, and use the areas passed by the rice noodles during the processing process as the rice noodle areas. By analyzing the contact situation between the local areas and the rice noodle areas in the monitoring screens, extract all characteristic periods in the monitoring screens; Control level calculation module: used to obtain the equipment areas corresponding to the processing equipment in the monitoring screens, obtain the target levels corresponding to each equipment area according to the target records corresponding to the processing equipment, and obtain the first control level corresponding to the monitoring screen according to the target levels; obtain the second control level corresponding to the monitoring screen according to the characteristic periods corresponding to the monitoring screen; Warning prompt module: used to obtain the total control level of each monitoring screen according to the first control level and the second control level; obtain the warning coefficient of the current screen display duration according to the screen display duration of each monitoring screen displayed in turn in the current traceability system and the total control level, and determine whether to give a warning prompt for the current screen display duration according to the warning coefficient.

7. The quality monitoring system for rice noodle processing according to claim 6, characterized in that, The target record extraction module includes a processing record acquisition unit and a target record extraction unit; Processing record acquisition unit: used to obtain the processing records corresponding to the historical processing equipment, obtain the processing times and processing equipment corresponding to the processing records, and extract the operating state parameters of the processing equipment at each moment during the processing time; Target record extraction unit: used to obtain the first, second, and third target values according to the operating state parameters at each moment, and then obtain the total target value of the processing record, and then extract all target records in the processing record.

8. The quality monitoring system for rice noodle processing according to claim 6, characterized in that, The control level calculation module includes a target level calculation unit and a control level calculation unit; Target level calculation unit: used to obtain all target records corresponding to the processing equipment, obtain the target duration of the processing equipment according to the processing time corresponding to the target records, and obtain the target level of the equipment area corresponding to the processing equipment; Control level calculation unit: used to obtain the total target level of the monitoring screen, and then obtain the first control level of the monitoring screen; Intercept the historical monitoring segments of a certain monitoring screen, obtain the total duration of the characteristic periods in the monitoring segments, and obtain the second control level of the monitoring screen.

Images