Intelligent garbage classification system and method based on machine vision

Abstract

This invention relates to the field of waste sorting, specifically to an intelligent waste sorting system and method based on machine vision. The system includes a waste disposal data acquisition module, a target recognition model construction module, a waste capacity recognition signal response module, a sorting bin recognition module, an estimated capacity consumption value analysis module, and an optimal bin response activation module. The waste disposal data acquisition module acquires waste disposal data and video data captured by monitoring equipment. The target recognition model construction module determines the target recognition model for eligible waste disposal. The waste capacity recognition signal response module responds to the target recognition model to make a judgment when the monitoring equipment detects a waste disposal user, transmitting a waste capacity recognition signal corresponding to the waste disposal type. The sorting bin recognition module determines the number of sorting bins that can be used for waste disposal of the corresponding waste type.

Description

Technical Field

[0001] This invention relates to the field of waste sorting technology, specifically to an intelligent waste sorting system and method based on machine vision. Background Technology

[0002] Waste sorting generally refers to a series of activities involving the sorting, storage, disposal, and transportation of waste according to certain regulations or standards, thereby transforming it into public resources. The purpose of sorting is to increase the resource and economic value of waste, striving to make the most of everything. Sorted waste collection can reduce the amount of waste requiring processing equipment, lower processing costs, and reduce the consumption of land resources, resulting in social, economic, and ecological benefits. People are now gradually realizing the necessity of waste sorting; this has led to the development of modern intelligent waste sorting equipment, such as smart waste recycling equipment installed in communities. Although users typically dispose of sorted waste on this equipment, most recycling machines have multiple sorting bins that are closed before being opened. Users cannot see the capacity of the bins, leading to situations where bins are opened but waste cannot be disposed of, reducing user comfort and negatively impacting waste sorting. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent waste sorting system and method based on machine vision to solve the problems mentioned in the background art.

[0004] To address the aforementioned technical problems, this invention provides the following technical solution: an intelligent waste sorting method based on machine vision, comprising the following analysis steps:

[0005] Step S100: Obtain the waste disposal data recorded by the user after logging into the smart waste sorting device, as well as the video data captured by the monitoring equipment associated with the smart waste sorting device; determine the target recognition model for the waste that can be disposed of in the smart waste sorting device based on the video data and the waste disposal data;

[0006] Step S200: When the monitoring equipment detects a user disposing of waste, the target recognition model makes a judgment and, based on the judgment result, transmits the waste capacity recognition signal corresponding to the waste disposal type to the analysis terminal of the intelligent waste sorting equipment.

[0007] Step S300: Based on the waste capacity recognition signal, the intelligent waste sorting equipment analysis terminal determines the number of sorting bins that can be used for the waste disposal type corresponding to the judgment result. When the number of sorting bins that can be used is one, the corresponding bin is opened. When the number of sorting bins that can be used is more than one, the intelligent waste sorting equipment analysis terminal extracts the historical waste disposal data of the corresponding disposal user and estimates the disposal capacity consumption value.

[0008] Step S400: The intelligent waste sorting equipment analyzer obtains the space capacity data of the allowed sorting bins, compares the space capacity data with the estimated disposal capacity consumption value, outputs the optimal allowed sorting bins, and responds by opening the corresponding bins.

[0009] Furthermore, the target recognition model for recyclable waste in intelligent waste sorting equipment is determined, including the following steps:

[0010] Step S110: Waste disposal data includes disposal user information, disposal time, disposal type, and disposal weight; all disposal user information recorded for the same disposal type is stored in information units, with each information unit corresponding to one disposal type; video data corresponding to each disposal user information in each information unit is obtained;

[0011] Step S120: In each information unit, the video data of each user is stored in the form of image frames at the same time interval in the corresponding user's frame set A. ; This represents the set of frames for the i-th user in the information unit. The first, second, ..., nth image frames represent the i-th user's delivery in the information unit. The first image frame is the image frame recorded when the monitoring device first captures the delivery object of the user, and the nth image frame is the image frame captured by the monitoring device when the user delivers the object to the sorting box.

[0012] Step S130: Extract the recognition area of ​​the projected subject in the j-th image frame of the frame set. , Calculate the sharpness of the subject recognition corresponding to the j-th image frame. , ,in Represents the total area of ​​the image frame, extracting the maximum sharpness. The corresponding image frame is the target frame for the corresponding user.

[0013] Extract the pixel value of each pixel in the target frame record of the i-th delivery user in the information unit. and number of pixels , pixel value and number of pixels Normalization yields and Using the formula:

[0014]

[0015] The target identification index of the i-th user in the calculation information unit ,in This represents summing the pixel values ​​of all pixels recording the subject in the target frame corresponding to the i-th delivery user; This represents the reference coefficient corresponding to the pixel value. This represents the reference coefficient corresponding to the number of pixels;

[0016] Step S140: Based on the target identification index of each user, construct the target identification model Y for the information unit. Where Q represents the type of waste disposal corresponding to the information unit. This indicates the identification index range corresponding to the delivery type. This indicates the minimum value of the target identification index recorded under the corresponding waste disposal type. This indicates the maximum value of the target identification index recorded under the corresponding waste disposal type.

[0017] Furthermore, step S200 includes the following steps:

[0018] Step S210: The monitoring equipment captures the user who was deployed when the number of image frames in the video data recorded by the monitoring equipment that indicate the user's appearance is greater than the number threshold. The monitoring equipment acquires the image frames of the video data recorded by the monitoring equipment and calculates the target recognition index R corresponding to each image frame.

[0019] Step S220: When there is a corresponding image frame in the video data and Existing independently, or Independent existence Extraction when not existing independently The waste disposal type Q recorded by the target recognition model is the first judgment result; Independent existence refers to the target recognition model recording any two types of waste disposal. The sets that intersect each other form an empty set; Independent existence means that R has one and only one record. middle;

[0020] When there exists a corresponding image frame in the video data When R does not exist independently, the recognition index interval recorded by the target recognition model to which R belongs is marked as the interval to be analyzed, using the formula: Calculate the proportion of image frames examined corresponding to the target recognition index in the interval to be analyzed; This indicates the number of image frames corresponding to the target recognition index within the interval to be analyzed. This represents the total number of image frames recorded in the video data; the maximum value of the selected evaluation ratio is used. The target identification model corresponding to the interval to be analyzed is the optimal identification model, and the waste disposal type Q recorded by the optimal identification model is extracted as the first judgment result;

[0021] Step S230: Obtain the second judgment result, which is the result of the user's selection of the type of waste on the interactive display screen of the intelligent waste sorting device; when the first judgment result is the same as the second judgment result, output either judgment result; when the first judgment result is different from the second judgment result, issue an early warning and output the second judgment result as the final judgment result; the early warning response refers to the maintenance and management of the target recognition model.

[0022] The purpose of prioritizing judgment rather than simply determining the type of garbage based on user selection is to obtain a more comprehensive data base for garbage capacity analysis, and determining user selection can further verify the accuracy of the system analysis.

[0023] Furthermore, the estimated capacity consumption value in step S300 includes the following analysis steps:

[0024] Step S310: Obtain the disposal time corresponding to the same type of waste disposal recorded by the user based on the intelligent waste sorting device, and calculate the interval period T corresponding to the user. The interval period refers to the number of days between adjacent disposal times; the disposal weight recorded at the next disposal time within the interval period is taken as the target weight value for the corresponding interval period. Calculate the unit cycle output value W corresponding to the interval period. Obtain all interval periods and corresponding unit period output values ​​W for the same waste disposal type from the same user, and calculate the average unit period output value. , M represents the number of intervals;

[0025] Step S320: Obtain the target identification index R and target weight value X corresponding to the same waste disposal type for the same user, establish a data set (R, X) for each disposal record, and construct a functional relationship model with R as the independent variable and X as the dependent variable based on the data set. ;

[0026] Step S330: Obtain the actual interval between when the monitoring device captures the user's delivery and the user's last delivery event. And the actual target recognition index calculated from the video data captured by the monitoring equipment. The actual target identification index Substitute into the functional relationship model Output the corresponding first target weight value And using the formula: Calculate the weight value of the second target. Output and The maximum value is used as the effective target weight value; selecting the maximum value can minimize the error caused by the calculation capacity and reduce the possibility of abnormal user delivery.

[0027] Step S340: Obtain the bin capacity data for each waste disposal by the user. The bin capacity data includes the remaining capacity of the bin before waste disposal. and the remaining capacity of the bin after garbage disposal Calculate the capacity effect value D for each unit change in weight. And calculate the average capacity impact value of the same user on the same campaign type. , ;

[0028] Step S350: Using the formula: Calculate the capacity consumption value V for the same user and corresponding waste disposal type.

[0029] Furthermore, step S400 includes the following steps:

[0030] Extract the allowable category bins whose record space capacity is greater than the delivery capacity consumption value V and designate them as target bins. When there is only one target bin, mark it as the optimal allowable category bin. When there is more than one target bin, select the target bin that is closest to the delivery user as the optimal allowable category bin.

[0031] A machine vision-based intelligent waste sorting system includes a data acquisition module, a target recognition model construction module, a waste capacity recognition signal response module, a sorting bin recognition module, an estimated capacity consumption value analysis module, and an optimal bin response opening module.

[0032] The waste disposal data acquisition module is used to acquire waste disposal data recorded by users after logging into the smart waste sorting device, as well as video data captured by the monitoring equipment associated with the smart waste sorting device;

[0033] The target recognition model building module is used to determine the target recognition model for recyclable waste in intelligent waste sorting equipment;

[0034] The waste capacity identification signal response module is used to respond to the target identification model to make a judgment when the monitoring equipment detects a user disposing of waste, and to transmit the waste capacity identification signal corresponding to the waste disposal type.

[0035] The sorting bin identification module is used to determine the number of sorting bins that can be used to identify the type of waste to be disposed of, corresponding to the judgment result.

[0036] The estimated capacity consumption analysis module is used to extract historical waste disposal data of the corresponding disposal user and estimate the disposal capacity consumption when the number of bins is greater than one.

[0037] The optimal container response activation module compares the space capacity data with the estimated deployment capacity consumption value, outputs the optimal category container that can be deployed, and activates the corresponding container.

[0038] Furthermore, the target identifiable model building module includes an information unit classification unit, a frame set construction unit, a recognition sharpness calculation unit, a target recognition index calculation unit, and a model building output unit;

[0039] The information unit classification unit is used to store all user information for the same delivery type in information units;

[0040] The frame set construction unit is used to store the video data of each user in the form of image frames at the same time interval in each information unit, forming a frame set for the corresponding user.

[0041] The sharpness calculation unit is used to calculate the sharpness of the subject corresponding to the image frame.

[0042] The target identification index calculation unit is used to calculate the target identification index of the users deployed in the information unit;

[0043] The model building output unit is used to build a target recognition model for the information unit based on the target recognition index of each user.

[0044] Furthermore, the waste capacity identification signal response module includes a judgment result analysis unit and a result verification unit;

[0045] The judgment result analysis unit is used to calculate the target recognition index corresponding to each image frame, determine whether the interval to be analyzed corresponding to the target recognition index is independent, and output the corresponding first judgment result.

[0046] The result verification unit is used to obtain the second judgment result and output a corresponding response when the first judgment result is consistent with or inconsistent with the second judgment result.

[0047] Furthermore, the estimated capacity consumption value analysis module includes a first target weight value calculation unit, a second target weight value calculation unit, an average capacity impact value calculation unit, and a deployment capacity consumption value calculation unit;

[0048] The first target weight value calculation unit is used to calculate the first target weight value based on a functional relationship model;

[0049] The second target weight value calculation unit is used to calculate the second target weight value using the average unit cycle output value and the actual interval cycle.

[0050] The average capacity impact value calculation unit is used to calculate the capacity impact value for each unit of weight change and to determine the average capacity impact value for the same user and the same delivery type.

[0051] The capacity consumption calculation unit is used to calculate the capacity consumption value of the same waste disposal user for the corresponding waste disposal type based on the maximum value of the first target weight and the second target weight and the average capacity influence value.

[0052] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention analyzes the garbage disposal data recorded by logged-in users on intelligent garbage sorting equipment and the video data captured by monitoring equipment to establish a target recognition model. When a user enters the monitoring range, the invention performs response analysis to determine the identified garbage disposal type and the corresponding number of bins. When the number is greater than one, the invention analyzes the garbage capacity that the bins can hold based on machine vision and data analysis to recommend bins that the user can successfully dispose of in one go. This allows users to conveniently dispose of garbage into the sorting bins in one go, even if they do not understand the capacity or cannot make a concrete judgment about the capacity. This improves the intelligence of the garbage sorting equipment in serving users and increases user comfort. Attached Figure Description

[0053] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0054] Figure 1 This is a schematic diagram of the structure of an intelligent waste sorting system and method based on machine vision according to the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] Please see Figure 1 This invention provides a technical solution: an intelligent waste sorting method based on machine vision, comprising the following analysis steps:

[0057] Step S100: Obtain the waste disposal data recorded by the user after logging into the smart waste sorting device, as well as the video data captured by the monitoring equipment associated with the smart waste sorting device; determine the target recognition model for the waste that can be disposed of in the smart waste sorting device based on the video data and the waste disposal data;

[0058] Determine the target recognition model for recyclable waste in intelligent waste sorting equipment, including the following steps:

[0059] Step S110: Waste disposal data includes disposal user information, disposal time, disposal type, and disposal weight; all disposal user information recorded for the same disposal type is stored in information units, with each information unit corresponding to one disposal type; video data corresponding to each disposal user information in each information unit is obtained;

[0060] Step S120: In each information unit, the video data of each user is stored in the form of image frames at the same time interval in the corresponding user's frame set A. ; This represents the set of frames for the i-th user in the information unit. The first, second, ..., nth image frames represent the i-th user's delivery in the information unit. The first image frame is the image frame recorded when the monitoring device first captures the delivery object of the user, and the nth image frame is the image frame captured by the monitoring device when the user delivers the object to the sorting box.

[0061] In this application, the main body of the waste disposal can be cardboard, plastic, etc., and the intelligent waste sorting equipment in this application is a waste recycling equipment that can identify the main body of the waste using machine vision. In this scenario, the waste disposed of by the user is usually of the same type.

[0062] Step S130: Extract the recognition area of ​​the projected subject in the j-th image frame of the frame set. , Calculate the sharpness of the subject recognition corresponding to the j-th image frame. , ,in Represents the total area of ​​the image frame, extracting the maximum sharpness. The corresponding image frame is the target frame for the corresponding user. When the number of frames recorded by the user in the same information unit is not one, the image frame corresponding to the maximum value is still selected from all image frames according to the above-mentioned method of analyzing the clarity.

[0063] Extract the pixel value of each pixel in the target frame record of the i-th delivery user in the information unit. and number of pixels , pixel value and number of pixels Normalization yields and Using the formula:

[0064]

[0065] The target identification index of the i-th user in the calculation information unit ,in This represents summing the pixel values ​​of all pixels recording the subject in the target frame corresponding to the i-th delivery user; This represents the reference coefficient corresponding to the pixel value. This represents the reference coefficient corresponding to the number of pixels;

[0066] Step S140: Based on the target identification index of each user, construct the target identification model Y for the information unit. Where Q represents the type of waste disposal corresponding to the information unit. This indicates the identification index range corresponding to the delivery type. This indicates the minimum value of the target identification index recorded under the corresponding waste disposal type. This indicates the maximum value of the target identification index recorded under the corresponding waste disposal type.

[0067] Step S200: When the monitoring equipment detects a user disposing of waste, the target recognition model makes a judgment and, based on the judgment result, transmits the waste capacity recognition signal corresponding to the waste disposal type to the analysis terminal of the intelligent waste sorting equipment.

[0068] Step S200 includes the following steps:

[0069] Step S210: The monitoring equipment captures the user who was deployed when the number of image frames in the video data recorded by the monitoring equipment that indicate the user's appearance is greater than the number threshold. The monitoring equipment acquires the image frames of the video data recorded by the monitoring equipment and calculates the target recognition index R corresponding to each image frame.

[0070] Step S220: When there is a corresponding image frame in the video data and Existing independently, or Independent existence Extraction when not existing independently The waste disposal type Q recorded by the target recognition model is the first judgment result; Independent existence refers to the target recognition model recording any two types of waste disposal. The sets that intersect each other form an empty set; Independent existence means that R has one and only one record. middle;

[0071] When there exists a corresponding image frame in the video data When R does not exist independently, the recognition index interval recorded by the target recognition model to which R belongs is marked as the interval to be analyzed, using the formula: Calculate the proportion of image frames examined corresponding to the target recognition index in the interval to be analyzed; This indicates the number of image frames corresponding to the target recognition index within the interval to be analyzed. This represents the total number of image frames recorded in the video data; the maximum value of the selected evaluation ratio is used. The target identification model corresponding to the interval to be analyzed is the optimal identification model, and the waste disposal type Q recorded by the optimal identification model is extracted as the first judgment result;

[0072] Step S230: Obtain the second judgment result, which is the result of the user's selection of the type of waste on the interactive display screen of the intelligent waste sorting device; when the first judgment result is the same as the second judgment result, output either judgment result; when the first judgment result is different from the second judgment result, issue an early warning and output the second judgment result as the final judgment result; the early warning response refers to the maintenance and management of the target recognition model.

[0073] The purpose of prioritizing judgment rather than simply determining the type of garbage based on user selection is to obtain a more comprehensive data base for garbage capacity analysis, and determining user selection can further verify the accuracy of the system analysis.

[0074] Step S300: Based on the waste capacity recognition signal, the intelligent waste sorting equipment analysis terminal determines the number of sorting bins that can be used for the waste disposal type corresponding to the judgment result. When the number of sorting bins that can be used is one, the corresponding bin is opened. When the number of sorting bins that can be used is greater than one, the intelligent waste sorting equipment analysis terminal extracts the historical waste disposal data of the corresponding user and estimates the disposal capacity consumption value. In this application, the case where the number of sorting bins that can be used is 0 is not considered, because in general, the system will prompt on the device interaction terminal when the number of bins is 0.

[0075] Step S300, which estimates the capacity consumption, includes the following analysis steps:

[0076] Step S310: Obtain the disposal time corresponding to the same type of waste disposal recorded by the user based on the intelligent waste sorting device, and calculate the interval period T corresponding to the user. The interval period refers to the number of days between adjacent disposal times; the disposal weight recorded at the next disposal time within the interval period is taken as the target weight value for the corresponding interval period. Calculate the unit cycle output value W corresponding to the interval period. Obtain all interval periods and corresponding unit period output values ​​W for the same waste disposal type from the same user, and calculate the average unit period output value. , M represents the number of intervals;

[0077] Step S320: Obtain the target identification index R and target weight value X corresponding to the same waste disposal type for the same user, establish a data set (R, X) for each disposal record, and construct a functional relationship model with R as the independent variable and X as the dependent variable based on the data set. For paper items, the greater the weight, the larger the volume that appears in the image. Therefore, the corresponding functional relationship model can be established as a linear regression model to describe the mathematical relationship between the image data of garbage disposal and the actual weight.

[0078] Step S330: Obtain the actual interval between when the monitoring device captures the user's delivery and the user's last delivery event. And the actual target recognition index calculated from the video data captured by the monitoring equipment. The actual target identification index Substitute into the functional relationship model Output the corresponding first target weight value And using the formula: Calculate the weight value of the second target. Output and The maximum value is used as the effective target weight value; selecting the maximum value can minimize the error caused by the calculation capacity and reduce the possibility of abnormal user delivery.

[0079] Step S340: Obtain the bin capacity data for each waste disposal by the user. The bin capacity data includes the remaining capacity of the bin before waste disposal. and the remaining capacity of the bin after garbage disposal The remaining capacity refers to the capacity obtained after the box has been compressed inside for a period of time and returned to a static state; calculate the capacity impact value D for each unit change in weight. And calculate the average capacity impact value of the same user on the same campaign type. , ;

[0080] Step S350: Using the formula: Calculate the capacity consumption value V for the same user and corresponding waste disposal type.

[0081] Step S400: The intelligent waste sorting equipment analyzer obtains the space capacity data of the allowed sorting bins, compares the space capacity data with the estimated disposal capacity consumption value, outputs the optimal allowed sorting bins, and responds by opening the corresponding bins.

[0082] Step S400 includes the following steps:

[0083] Extract the allowable category bins whose record space capacity is greater than the delivery capacity consumption value V and designate them as target bins. When there is only one target bin, mark it as the optimal allowable category bin. When there is more than one target bin, select the target bin that is closest to the delivery user as the optimal allowable category bin.

[0084] A machine vision-based intelligent waste sorting system includes a data acquisition module, a target recognition model construction module, a waste capacity recognition signal response module, a sorting bin recognition module, an estimated capacity consumption value analysis module, and an optimal bin response opening module.

[0085] The waste disposal data acquisition module is used to acquire waste disposal data recorded by users after logging into the smart waste sorting device, as well as video data captured by the monitoring equipment associated with the smart waste sorting device;

[0086] The target recognition model building module is used to determine the target recognition model for recyclable waste in intelligent waste sorting equipment;

[0087] The waste capacity identification signal response module is used to respond to the target identification model to make a judgment when the monitoring equipment detects a user disposing of waste, and to transmit the waste capacity identification signal corresponding to the waste disposal type.

[0088] The sorting bin identification module is used to determine the number of sorting bins that can be used to identify the type of waste to be disposed of, corresponding to the judgment result.

[0089] The estimated capacity consumption analysis module is used to extract historical waste disposal data of the corresponding disposal user and estimate the disposal capacity consumption when the number of bins is greater than one.

[0090] The optimal container response activation module compares the space capacity data with the estimated deployment capacity consumption value, outputs the optimal category container that can be deployed, and activates the corresponding container.

[0091] The target recognition model construction module includes an information unit classification unit, a frame set construction unit, a recognition sharpness calculation unit, a target recognition index calculation unit, and a model construction output unit.

[0092] The information unit classification unit is used to store all user information for the same delivery type in information units;

[0093] The frame set construction unit is used to store the video data of each user in the form of image frames at the same time interval in each information unit, forming a frame set for the corresponding user.

[0094] The sharpness calculation unit is used to calculate the sharpness of the subject corresponding to the image frame.

[0095] The target identification index calculation unit is used to calculate the target identification index of the users deployed in the information unit;

[0096] The model building output unit is used to build a target recognition model for the information unit based on the target recognition index of each user.

[0097] The waste capacity identification signal response module includes a judgment result analysis unit and a result verification unit;

[0098] The judgment result analysis unit is used to calculate the target recognition index corresponding to each image frame, determine whether the interval to be analyzed corresponding to the target recognition index is independent, and output the corresponding first judgment result.

[0099] The result verification unit is used to obtain the second judgment result and output a corresponding response when the first judgment result is consistent with or inconsistent with the second judgment result.

[0100] The estimated capacity consumption value analysis module includes a first target weight value calculation unit, a second target weight value calculation unit, an average capacity impact value calculation unit, and a deployment capacity consumption value calculation unit.

[0101] The first target weight value calculation unit is used to calculate the first target weight value based on a functional relationship model;

[0102] The second target weight value calculation unit is used to calculate the second target weight value using the average unit cycle output value and the actual interval cycle.

[0103] The average capacity impact value calculation unit is used to calculate the capacity impact value for each unit of weight change and to determine the average capacity impact value for the same user and the same delivery type.

[0104] The capacity consumption calculation unit is used to calculate the capacity consumption value of the same waste disposal user for the corresponding waste disposal type based on the maximum value of the first target weight and the second target weight and the average capacity influence value.

[0105] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0106] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A machine vision-based intelligent waste sorting method, characterized in that, The analysis includes the following steps: Step S100: Obtain the waste disposal data recorded by the user after logging into the smart waste sorting device, as well as the video data captured by the monitoring device associated with the smart waste sorting device; Based on video data and waste disposal data, a target recognition model for disposable waste in intelligent waste sorting equipment is determined. The target recognition model includes a target recognition index R; Step S200: When the monitoring equipment detects a user disposing of waste, the target recognition model makes a judgment and, based on the judgment result, transmits the waste capacity recognition signal corresponding to the waste disposal type to the analysis terminal of the intelligent waste sorting equipment. Step S300: Based on the waste capacity recognition signal, the intelligent waste sorting equipment analysis terminal determines the number of sorting bins that can be used for the waste disposal type corresponding to the judgment result. When the number of sorting bins that can be used is one, the corresponding bin is opened. When the number of sorting bins that can be used is more than one, the intelligent waste sorting equipment analysis terminal extracts the historical waste disposal data of the corresponding disposal user and estimates the disposal capacity consumption value. The estimated capacity consumption value in step S300 includes the following analysis steps: Step S310: Obtain the disposal time corresponding to the same type of waste disposal recorded by the user based on the intelligent waste sorting device, and calculate the interval period T corresponding to the user, where the interval period refers to the number of days between adjacent disposal times; use the disposal weight recorded at the next disposal time within the interval period as the target weight value for the corresponding interval period. Calculate the unit cycle output value W corresponding to the interval period. Obtain all interval periods and corresponding unit period output values ​​W for the same waste disposal type from the same user, and calculate the average unit period output value. , M represents the number of intervals; Step S320: Obtain the target identification index R and target weight value X corresponding to the same waste disposal type for the same user, establish a data set (R, X) for each disposal record, and construct a functional relationship model with R as the independent variable and X as the dependent variable based on the data set. ; Step S330: Obtain the actual interval between when the monitoring device captures the user's delivery and the user's last delivery event. And the actual target recognition index calculated from the video data captured by the monitoring equipment. The actual target identification index Substitute into the functional relationship model Output the corresponding first target weight value And using the formula: Calculate the weight value of the second target. Output and The maximum value is taken as the effective target weight value; Step S340: Obtain the bin capacity data for each waste disposal by the user, the bin capacity data including the remaining capacity of the bin before waste disposal. and the remaining capacity of the bin after garbage disposal ; Calculate the capacity effect value D for each unit change in weight. And calculate the average capacity impact value of the same user on the same campaign type. , ; Step S350: Using the formula: Calculate the capacity consumption value V for the same user and corresponding waste disposal type; Step S400: The intelligent waste sorting equipment analyzer obtains the space capacity data of the allowed sorting bins, compares the space capacity data with the estimated disposal capacity consumption value, outputs the optimal allowed sorting bins and responds to open the corresponding bins.

2. The intelligent waste sorting method based on machine vision according to claim 1, characterized in that: The process of determining the target identification model for disposable waste in the intelligent waste sorting equipment includes the following steps: Step S110: The waste disposal data includes disposal user information, disposal time, disposal type, and disposal weight; all disposal user information recorded for the same disposal type is stored in information units, with each information unit corresponding to a disposal type; video data corresponding to each disposal user information in each information unit is obtained; Step S120: In each information unit, the video data of each user is stored in the form of image frames at the same time interval in the corresponding user's frame set A. ; This represents the set of frames for the i-th user in the information unit. The first, second, ..., nth image frames of the i-th user in the information unit are represented. The first image frame is the image frame recorded when the monitoring device first captures the user's object being placed in the sorting bin. The nth image frame is the image frame captured by the monitoring device when the user places the object into the sorting bin. Step S130: Extract the recognition area of ​​the projected subject in the j-th image frame of the frame set. , Calculate the sharpness of the subject recognition corresponding to the j-th image frame. , ,in Represents the total area of ​​the image frame, extracting the maximum sharpness. The corresponding image frame is the target frame for the corresponding user. Extract the pixel value of each pixel in the target frame record of the i-th delivery user in the information unit. and number of pixels , pixel value and number of pixels Normalization yields and Using the formula: The target identification index of the i-th user in the calculation information unit ,in This represents summing the pixel values ​​of all pixels recording the subject in the target frame corresponding to the i-th delivery user; This represents the reference coefficient corresponding to the pixel value. This represents the reference coefficient corresponding to the number of pixels; Step S140: Based on the target identification index of each user, construct the target identification model Y for the information unit. Where Q represents the type of waste disposal corresponding to the information unit. This indicates the identification index range corresponding to the delivery type. This indicates the minimum value of the target identification index recorded under the corresponding waste disposal type. This indicates the maximum value of the target identification index recorded under the corresponding waste disposal type.

3. The intelligent waste sorting method based on machine vision according to claim 2, characterized in that: Step S200 includes the following steps: Step S210: The monitoring device captures the user when the number of image frames in the video data recorded by the monitoring device that mark the user is greater than the number threshold; obtain the image frames of the video data recorded by the monitoring device, and calculate the target recognition index R corresponding to each image frame; Step S220: When there is a corresponding image frame in the video data and Existing independently, or Independent existence Extraction when not existing independently The waste disposal type Q recorded by the target recognition model is the first judgment result; Independent existence refers to the target recognition model recording any two types of waste disposal. The intersection between them is an empty set; the aforementioned Independent existence means that R has one and only one record. middle; When there exists a corresponding image frame in the video data When R does not exist independently, the recognition index interval recorded by the target recognition model to which R belongs is marked as the interval to be analyzed, using the formula: Calculate the proportion of image frames examined corresponding to the target recognition index in the interval to be analyzed; This indicates the number of image frames corresponding to the target recognition index within the interval to be analyzed. This represents the total number of image frames recorded in the video data; the maximum value of the selected evaluation ratio is used. The target identification model corresponding to the interval to be analyzed is the optimal identification model, and the waste disposal type Q recorded by the optimal identification model is extracted as the first judgment result; Step S230: Obtain the second judgment result, which is the result of the user's selection of the type of waste on the interactive display screen of the intelligent waste sorting device; when the first judgment result is the same as the second judgment result, output either judgment result; when the first judgment result is different from the second judgment result, issue an early warning and output the second judgment result as the final judgment result; the early warning response refers to the maintenance and management of the target recognition model.

4. The intelligent waste sorting method based on machine vision according to claim 3, characterized in that: Step S400 includes the following steps: Extract the allowable category bins whose record space capacity is greater than the delivery capacity consumption value V and designate them as target bins. When there is only one target bin, mark it as the optimal allowable category bin. When there is more than one target bin, select the target bin that is closest to the delivery user as the optimal allowable category bin.

5. A machine vision-based intelligent waste sorting system, employing the machine vision-based intelligent waste sorting method according to any one of claims 1-4, characterized in that, It includes a data acquisition module, a target identification model construction module, a waste capacity identification signal response module, a sorting bin identification module, an estimated capacity consumption value analysis module, and an optimal bin response activation module; The data acquisition module is used to acquire the waste disposal data recorded by the user after logging into the smart waste sorting device, as well as the video data captured by the monitoring device associated with the smart waste sorting device. The target identifiable model construction module is used to determine the target identification model of the garbage that can be disposed of in the intelligent garbage sorting equipment; The waste capacity identification signal response module is used to respond to the target identification model to make a judgment when the monitoring equipment detects a user disposing of waste, and to transmit the waste capacity identification signal corresponding to the waste disposal type. The sorting bin identification module is used to determine the number of sorting bins that can be used to identify the waste disposal type corresponding to the judgment result. The estimated capacity consumption value analysis module is used to extract the historical waste disposal data of the corresponding disposal user and estimate the disposal capacity consumption value when the number of bins is greater than one. The optimal container response opening module is used to compare the space capacity data with the estimated deployment capacity consumption value, output the optimal category container that can be deployed, and respond by opening the corresponding container.

6. The intelligent waste sorting system based on machine vision according to claim 5, characterized in that: The target identifiable model construction module includes an information unit classification unit, a frame set construction unit, an identification clarity calculation unit, a target identification index calculation unit, and a model construction output unit. The information unit classification unit is used to store all user information recorded for the same delivery type in information units; The frame set construction unit is used to store the video data of each user in the form of image frames at the same time interval in each information unit as the frame set of the corresponding user. The recognition clarity calculation unit is used to calculate the recognition clarity of the subject corresponding to the image frame; The target identification index calculation unit is used to calculate the target identification index of the user in the information unit; The model construction output unit is used to construct the target recognition model of the information unit based on the target recognition index of each user.

7. The intelligent waste sorting system based on machine vision according to claim 6, characterized in that: The waste volume identification signal response module includes a judgment result analysis unit and a result verification unit; The judgment result analysis unit is used to calculate the target recognition index corresponding to each image frame, determine whether the interval to be analyzed corresponding to the target recognition index is independent, and output the corresponding first judgment result. The result verification unit is used to obtain the second judgment result and output a corresponding response when the first judgment result is consistent with or inconsistent with the second judgment result.

8. The intelligent waste sorting system based on machine vision according to claim 7, characterized in that: The estimated capacity consumption value analysis module includes a first target weight value calculation unit, a second target weight value calculation unit, an average capacity impact value calculation unit, and a deployment capacity consumption value calculation unit. The first target weight value calculation unit is used to calculate the first target weight value based on a functional relationship model; The second target weight value calculation unit is used to calculate the second target weight value using the average unit cycle output value and the actual interval cycle; The average capacity impact value calculation unit is used to calculate the capacity impact value for each unit of weight change, and to determine the average capacity impact value for the same user and the same delivery type. The capacity consumption calculation unit is used to calculate the capacity consumption value of the same waste disposal user for the corresponding waste disposal type based on the maximum value of the first target weight and the second target weight and the average capacity influence value.

Images