A garbage classification behavior data visualization analysis management system and method

Abstract

The present application relates to a data visualization analysis management system and method for garbage classification behavior, including: determining the instantaneous pressure-thickness change rate corresponding to the corresponding extrusion surface of the classification trash can based on pressure data and thickness data; obtaining and sending an instruction to the classification trash can to stop extrusion of the corresponding extrusion surface based on the relationship between the instantaneous pressure-thickness change rate and a predetermined value, so that it stops extrusion; identifying and sending the type of garbage under each extrusion surface to the classification trash can based on image data, so that the classification trash can puts the corresponding garbage into the corresponding classification compartment according to the type, wherein the image data is collected after all the extrusion surfaces stop extruding and are away from the garbage; determining and sending the classification evaluation of the garbage based on the type to the classification trash can, so that the classification trash can displays the classification evaluation. The present application can improve the willingness and execution of garbage classification and can carry out refined management of garbage classification behavior.

Description

Technical Field

[0001] The present application relates to the field of data analysis and management, and in particular to a garbage classification behavior data visualization analysis and management system and method. Background Art

[0002] Currently, the collection of garbage classification data mostly relies on manual spot checks, which can only obtain basic garbage classification data such as the total amount of garbage disposed of, the frequency of garbage disposal, and the number of garbage transportation times. The management of garbage classification behavior based on basic garbage classification data is bound to be relatively extensive, and there will be blind spots in the management of garbage classification behavior, which will lead to unsatisfactory garbage classification results, and in turn affect the promotion and popularization of garbage classification.

[0003] Therefore, how to improve residents' willingness and execution of garbage sorting, and to be able to carry out refined management of garbage sorting behavior, so as to achieve effective promotion and popularization of garbage sorting, is a technical problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0004] This application provides a garbage classification behavior data visualization analysis management system and method to improve residents' willingness and execution of garbage classification, and can carry out refined management of garbage classification behavior, thereby achieving effective promotion and popularization of garbage classification.

[0005] To solve the above technical problems, this application provides the following technical solutions:

[0006] A method for visual analysis and management of garbage classification behavior data, applied to a back-end server side, comprises the following steps: in response to receiving pressure data and thickness data, determining the instantaneous pressure-thickness change rate corresponding to the corresponding extrusion surface of the classification garbage bin according to the pressure data and thickness data; obtaining and sending an instruction for the corresponding extrusion surface to stop extrusion according to the relationship between the instantaneous pressure-thickness change rate and a predetermined value to the classification garbage bin, so that the corresponding extrusion surface stops extrusion; in response to receiving image data, identifying the type of garbage under each extrusion surface according to the image data and sending it to the classification garbage bin, so that the classification garbage bin puts the corresponding type of garbage into the corresponding classification compartment according to the type, wherein the image data is collected from the garbage after all the extrusion surfaces stop extruding and are away from the garbage; determining the classification evaluation of the garbage according to the type of the garbage and sending it to the classification garbage bin, so that the classification garbage bin displays the classification evaluation.

[0007] The garbage classification behavior data visualization analysis and management method as described above is applied to the back-end server side, wherein, preferably, determining the instantaneous pressure-thickness change rate includes the following sub-steps: judging whether there are extremely large outliers in the pressure data corresponding to the corresponding extrusion surface based on the sum of the mean value of the pressure data corresponding to each extrusion surface and a predetermined multiple of the standard deviation; if not, determining the instantaneous pressure-thickness change rate of the extrusion surface based on the change in the pressure data and the change in the thickness data corresponding to the extrusion surface.

[0008] The garbage classification behavior data visualization analysis and management method as described above is applied to the back-end server side, wherein, preferably, determining the instantaneous pressure-thickness change rate also includes the following sub-steps: if any, determining the normal instantaneous pressure-thickness change rate of the extrusion surface based on the change in normal pressure data and the corresponding thickness data corresponding to the extrusion surface; correcting the normal instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the maximum outlier value to the pressure data and the ratio of the number of maximum outliers to the number of pressure data corresponding to the extrusion surface to obtain the instantaneous pressure-thickness change rate of the extrusion surface.

[0009] The garbage classification behavior data visualization analysis and management method as described above is applied to the back-end server side, wherein, preferably, the type of garbage identified includes the following sub-steps: dividing the image data into multiple image data units according to the arrangement rules of the extrusion surface, and preprocessing each image data unit; subjecting each preprocessed image data unit to multiple convolutions and downsampling to obtain feature maps of different scales; performing convolution operations on the feature maps of different scales to obtain corresponding new feature maps, and fusing the new feature maps of different scales; inputting the fused feature map into the classifier to obtain the probability distribution of the fused feature map belonging to different types of garbage, so as to identify the type of garbage in the image data unit corresponding to the corresponding extrusion surface.

[0010] A method for visual analysis and management of garbage classification behavior data, applied to the classified garbage bin side, comprises the following steps: squeezing garbage, collecting and sending pressure data and thickness data corresponding to each squeezing surface to a back-end server, so that the back-end server determines the instantaneous pressure-thickness change rate corresponding to the corresponding squeezing surface; in response to receiving an instruction to stop squeezing, stopping squeezing of the corresponding squeezing surface, wherein the instruction to stop squeezing is obtained based on the relationship between the instantaneous pressure-thickness change rate and a predetermined value; in response to all squeezing surfaces stopping squeezing, all squeezing surfaces are moved away from the squeezed garbage, collecting and sending image data of the garbage to the back-end server, so that the back-end server identifies the type of garbage under each squeezing surface based on the image data; in response to receiving a classification evaluation of the garbage and the type of the garbage, placing the corresponding type of garbage into the corresponding classification compartment according to the type of the garbage, and displaying the classification evaluation, wherein the classification evaluation is determined by the back-end server based on the type of garbage under each squeezing surface.

[0011] The garbage classification behavior data visualization analysis and management method as described above is applied to the classification garbage bin side, wherein, preferably, determining the instantaneous pressure-thickness change rate includes the following sub-steps: judging whether there is a large abnormal value in the pressure data corresponding to the corresponding extrusion surface based on the sum of the mean value of the pressure data corresponding to each extrusion surface and the predetermined multiple of the standard deviation; if not, determining the instantaneous pressure-thickness change rate of the extrusion surface based on the change amount of the pressure data and the change amount of the thickness data corresponding to the extrusion surface.

[0012] The garbage classification behavior data visualization analysis and management method as described above is applied to the classification garbage bin side, wherein, preferably, determining the instantaneous pressure-thickness change rate also includes the following sub-steps: if any, determining the normal instantaneous pressure-thickness change rate of the extrusion surface based on the change in normal pressure data and the corresponding thickness data corresponding to the extrusion surface; correcting the normal instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the maximum outlier value to the pressure data and the ratio of the number of maximum outliers to the number of pressure data corresponding to the extrusion surface to obtain the instantaneous pressure-thickness change rate of the extrusion surface.

[0013] The garbage classification behavior data visualization analysis and management method as described above is applied to the classified garbage bin side, wherein, preferably, the type of garbage identified includes the following sub-steps: dividing the image data into multiple image data units according to the arrangement rule of the extrusion surface, and preprocessing each image data unit; subjecting each preprocessed image data unit to multiple convolutions and downsampling to obtain feature maps of different scales; performing convolution operations on the feature maps of different scales to obtain corresponding new feature maps, and fusing the new feature maps of different scales; inputting the fused feature map into the classifier to obtain the probability distribution of the fused feature map belonging to different types of garbage, so as to identify the type of garbage in the image data unit corresponding to the corresponding extrusion surface.

[0014] A garbage classification behavior data visualization analysis and management system includes: a classification garbage bin and a back-end server; the back-end server executes any of the above-mentioned methods applied to the back-end server side, and the classification garbage bin executes any of the above-mentioned methods applied to the classification garbage bin side.

[0015] The above-mentioned garbage classification behavior data visualization analysis and management system, wherein preferably, the classification garbage bin includes: a garbage squeezing device, a pressure collection sensor unit and a thickness collection sensor; the garbage squeezing device has An extrusion surface of uniform size, is the number of rows of extruded surfaces, is the number of columns of the extrusion surface; each extrusion surface is composed of a plurality of pressure acquisition sensing units to form a pressure sensing array, and the distribution density of the pressure acquisition sensing units on the middle extrusion surface is several times the distribution density of the pressure acquisition sensing units on the edge extrusion surface; each extrusion surface is configured with a corresponding thickness acquisition sensor.

[0016] Compared with the above-mentioned background technology, the garbage classification behavior data visualization analysis management system and method provided by this application can improve residents' willingness and execution of garbage classification, and can carry out refined management of garbage classification behavior, thereby realizing the effective promotion and popularization of garbage classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0018] Figure 1 This is a schematic diagram of the garbage classification behavior data visualization analysis and management system provided by this application;

[0019] Figure 2 This is a flowchart of the garbage classification behavior data visualization analysis and management method provided by this application;

[0020] Figure 3 is a flow chart for determining the instantaneous pressure-thickness change rate provided by the present application;

[0021] Figure 4 This is a flowchart for identifying garbage types provided by this application. DETAILED DESCRIPTION

[0022] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention. Example 1

[0023] like Figure 1 As shown, the present application provides a garbage classification behavior data visualization analysis and management system, including: a classification garbage bin 110 and a back-end server 120; wherein, one back-end server 120 serves multiple classification garbage bins 110; the back-end server 120 is arranged in a control center, and the communication module 118 of the classification garbage bin 110 is communicated with the back-end server 120.

[0024] A classified trash can 110 is placed in a residential area in a city or village; the classified trash can 110 includes: a trash squeezing device 111, a classified display platform 112, a pressure acquisition sensor unit 113, a thickness acquisition sensor 114, an image acquisition device 115, a trash segmentation device 116, a control module 117, a communication module 118 and multiple classified compartments 119.

[0025] The garbage squeezing device 111 has An extrusion surface of uniform size, is the number of rows of extruded surfaces, is the number of columns of the extrusion surface. On each extrusion surface, a pressure sensing array is formed by a plurality of pressure acquisition sensing units 113 (for example, piezoresistive sensing units). The distribution density of the pressure acquisition sensing units 113 on the middle extrusion surface is multiple times that of the pressure acquisition sensing units 113 on the edge extrusion surfaces, preferably 4 times in this application, to adapt to the density of garbage at the center and the density of garbage at the edge after the garbage is squeezed and spread out. Among them, the extrusion surface in the middle is all the extrusion surfaces opposite to the area directly below the delivery window of the classified trash bin 110. Each extrusion surface is also configured with a corresponding thickness acquisition sensor 114, which can be a laser rangefinder (with an accuracy of ), half of the product of the flight time of the laser rangefinder light and the propagation speed of light is taken as the measured thickness.

[0026] Example 2

[0027] like Figure 2 As shown, this application provides a method for visual analysis and management of garbage classification behavior data, including the following steps:

[0028] Step S210: All the squeezing surfaces of the classified trash bin squeeze the thrown trash, collect the pressure data and thickness data corresponding to each squeezing surface, and send the pressure data and thickness data to the backend server;

[0029] When residents put out their garbage, they place it through the drop-in window of the classified trash bin 110 onto the classified display platform 112 located inside the classified trash bin 110. After the garbage is placed on the classified display platform 112, the control module 117 controls each squeezing surface of the garbage squeezing device 111 to squeeze the garbage on the classified display platform 112, thereby spreading the garbage on the classified display platform 112. Since the classified trash bins 110 in residential areas generally contain recyclable garbage, hazardous waste, kitchen waste, and other garbage, which are not very large in size, there is sufficient area on the classified display platform 112 for spreading the garbage.

[0030] Furthermore, the control module 117 controls all the pressure collection sensing units 113 constituting the pressure sensing array on each extrusion surface to collect pressure data at predetermined time intervals. ,

[0031] in, For the Rank The pressure data collected by the first pressure collection sensor unit 113 on the extrusion surface of the column, For the Rank The pressure data collected by the second pressure collection sensor unit 113 on the extrusion surface of the column, For the Rank The first row on the extrusion surface The pressure data collected by the pressure collection sensor unit 113, The value range is 1 to , The value range is 1 to , The value range is 1 to , For the Rank At the same time, the control module 117 controls each thickness collection sensor 114 to collect the thickness data of the garbage under the corresponding extrusion surface according to the predetermined time interval. ,in, For the Rank Thickness data of garbage under the extrusion surface of the column.

[0032] When collecting pressure data and thickness data After that, the control module 117 controls the communication module 118 to immediately transmit the collected pressure data and thickness data Transmitted to the backend server 120.

[0033] Step S220: After receiving the pressure data and thickness data corresponding to each extrusion surface, the backend server determines the instantaneous pressure-thickness change rate corresponding to each extrusion surface based on the pressure data and thickness data corresponding to each extrusion surface;

[0034] like Figure 3 As shown, step S220 includes the following sub-steps:

[0035] Step S221: judging whether there is a large abnormal value in the pressure data corresponding to each extrusion surface based on the sum of the mean value of the pressure data corresponding to each extrusion surface received this time and the standard deviation of a predetermined multiple;

[0036] The backend server 120 receives the pressure data and thickness data , it will immediately determine whether there is a large abnormal value in all the pressure data corresponding to each extrusion surface based on the mean and multiple standard deviation of all the pressure data corresponding to each extrusion surface received this time. For example: if the pressure data , then the pressure data is considered is a large outlier, which is expressed as , For the Rank The pressure data collected by the pressure collection sensing unit 113 on the extrusion surface of the column Otherwise, there is no extreme outlier.

[0037] Step S222: If there is no extreme outlier, determine the instantaneous pressure-thickness change rate of the extrusion surface based on the change in pressure data and thickness data corresponding to the extrusion surface;

[0038] If there is no extreme outlier, it means that the type of garbage squeezed by the extrusion surface is consistent. Then the backend server 120 determines the instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the change in all pressure data and thickness data corresponding to the extrusion surface received this time. ,

[0039] in, For the Rank The instantaneous pressure-thickness change rate of the extrusion surface of the column, for The differential representation of The amount of change, for The differential representation of The amount of change.

[0040] Step S223: If there is a significant anomaly, determine the normal instantaneous pressure-thickness change rate of the extrusion surface based on the change in normal pressure data and the change in corresponding thickness data corresponding to the extrusion surface;

[0041] If there is a large anomaly, it means that the type of garbage squeezed by the extrusion surface is inconsistent. Then the back-end server 120 determines the normal instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the change in all normal pressure data corresponding to the extrusion surface to the change in the corresponding thickness data. ,

[0042] in, is the number of extremely large outliers.

[0043] Step S224: Correcting the normal instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the maximum outlier value to the pressure data and the ratio of the number of maximum outliers to the number of pressure data corresponding to the extrusion surface to obtain the instantaneous pressure-thickness change rate of the extrusion surface;

[0044] Furthermore, the backend server 120 corrects the normal instantaneous pressure-thickness change rate of the extrusion surface according to the ratio of all the maximum abnormal values ​​to all the pressure data and the ratio of the number of all the maximum abnormal values ​​to the number of all the pressure data corresponding to the extrusion surface. , the instantaneous pressure-thickness change rate of the extrusion surface is expressed as ,

[0045] ,

[0046] in, is the influence weight of the number of extremely large outliers on the correction of instantaneous pressure-thickness change rate, is the influence weight of the maximum outlier on the correction of the instantaneous pressure-thickness change rate, and are all decimals between 0 and 1, and .

[0047] Step S230: The backend server determines an instruction for the corresponding extrusion surface to stop extrusion based on the relationship between the instantaneous pressure-thickness change rate and a predetermined value, and sends the instruction for the corresponding extrusion surface to stop extrusion to the classified trash can;

[0048] When obtaining the instantaneous pressure-thickness change rate of the extrusion surface After that, the backend server 120 determines the instantaneous pressure-thickness change rate of the extrusion surface Is the instantaneous pressure-thickness change rate on the extrusion surface greater than the preset value? When it is greater than a predetermined value, for example: When the back-end server 120 sends an instruction to the communication module 118 of the classified trash can 110 to stop squeezing the corresponding squeezing surface.

[0049] By correcting the normal instantaneous pressure-thickness change rate of the extrusion surface, when the types of garbage extruded by the extrusion surface are inconsistent, the instantaneous pressure-thickness change rate can be appropriately reduced according to the parameters of the harder garbage. This can increase the actual pressure when the extrusion surface stops extruding, reduce the impact of harder garbage on the degree of extrusion of other garbage, and ensure that the actual pressure when the extrusion surface stops extruding is not too large, thereby avoiding damage to the garbage extrusion device 111 and the corresponding pressure collection sensor unit 113.

[0050] After receiving the pressure data and thickness data next time, the back-end server 120 will continue to perform the above operations (step S210, step S220 and step S230) until the back-end server 120 sends the instructions to control all extrusion surfaces to stop extrusion to the communication module 118 of the classified trash can 110.

[0051] Step S240: The classified trash can receives the instruction to stop squeezing and stops squeezing the corresponding squeezing surface;

[0052] In response to receiving the command to stop squeezing, the control module 117 controls the corresponding squeezing surface of the garbage squeezing device 111 to immediately stop squeezing. The pressure sensing unit 113 and thickness sensing unit 114 corresponding to the squeezing surface that has stopped squeezing will stop collecting corresponding data; the pressure sensing unit 113 and thickness sensing unit 114 corresponding to the squeezing surface that has not stopped squeezing will continue collecting data.

[0053] Step S250: In response to all the squeezing surfaces stopping squeezing, the classified trash can moves all the squeezing surfaces away from the squeezing and spreading trash, collects image data of the spread trash, and sends the collected image data to the backend server;

[0054] After all the squeezing surfaces of the garbage squeezing device 111 stop squeezing, the garbage is considered to have been squeezed and spread out. Therefore, after a certain period of time, the control module 117 controls all the squeezing surfaces of the garbage squeezing device 111 to move away from the squeezed and spread out garbage. The control module 117 also controls the image acquisition device 115 to capture image data of the spread out garbage and transmits the captured image data to the backend server 120 via the communication module 118. The image acquisition device 115 is a high-resolution line array camera (for example, 2048 pixels, 0.1 mm / pixel), and the captured image data includes surface texture and color.

[0055] Step S260: The backend server receives the image data and identifies the type of garbage under each extrusion surface based on the image data;

[0056] like Figure 4 As shown, step S260 includes the following sub-steps:

[0057] Step S261: Divide the image data into multiple image data units according to the arrangement rule of the extrusion surface, and pre-process each image data unit;

[0058] After receiving the collected image data, the backend server 120 divides the image data into multiple image data units according to the arrangement rules of the extrusion surface. ,in For the Rank The image data unit corresponding to the extrusion surface of the column, The value range is 1 to , The value range is 1 to Then, for each image data unit All preprocessing is performed, such as normalization, cropping, enhancement, etc., to obtain the corresponding preprocessed image data units .

[0059] Step S262: performing multiple convolutions and downsampling on each preprocessed image data unit to obtain feature maps of different scales;

[0060] Each preprocessed image data unit All passed Sub-convolution and downsampling both obtain feature maps of different scales ,

[0061] in, is the preprocessed image data unit After the first convolution and downsampling, the feature map is obtained. is the preprocessed image data unit After the second convolution and downsampling, the feature map is obtained. is the preprocessed image data unit After the Sub-convolution and downsampling obtain feature maps; feature maps The size of the preprocessed image data unit of .

[0062] Step S263: performing convolution operations on the feature maps of different scales to obtain corresponding new feature maps, and fusing the new feature maps of different scales;

[0063] Feature maps at different scales Perform convolution operation on it to get a new feature map , where the convolution operation follows the formula ,in is the convolution kernel, is the convolution operation, is the bias term of the convolution operation. Perform fusion to obtain the fused feature map , is the number of convolution and downsampling.

[0064] Step S264: Input the fused feature map into a classifier to obtain a probability distribution of the fused feature map belonging to different types of garbage, so as to identify the type of garbage in the image data unit corresponding to the corresponding extrusion surface;

[0065] The fused feature map is input into the classifier to obtain the probability distribution of the fused feature map belonging to different types of garbage. The garbage type corresponding to the maximum distribution probability is used as the type of garbage in the image data unit corresponding to the corresponding extrusion surface. The classifier is represented as ,

[0066] The weight of the classifier is , the bias of the classifier is , The function is used to convert the input into the probability distribution of each category. is the output, i.e. the probability distribution of each category.

[0067] Step S270: The backend server determines the classification evaluation of the garbage according to the type of garbage under each squeeze surface, and sends the type of garbage under each squeeze surface and the classification evaluation to the classification garbage bin;

[0068] After the back-end server 120 determines the type of garbage under each extrusion surface, it sends the type of garbage under each extrusion surface to the communication module 118; and the back-end server 120 also uses the ratio of the maximum number of extrusion surfaces that squeeze the same type of garbage to the number of all extrusion surfaces that squeeze the garbage as the residents' classification evaluation of the garbage, and also sends the classification evaluation to the communication module 118.

[0069] The calculation formula for classification evaluation is as follows:

[0070] ;

[0071] in, For classification evaluation; For the first type, For the second type, For the type; is the number of extruded surfaces of garbage belonging to the first category, The number of extruded surfaces belonging to the second category of garbage, For garbage belongs to The number of extrusion faces of the type, For garbage belongs to The number of extrusion surfaces of the type; the number of types described for the waste; To take the maximum value.

[0072] Step S280: The classified trash bin places the garbage under each extrusion surface into the corresponding classification compartment according to the type of garbage under each extrusion surface, and displays the classification evaluation through graphics and / or voice;

[0073] The classification evaluation is received in the classification trash can 110, and the control module 117 displays the classification evaluation to the user through graphics and / or voice to enhance the willingness and execution of residents in garbage classification, thereby achieving effective promotion and popularization of garbage classification; and the control module 117 controls the garbage segmentation device 116 according to the type of garbage under each extrusion surface to put the garbage under each extrusion surface into the corresponding classification compartment 119, thereby achieving refined management of garbage classification behavior, thereby achieving effective promotion and popularization of garbage classification.

[0074] In addition, since this application arranges the operations of determining and generating stop squeezing instructions, identifying garbage types, and calculating classification evaluations on the back-end server 120, and the back-end server 120 has a larger memory, the execution speed of the overall system is also guaranteed.

[0075] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

[0076] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A method for visual analysis and management of garbage classification behavior data, applied to the backend server side, characterized in that: The steps include: In response to receiving the pressure data and the thickness data, determining an instantaneous pressure-thickness change rate corresponding to a corresponding extrusion surface of the classified trash can according to the pressure data and the thickness data; The garbage squeezing device has An extrusion surface of uniform size, is the number of rows of extruded surfaces, is the number of columns of the extrusion surface. On each extrusion surface, a pressure sensing array is formed by multiple pressure acquisition sensing units, and each extrusion surface is equipped with a corresponding thickness acquisition sensor; Each squeezing surface of the garbage squeezing device squeezes the garbage placed on the classification display platform to spread the garbage on the classification display platform, and all pressure collection sensor units constituting the pressure sensor array on each squeezing surface collect pressure data, and each thickness collection sensor collects thickness data of the garbage under the corresponding squeezing surface; Determining the instantaneous pressure-thickness change rate includes the following sub-steps: Determine whether there is a large abnormal value in the pressure data corresponding to the corresponding extrusion surface based on the sum of the mean value of the pressure data corresponding to each extrusion surface and the standard deviation of a predetermined multiple; If it does not exist, the type of garbage squeezed by the extrusion surface is consistent, and the instantaneous pressure-thickness change rate of the extrusion surface is determined according to the change in pressure data and thickness data corresponding to the extrusion surface; If it exists, the types of garbage squeezed by the squeezing surface are inconsistent, and the normal instantaneous pressure-thickness change rate of the squeezing surface is determined according to the change in the normal pressure data and the corresponding thickness data corresponding to the squeezing surface; Correcting the normal instantaneous pressure-thickness change rate of the extrusion surface based on the ratio of the maximum outlier value to the pressure data and the ratio of the number of maximum outliers to the number of pressure data corresponding to the extrusion surface to obtain the instantaneous pressure-thickness change rate of the extrusion surface; By correcting the normal instantaneous pressure-thickness change rate of the extrusion surface, when the types of garbage squeezed by the extrusion surface are inconsistent, the instantaneous pressure-thickness change rate is reduced according to the parameters of the harder garbage, thereby increasing the actual pressure when the extrusion surface stops squeezing; According to the relationship between the instantaneous pressure-thickness change rate and the predetermined value, an instruction to stop extrusion of the corresponding extrusion surface is obtained and sent to the classified trash can, so that the corresponding extrusion surface stops extrusion; In response to receiving the image data, identifying and transmitting to the classified trash bin the type of garbage under each squeezing surface according to the image data, so that the classified trash bin deposits the corresponding type of garbage into the corresponding classified compartment according to the type, wherein the image data is collected after all squeezing surfaces stop squeezing and are away from the garbage; The classification evaluation of the garbage is determined according to the type of garbage and sent to the classification garbage bin so that the classification garbage bin displays the classification evaluation.

2. The method for visual analysis and management of garbage classification behavior data according to claim 1 is applied to the backend server side, characterized in that: Identifying the types of garbage includes the following sub-steps: Dividing the image data into a plurality of image data units according to the arrangement rule of the extrusion surface, and preprocessing each image data unit; Each preprocessed image data unit is subjected to multiple convolutions and downsampling to obtain feature maps of different scales; Perform convolution operations on feature maps of different scales to obtain corresponding new feature maps, and then fuse the new feature maps of different scales; The fused feature map is input into the classifier to obtain the probability distribution of the fused feature map belonging to different types of garbage, so as to identify the type of garbage in the image data unit corresponding to the corresponding extrusion surface.

3. A method for visual analysis and management of garbage classification behavior data, applied to the classification garbage bin side, characterized in that: The steps include: The garbage squeezing device has An extrusion surface of uniform size, is the number of rows of extruded surfaces, is the number of columns of the extrusion surface. On each extrusion surface, a pressure sensing array is formed by multiple pressure acquisition sensing units, and each extrusion surface is equipped with a corresponding thickness acquisition sensor; Each squeezing surface of the garbage squeezing device squeezes the garbage placed on the classification display platform to spread the garbage on the classification display platform, and all pressure collection sensor units constituting the pressure sensor array on each squeezing surface collect pressure data, and each thickness collection sensor collects thickness data of the garbage under the corresponding squeezing surface, and transmits the pressure data and thickness data corresponding to each squeezing surface to the back-end server, so that the back-end server determines the instantaneous pressure-thickness change rate corresponding to the corresponding squeezing surface; Determining the instantaneous pressure-thickness change rate includes the following sub-steps: Determine whether there is a large abnormal value in the pressure data corresponding to the corresponding extrusion surface based on the sum of the mean value of the pressure data corresponding to each extrusion surface and the standard deviation of a predetermined multiple; If it does not exist, the type of garbage squeezed by the extrusion surface is consistent, and the instantaneous pressure-thickness change rate of the extrusion surface is determined according to the change in pressure data and thickness data corresponding to the extrusion surface; If it exists, the types of garbage squeezed by the squeezing surface are inconsistent, and the normal instantaneous pressure-thickness change rate of the squeezing surface is determined according to the change in the normal pressure data and the corresponding thickness data corresponding to the squeezing surface; Correcting the normal instantaneous pressure-thickness change rate of the extrusion surface according to the ratio of the maximum outlier value to the pressure data and the ratio of the number of maximum outliers to the number of pressure data corresponding to the extrusion surface to obtain the instantaneous pressure-thickness change rate of the extrusion surface; By correcting the normal instantaneous pressure-thickness change rate of the extrusion surface, when the types of garbage squeezed by the extrusion surface are inconsistent, the instantaneous pressure-thickness change rate is reduced according to the parameters of the harder garbage, thereby increasing the actual pressure when the extrusion surface stops squeezing; In response to receiving a stop extrusion instruction, stopping extrusion of the corresponding extrusion surface, wherein the stop extrusion instruction is obtained based on a relationship between the instantaneous pressure-thickness change rate and a predetermined value; In response to all the squeezing surfaces stopping squeezing, all the squeezing surfaces are moved away from the squeezed garbage, and image data of the garbage is collected and sent to a back-end server, so that the back-end server identifies the type of garbage under each squeezing surface based on the image data; In response to receiving the classification evaluation and type of garbage, the corresponding type of garbage is placed into the corresponding classification compartment according to the type of garbage, and the classification evaluation is displayed, wherein the classification evaluation is determined by the back-end server according to the type of garbage under each extrusion surface.

4. The method for visual analysis and management of garbage classification behavior data according to claim 3 is applied to the classification garbage bin side, characterized in that: Identifying the types of garbage includes the following sub-steps: Dividing the image data into a plurality of image data units according to the arrangement rule of the extrusion surface, and preprocessing each image data unit; Each preprocessed image data unit is subjected to multiple convolutions and downsampling to obtain feature maps of different scales; Perform convolution operations on feature maps of different scales to obtain corresponding new feature maps, and then fuse the new feature maps of different scales; The fused feature map is input into the classifier to obtain the probability distribution of the fused feature map belonging to different types of garbage, so as to identify the type of garbage in the image data unit corresponding to the corresponding extrusion surface.

5. A garbage classification behavior data visualization analysis and management system, characterized by: include: Classification trash can and backend server; The back-end server executes the method described in claim 1 or 2 above, and the classified trash can executes the method described in claim 3 or 4 above.

6. The garbage classification behavior data visualization analysis and management system according to claim 5 is characterized in that: The distribution density of the pressure acquisition sensor units on the middle extrusion surface is multiple times the distribution density of the pressure acquisition sensor units on the edge extrusion surface.

Images