An internet of things data processing system based on intelligent garbage classification

Abstract

The present application relates to garbage classification technical field, especially to a kind of based on wisdom garbage classification Internet of Things data processing system.The system includes: parameter measurement module, it is to measure the garbage volume and recovery volume and garbage weight and dry weight of single time loading garbage classification device's kitchen garbage;Image acquisition module, it is to collect garbage initial image, smoke image, drying product image and dispersion product image;Garbage processing module, it is to determine whether to dry kitchen garbage;Image recognition module, it is to carry out image segmentation and image recombination to drying product image and generate recombination image and clean feature garbage;Analysis processing module, it is to determine whether to carry out secondary drying.The present application utilizes above-mentioned each module cooperation, in effectively improving garbage foreign matter identification efficiency, further improve the accuracy of based on wisdom garbage classification Internet of Things data processing system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of garbage classification, and particularly relates to an Internet of Things data processing system based on intelligent garbage classification. BACKGROUND

[0002] In recent years, in the practice of garbage classification, the classification of kitchen waste has always been a difficult point. Due to the residents' vague cognition of the classification standard, a large number of sundries not belonging to the kitchen waste category are often mixed in, such as used paper towels, broken plastic bags, expired medicine packaging and even broken ceramic fragments. These other garbage and the real kitchen waste are mixed together to form a headache mixed garbage. This classification error directly leads to two chain problems: on the one hand, when these mixed kitchen waste with foreign matters are concentrated and poured, irregular stacking is easily formed in the garbage can, which adds an additional burden to the subsequent cleaning and transportation work; on the other hand, after entering the processing link, these mixed other garbage will seriously interfere with the resource utilization process of kitchen waste.

[0003] Chinese patent application publication No. CN118429733A discloses a multi-head attention driven kitchen waste multi-label classification method. The application discloses a multi-head attention driven kitchen waste multi-label classification method, which comprises the following steps: constructing a kitchen waste multi-label classification data set, including a plurality of images of different categories of kitchen waste, and the image label includes one or more categories; constructing a multi-head attention driven graph convolution lightweight network model, including a feature extraction module, a multi-head attention module and a dynamic graph convolution module; wherein the feature extraction module extracts features from the input image, and then sends the features to the multi-head attention module to strengthen the category perception area of the feature map, and then sends the features to the dynamic graph convolution module to adaptively capture the category perception area; training the graph convolution lightweight network model using the constructed kitchen waste multi-label classification data set; finally, using the trained classification model to perform multi-label classification on the kitchen waste image to be predicted. The application enhances the recognition ability and improves the multi-label classification effect while reducing the performance loss caused by the reduction of model parameters.

[0004] However, the above method has the following problems: it fails to accurately identify and effectively process a large number of stacked kitchen waste rich in moisture. SUMMARY

[0005] Therefore, the present application provides an Internet of Things data processing system based on intelligent garbage classification to overcome the problem that the prior art fails to accurately identify and effectively process a large number of stacked kitchen waste rich in moisture.

[0006] To achieve the above purpose, the present application provides an Internet of Things data processing system based on intelligent garbage classification, which comprises:

[0007] a parameter measurement module including a volume detection unit configured to measure a garbage volume of the kitchen garbage and a characteristic garbage volume of the characteristic garbage, and a weight detection unit configured to measure a garbage weight of the kitchen garbage and a dried weight of the kitchen garbage after drying;

[0008] an image acquisition module configured to acquire a garbage initial image of the kitchen garbage, a smoke image during drying, a drying product image of a first product generated at the end of drying, and a dispersion product image of a second product generated at the end of dispersion;

[0009] a garbage processing module connected to the parameter measurement module and the image acquisition module, and including a drying unit configured to determine whether to dry the kitchen garbage based on the garbage volume and the garbage initial image, and a dispersion unit configured to disperse the first product;

[0010] an image recognition module connected to the image acquisition module and the garbage processing module, and including an image processing unit configured to perform image segmentation and image recombination on the drying product image to generate a recombined image, an image screening unit configured to screen a characteristic image based on the recombined image, and an image comparison unit configured to compare the characteristic image and the dispersion product image, and determine whether to clean the characteristic garbage corresponding to the characteristic image based on a comparison result.

[0011] Further, the system further includes:

[0012] an analysis processing module connected to the parameter measurement module, the image acquisition module, the garbage processing module, and the image recognition module, and including a compression processing unit configured to determine whether to perform secondary drying and compression based on a result of determining not to clean the characteristic garbage in combination with the smoke image, a data uploading unit configured to clean and recycle the characteristic garbage based on a result of determining to clean the characteristic garbage, generate a characteristic value according to a volume difference and a weight difference, and determine whether to compress the recombined image into a plurality of data packets and transmit the data packets to a characteristic image library based on the characteristic value.

[0013] Further, the image processing unit is configured to perform image segmentation and image recombination on the drying product image, wherein,

[0014] the drying product image is segmented into a plurality of unit regions based on semantic and instance segmentation;

[0015] a plurality of edge seed points are arranged at edges of the unit regions;

[0016] similar pixels adjacent to the edge seed points are identified;

[0017] generating a reorganization image by merging the regions corresponding to the similar pixel pairs into the unit region.

[0018] Further, the image screening unit screens a feature image based on the reorganization image, wherein

[0019] determining a feature boundary line of the reorganization image;

[0020] calculating image similarity between the feature boundary line and image boundary lines of a plurality of existing feature images in a feature image library;

[0021] screening an existing feature image corresponding to a maximum value of the image similarity as the feature image.

[0022] Further, the image comparison unit generates a comparison result based on the feature image and the dispersion product image, wherein

[0023] performing semantic segmentation and instance segmentation on the feature image to generate a feature instance;

[0024] performing semantic segmentation and instance segmentation on the dispersion product image to generate a plurality of dispersion instances, respectively;

[0025] sequentially calculating coincidence degrees between the feature instance and the plurality of dispersion instances.

[0026] Further, the image comparison unit determines whether to clean up a feature garbage corresponding to the feature image based on a comparison result of the coincidence degree and a preset coincidence degree, wherein

[0027] if the coincidence degree is greater than or equal to the preset coincidence degree, it is determined to clean up the feature garbage corresponding to the feature image;

[0028] the preset coincidence degree is positively correlated with a garbage volume of the kitchen garbage.

[0029] Further, the compression processing unit, in response to a determination result that the feature garbage is not cleaned up, identifies whether the smoke in the smoke image is white, wherein

[0030] if the smoke is white, it is determined to compress the feature garbage after secondary drying;

[0031] if the smoke is not white, it is determined to compress the feature garbage.

[0032] Further, the data uploading unit, in response to a determination result that the feature garbage is cleaned up, cleans up and recycles the feature garbage, calculates a difference between the garbage volume and a recycling volume as a volume difference, and calculates a difference between the garbage weight and a drying weight as a weight difference.

[0033] Further, the data uploading unit compares the feature value with a preset feature value, and determines whether to compress the reorganized image into a plurality of data packets and transmit to a feature image library according to a comparison result.

[0034] If the feature value is less than or equal to the preset feature value, it is determined to compress the reorganized image into a plurality of data packets and transmit to the feature image library.

[0035] The preset feature value is positively correlated with the garbage volume of the kitchen garbage.

[0036] Further, the data uploading unit is used to compress the reorganized image into a plurality of data packets and transmit to the feature image library.

[0037] The image boundary line corresponding to the reorganized image is compressed into a first data packet.

[0038] The feature instance corresponding to the reorganized image is compressed into a second data packet.

[0039] The reorganized image, the first data packet and the second data packet are sequentially transmitted to the feature image library.

[0040] Compared with the prior art, the beneficial effects of the present application are that the kitchen garbage loaded into the garbage classification device is dried and dispersed after drying, and the moisture in the kitchen garbage is fully removed. At the same time, the kitchen garbage in the stacked state is dispersed and processed, the components of the kitchen garbage can be clearly identified, the efficiency of subsequent analysis and processing of the kitchen garbage is improved, the drying process can fully remove the moisture in the kitchen garbage, greatly reduce the weight and volume of the garbage, prolong the processing window period of the kitchen garbage, reduce the pollution of harmful leachate generated by corruption to the environment, and the kitchen garbage in the stacked state will block each other and be difficult to distinguish the specific components. After dispersion, various substances can be separated, which is convenient for equipment to clearly identify, lays a foundation for subsequent classification and analysis, reduces detection errors and repeated operations caused by moisture interference and material accumulation, thereby shortens the analysis time, improves the overall processing efficiency, and effectively improves the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0041] Further, the present application determines whether the drying temperature of the kitchen garbage is too high by combining the collection of smoke during the drying process of the kitchen garbage, and if the drying temperature of the kitchen garbage is too high during the first drying process, it indicates that the kitchen garbage that does not need foreign matter cleaning is not suitable for secondary drying, preventing the generation of toxic gases at high temperatures, and under the condition that the kitchen garbage that does not need foreign matter cleaning is determined to be secondary dried, secondary drying helps to further reduce the moisture content of the kitchen garbage, and helps subsequent processing, after cleaning and recycling the kitchen garbage that needs foreign matter cleaning, the classification quality of the kitchen garbage is evaluated according to the volume change value of the kitchen garbage after cleaning and the weight difference before and after drying, and the reorganized image of the kitchen garbage during the processing process is uploaded to the feature image library for comparison in the next identification process, further improving the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0042] Further, the present application screens feature images based on reorganized images by image segmentation and image reorganization of the drying product image, although the state of the kitchen garbage after drying is more stable, there may still be impurities or adhesion of different components, through image segmentation technology, the target kitchen component can be separated from the interference and background area, excluding irrelevant information interference, the segmented image may have information fragmentation problems due to local occlusion, angle deviation, etc., image reorganization restores the complete form of the object or corrects the angle deviation through splicing, so that subsequent analysis can be based on a relatively complete sample, reducing the feature extraction error caused by incomplete form, the image after segmentation, reorganization and screening has less noise and more prominent features, which can enable the model to more accurately learn the essential differences between different kitchen garbage, thereby reducing the classification error rate, the reorganized image contains a large amount of information, but not all areas are valuable for classification, by screening feature images, the key area that best represents the material properties can be extracted, focusing the analysis on the part with significant classification features, avoiding redundant information consuming computing resources or interfering with judgment, and further improving the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0043] Further, the present application compares the feature image and the dispersed product image after dispersion, and further determines whether there is a foreign matter that needs to be treated in the kitchen waste on the premise of screening the foreign matter in the kitchen waste according to the existing feature image database. The traditional foreign matter identification may depend on artificial visual inspection, and is affected by subjective experience, fatigue and other factors, and is prone to missed judgment or misjudgment. Through image comparison, the system can quantitatively determine the foreign matter based on a preset feature matching algorithm. The whole process has a unified standard, reduces the uncertainty caused by human intervention. If foreign matter is mixed into the kitchen waste, the kitchen waste composition data counted by the Internet of Things system will be distorted, and then the subsequent garbage collection and transportation scheduling and resource treatment scheme will be affected. After the comparison and identification and the removal of the foreign matter, the kitchen waste data obtained by the system is more pure, which provides accurate basic data for garbage classification statistics and processing technology optimization, and further improves the accuracy of the Internet of Things data processing system based on intelligent garbage classification. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 It is a structural block diagram of the Internet of Things data processing system based on intelligent garbage classification of the present application;

[0045] Figure 2 It is a structural block diagram of the analysis and processing module of the present application;

[0046] Figure 3 It is a logic diagram for determining whether to clean the feature garbage corresponding to the feature image;

[0047] Figure 4 It is a logic diagram for determining whether to compress the reorganized image into a plurality of data packets and transmit to the feature image library. DETAILED DESCRIPTION

[0048] In order to make the purpose and advantages of the present application clearer and more apparent, the present application will be further described below in conjunction with examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0049] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application, and are not used to limit the protection scope of the present application.

[0050] It should be noted that in the description of the present application, the terms "up", "down", "left", "right", "in", "out" and the like indicate the direction or positional relationship terms based on the direction or positional relationship shown in the drawings, which are only for the convenience of description, and do not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0051] Moreover, it needs to be explained that in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connecting" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected, it can be mechanical connection, or electrical connection, it can be directly connected, or indirectly connected through intermediate medium, it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0052] Please refer to Figure 1 As shown in the structure block diagram of the Internet of Things data processing system based on intelligent garbage classification of the present application, the present application provides an Internet of Things data processing system based on intelligent garbage classification, comprising:

[0053] The parameter measurement module comprises a volume detection unit for measuring the garbage volume of the kitchen garbage loaded into the garbage classification device at a time and the recycling volume of the characteristic garbage, and a weight detection unit for measuring the garbage weight of the kitchen garbage and the dry weight after drying;

[0054] The image acquisition module is used to acquire the garbage initial image of the kitchen garbage, the smoke image during drying and the drying product image of the first product generated after drying, and the dispersion product image of the second product generated after dispersion;

[0055] The garbage processing module is connected with the parameter measurement module and the image acquisition module respectively, and comprises a drying unit for determining whether to dry the kitchen garbage based on the garbage volume and the garbage initial image, and a dispersion unit for dispersing the first product;

[0056] The image recognition module is connected with the image acquisition module and the garbage processing module respectively, and comprises an image processing unit for image segmentation and image recombination of the drying product image to generate a recombined image, an image screening unit for screening a characteristic image based on the recombined image, and an image comparison unit for comparing the characteristic image and the dispersion product image, and determining whether to clean the characteristic garbage corresponding to the characteristic image based on the comparison result.

[0057] Specifically, the kitchen garbage loaded into the garbage classification device can be weighed by a bottom weighing sensor or a suspended scale, the surface of the garbage can be scanned by a 3D camera to obtain point cloud data, and then the 3D model of the garbage is reconstructed by an algorithm to calculate the volume of the model. For those skilled in the art, all of the above are prior art, and will not be described here.

[0058] Specifically, the present application can fully remove the moisture in the kitchen waste by drying and dispersing the kitchen waste loaded into the garbage classification device, and can clearly identify the components of the kitchen waste by dispersing the stacked kitchen waste, thereby improving the efficiency of subsequent analysis and processing of the kitchen waste. The drying process can fully remove the moisture in the kitchen waste, greatly reduce the weight and volume of the garbage, extend the processing window of the kitchen waste, reduce the pollution of harmful leachate generated by corruption to the environment, and reduce the pollution of harmful leachate generated by corruption to the environment. The stacked kitchen waste blocks each other and is difficult to distinguish the specific components. After the dispersion treatment, the various substances can be separated, which is convenient for the equipment to clearly identify, lays the foundation for subsequent classification and analysis, reduces the detection error and repeated operation caused by moisture interference and material accumulation, thereby shortens the analysis time, improves the overall processing efficiency, and effectively improves the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0059] Referring to Figure 2 As shown in the figure, it is a structural block diagram of the analysis and processing module of the present application, and further comprises:

[0060] The analysis and processing module is connected with the parameter measurement module, the image acquisition module, the garbage processing module and the image recognition module respectively, and comprises a compression processing unit for determining whether to perform secondary drying and compression based on the determination result of not cleaning the characteristic garbage combined with the smoke image, a data uploading unit for cleaning and recycling the characteristic garbage based on the determination result of cleaning the characteristic garbage, generating a characteristic value according to the volume difference and the weight difference, and determining whether to compress the recombined image into a plurality of data packets and transmit them to the characteristic image library based on the characteristic value.

[0061] Specifically, the present application combines the collection of smoke during the drying process of the kitchen waste, determines whether the drying temperature of the kitchen waste is too high through the state of the smoke, and if the drying temperature of the kitchen waste is too high during the first drying process, it indicates that the kitchen waste which does not need to be cleaned of foreign matter is not suitable for secondary drying, thereby preventing the generation of toxic gases at high temperature. Under the condition that the kitchen waste which does not need to be cleaned of foreign matter is determined to be subjected to secondary drying, secondary drying helps to further reduce the moisture in the kitchen waste and helps subsequent processing. After cleaning and recycling the kitchen waste which needs to be cleaned of foreign matter, the classification quality of the kitchen waste is evaluated according to the volume change value of the kitchen waste after cleaning and the weight difference before and after drying, and the recombined image during the processing of the kitchen waste is uploaded to the characteristic image library for comparison in the next identification process, thereby further improving the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0062] Specifically, the image processing unit is used for image segmentation and image recombination of the drying product image, wherein,

[0063] to segment the drying product image into several unit regions based on semantic and instance segmentation;

[0064] to set several edge seed points at the edge of the unit region;

[0065] to identify similar pixels adjacent to the edge seed points;

[0066] to merge the regions corresponding to the similar pixels into the unit region to generate a reorganized image.

[0067] Specifically, the image processing unit identifies the text information in the drying product image, preliminarily divides the region range of different categories based on the text information, further distinguishes several unit regions of different individuals in the same category based on semantic segmentation, extracts the edge contour of the unit region, and uniformly selects several pixel points on the edge contour as edge seed points; defines that the pixels with an RGB difference less than 20 from the edge seed points are judged as similar pixels, and the similar pixels are included in the unit region, and this step is repeated until there are no adjacent pixels that meet the similar condition, and a reorganized image is generated. It can be understood that the color of the dried kitchen waste is mostly yellow-brown, light brown, gray-white and the like, the overall color tone is dull, and the color span is small. If the threshold is set too small (such as 5), the pixels of the same region will be misjudged as dissimilar due to slight color fluctuations, resulting in too fragmented region segmentation; if the threshold is set too large (such as 50), different category objects may be misjudged as similar, resulting in incorrect region merging; therefore, the threshold of 20 can balance the color fluctuations and category differences, and is suitable for the low contrast and small span color characteristics of the drying product.

[0068] Specifically, the image screening unit is used to screen a feature image based on the reorganized image, wherein,

[0069] to determine the feature boundary line of the reorganized image;

[0070] to calculate the image similarity of the feature boundary line and the image boundary line of several existing feature images in the feature image library;

[0071] to screen the existing feature image corresponding to the maximum image similarity as the feature image.

[0072] Specifically, the feature boundary line of the reorganized image is extracted, the feature boundary line and the image boundary line are subjected to scale normalization and rotation alignment, the feature boundary line and the image boundary line are placed in the same coordinate system, several parallel lines parallel to the Y-axis are divided along the Y-axis of the coordinate system at a preset distance, and the parallel lines intersect the feature boundary line and the image boundary line at several points, respectively. The similarity values of the horizontal coordinates and the vertical coordinates of the corresponding points are calculated, and the image similarity is obtained by adding them up.

[0073] It can be understood that the similarity value calculation formula of the abscissa and the ordinate is the similarity value MP of the abscissa (or ordinate) list YB=(YB1, YB2, …, YBj, …, YBm) of the feature boundary line point position and the abscissa (or ordinate) list EB=(EB1, EB2, …, EBj, …, EBm) of the image boundary line point position; wherein, j=1, 2, …, m;

[0074] The similarity value is MP=∑mj=1YBj*EBj / (sqrt(∑mj=1(YBj)2)*sqrt(∑mj=1(EBj)2)).

[0075] It can be understood that the preset distance can satisfy the intersection of the Y-axis parallel line and the feature boundary line and the image boundary line, and details are not repeated here.

[0076] Specifically, the application carries out image segmentation and image recombination on the dried product image, filters feature images based on the recombined image, and the kitchen waste is more stable after drying, but impurities or adhesion of different components may still exist. Through the image segmentation technology, the target kitchen component can be separated from the interference and the background area, the interference of irrelevant information is excluded, the segmented image may have information fragmentation problems due to local shielding, angle deviation, etc., the image recombination restores the complete form of the object or corrects the angle deviation through splicing, so that the subsequent analysis can be based on a relatively complete sample, the feature extraction error caused by incomplete form is reduced, the image after segmentation, recombination and filtering has less noise and more prominent features, so that the model can more accurately learn the essential differences of different kitchen wastes, thereby reducing the classification error rate. The recombined image contains a large amount of information, but not all regions are valuable for classification. By filtering the feature image, the key region representing the material properties can be extracted, the focus of analysis is concentrated on the part with significant classification features, redundant information consumption of calculation resources or interference judgment is avoided, and the accuracy of the Internet of Things data processing system based on intelligent garbage classification is further improved.

[0077] Please refer to Figure 3 As shown in the figure, it is a logic diagram for determining whether to clean the feature garbage corresponding to the feature image, the image comparison unit is used to generate a comparison result based on the feature image and the dispersion product image, wherein

[0078] The feature image is subjected to semantic segmentation and instance segmentation to generate a feature instance;

[0079] The dispersion product image is subjected to semantic segmentation and instance segmentation to generate a plurality of dispersion instances;

[0080] The feature instance and the plurality of dispersion instances are sequentially calculated for coincidence degree.

[0081] Specifically, the image comparison unit determines the foreign matter contained in the feature image, i.e., the feature instance, by semantic segmentation and instance segmentation, determines the foreign matter contained in the dispersion product image, i.e., the dispersion instance, performs scale normalization and rotation alignment on the feature instance and the dispersion instance, and the area ratio of the overlapping area of the two to the projection area of the two is the coincidence degree.

[0082] Specifically, the image comparison unit determines whether to clean the feature garbage corresponding to the feature image based on a comparison result of the coincidence degree and a preset coincidence degree, wherein,

[0083] If the coincidence degree is greater than or equal to the preset coincidence degree, it is determined to clean the feature garbage corresponding to the feature image.

[0084] If the coincidence degree is less than the preset coincidence degree, it is determined not to clean the feature garbage corresponding to the feature image.

[0085] In one specific embodiment, the coincidence degree is set to 90%, and if the coincidence degree is 98% greater than the preset coincidence degree, it is determined to clean the feature garbage corresponding to the feature image.

[0086] If the coincidence degree is 84% less than the preset coincidence degree, it is determined not to clean the feature garbage corresponding to the feature image.

[0087] The preset coincidence degree is positively correlated with the garbage volume of the kitchen waste.

[0088] It can be understood that the larger the garbage volume of the kitchen waste, the greater the probability of containing foreign matter, the higher the demand for foreign matter determination, and therefore the preset coincidence degree is positively correlated with the garbage volume of the kitchen waste.

[0089] Preferably, the garbage volume of the kitchen waste is 0.2 cubic meters, and the preset coincidence degree is 80%;

[0090] The garbage volume of the kitchen waste is 0.5 cubic meters, and the preset coincidence degree is 85%;

[0091] The garbage volume of the kitchen waste is 0.8 cubic meters, and the preset coincidence degree is 90%.

[0092] Specifically, the present application compares the feature image and the dispersed dispersion product image, further determines whether there is a foreign matter in the kitchen waste that needs to be treated on the premise of screening the foreign matter in the kitchen waste according to the existing feature image database, the traditional foreign matter identification may rely on artificial visual inspection, is affected by subjective experience, fatigue and other factors, and is prone to missed judgment or misjudgment, and through image comparison, the system can quantitatively determine the foreign matter based on a preset feature matching algorithm, the whole process has a unified standard, and the uncertainty caused by human intervention is reduced, if the foreign matter is mixed into the kitchen waste, the kitchen waste composition data counted by the Internet of Things system will be distorted, and then the subsequent garbage collection and transportation scheduling and resource treatment scheme will be affected, after the foreign matter is identified and removed through comparison, the kitchen waste data obtained by the system is more pure, which provides accurate basic data for garbage classification statistics, treatment process optimization and the like, and further improves the accuracy of the Internet of Things data processing system based on intelligent garbage classification.

[0093] Specifically, the compression processing unit determines whether the smoke in the smoke image is white in response to a condition of a determination result that the feature garbage is not cleaned.

[0094] If the smoke is white, it is determined to compress the feature garbage after secondary drying.

[0095] If the smoke is not white, it is determined to compress the feature garbage.

[0096] Specifically, the data uploading unit cleans and recycles the feature garbage in response to a condition of a determination result that the feature garbage is cleaned, calculates a difference between the garbage volume and the recycling volume as a volume difference, and calculates a difference between the garbage weight and the dry weight as a weight difference.

[0097] Specifically, the feature value=ax volume difference+bx weight difference, wherein a and b are weighting values, a+b=1, the weight loss caused by water evaporation has a greater impact on the overall feature value, so a is generally taken as 0.4, and b is taken as 0.6; the unit of the volume difference is cubic decimeter, and the unit of the weight difference is kilogram, only the numerical value is taken in the above formula calculation.

[0098] Referring to Figure 4 The figure is a logic diagram for determining whether to compress the reorganized image into a plurality of data packets and transmit to the feature image library, the data uploading unit compares the feature value with the preset feature value, and determines whether to compress the reorganized image into a plurality of data packets and transmit to the feature image library according to the comparison result, wherein,

[0099] If the feature value is less than or equal to the preset feature value, it is determined to compress the reorganized image into a plurality of data packets and transmit to the feature image library;

[0100] If the feature value is greater than the preset feature value, it is determined not to upload the reorganized image;

[0101] In one specific embodiment, the characteristic value is set to 100, if the characteristic value is 86, which is less than the preset characteristic value, it is determined that the reorganized image is compressed into a plurality of data packets and transmitted to the characteristic image library;

[0102] If the characteristic value is 142, which is greater than the preset characteristic value, it is determined that the reorganized image is not uploaded;

[0103] The preset characteristic value is positively correlated with the garbage volume of the kitchen garbage.

[0104] It can be understood that the larger the garbage volume of the kitchen garbage, the more water it contains, the greater the probability of containing foreign matter, and the larger the total volume of the foreign matter. Therefore, the preset characteristic value is positively correlated with the garbage volume of the kitchen garbage.

[0105] Preferably, the garbage volume of the kitchen garbage is 0.2 cubic meters, and the preset characteristic value is 60;

[0106] The garbage volume of the kitchen garbage is 0.5 cubic meters, and the preset characteristic value is 100;

[0107] The garbage volume of the kitchen garbage is 0.8 cubic meters, and the preset characteristic value is 180.

[0108] Specifically, the data uploading unit is used to compress the reorganized image into a plurality of data packets and transmit them to the characteristic image library, wherein,

[0109] The image boundary line corresponding to the reorganized image is compressed into the first data packet;

[0110] The characteristic instance corresponding to the reorganized image is compressed into the second data packet;

[0111] The reorganized image, the first data packet and the second data packet are sequentially transmitted to the characteristic image library.

[0112] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after such changes or replacements will fall within the protection scope of the present application.

[0113] The above description is only the preferred embodiments of the present application and is not intended to limit the present application; for those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An Internet of Things data processing system based on intelligent garbage classification, characterized in that, The kitchen waste classification device comprises: a parameter measurement module comprising a volume detection unit for measuring the volume of kitchen waste loaded into the kitchen waste classification device and the volume of characteristic kitchen waste, and a weight detection unit for measuring the weight of the kitchen waste and the dry weight after drying; an image acquisition module for acquiring an initial image of the kitchen waste, a smoke image during drying, a drying product image of the first product generated at the end of drying, and a dispersion product image of the second product generated at the end of dispersion; a waste treatment module connected to the parameter measurement module and the image acquisition module, comprising a drying unit for determining whether to dry the kitchen waste based on the volume of the kitchen waste and the initial image of the kitchen waste, and a dispersion unit for dispersing the first product; an image recognition module connected to the image acquisition module and the waste treatment module, comprising an image processing unit for image segmentation and image recombination of the drying product image to generate a recombined image, an image screening unit for screening a characteristic image based on the recombined image, and an image comparison unit for comparing the characteristic image and the dispersion product image to determine whether to clean the characteristic kitchen waste corresponding to the characteristic image based on the comparison result; the image processing unit identifies text information in the drying product image, divides the area range of different categories based on the text information, distinguishes a plurality of unit areas of different individuals in the same category based on semantic segmentation, extracts the edge contour of each unit area, and uniformly selects a plurality of pixel points on the edge contour as edge seed points; the pixel points with a RGB difference from the edge seed points less than a preset pixel threshold are defined as similar pixel points, the similar pixel points are included in the unit area, and the step is repeated until no adjacent pixel points meeting the similar condition are included in the unit area to generate a recombined image. 2.The Internet of Things data processing system based on intelligent garbage classification according to claim 1, wherein, Further comprising: an analysis processing module connected to the parameter measurement module, the image acquisition module, the waste treatment module, and the image recognition module, comprising a compression processing unit for determining whether to perform secondary drying and compression based on the smoke image and the determination result that the characteristic kitchen waste is not cleaned, a data uploading unit for cleaning and recycling the characteristic kitchen waste based on the determination result that the characteristic kitchen waste is cleaned, generating a characteristic value according to the volume difference and the weight difference, and determining whether to compress the recombined image into a plurality of data packets and transmit the data packets to a characteristic image library. 3.The Internet of Things data processing system based on intelligent garbage classification according to claim 1, characterized in that, The image processing unit is used to perform image segmentation and image recombination on the drying product image, wherein the drying product image is segmented into a plurality of unit areas based on semantics and instances; a plurality of edge seed points are set at the edges of the unit areas; similar pixels adjacent to the edge seed points are identified; the areas corresponding to the similar pixels are merged into the unit areas to generate a recombined image. 4.The Internet of Things data processing system based on intelligent garbage classification according to claim 3, characterized in that, The image screening unit is used to screen a characteristic image based on the recombined image, wherein a characteristic boundary line of the recombined image is determined; calculating image similarity between the feature boundary line and image boundary lines of a plurality of existing feature images in a feature image library; screening an existing feature image corresponding to the maximum image similarity as a feature image. 5.The Internet of Things data processing system based on intelligent garbage classification according to claim 4, characterized in that, The image comparison unit is configured to generate a comparison result based on the feature image and the dispersion product image, wherein performing semantic segmentation and instance segmentation on the feature image to generate a feature instance; performing semantic segmentation and instance segmentation on the dispersion product image to generate a plurality of dispersion instances, respectively; calculating the degree of overlap between the feature instance and the plurality of dispersion instances in sequence. 6.The Internet of Things data processing system based on intelligent garbage classification according to claim 5, characterized in that, The image comparison unit determines whether to clean up the feature garbage corresponding to the feature image based on a comparison result of the degree of overlap and a preset degree of overlap, wherein if the degree of overlap is greater than or equal to the preset degree of overlap, it is determined to clean up the feature garbage corresponding to the feature image; the preset degree of overlap is positively correlated with the garbage volume of the kitchen waste. 7.The Internet of Things data processing system based on intelligent garbage classification according to claim 2, characterized in that, The compression processing unit determines whether the smoke in the smoke image is white in response to a determination result that the feature garbage is not cleaned up, wherein if the smoke is white, it is determined to compress the feature garbage after secondary drying; if the smoke is not white, it is determined to compress the feature garbage. 8.The Internet of Things data processing system based on intelligent garbage classification according to claim 2, characterized in that, The data uploading unit cleans up and recycles the feature garbage in response to a determination result that the feature garbage is cleaned up, calculates a volume difference between the garbage volume and the recycling volume, and calculates a weight difference between the garbage weight and the drying weight. 9.The Internet of Things data processing system based on intelligent garbage classification according to claim 8, characterized in that, The data uploading unit compares the feature value with a preset feature value, and determines whether to compress the reorganized image into a plurality of data packets and transmit them to the feature image library according to the comparison result, wherein if the feature value is less than or equal to the preset feature value, it is determined to compress the reorganized image into a plurality of data packets and transmit them to the feature image library; the preset feature value is positively correlated with the garbage volume of the kitchen waste. 10.The Internet of Things data processing system based on intelligent garbage classification according to claim 9, characterized in that, The data uploading unit is configured to compress the reorganized image into a plurality of data packets and transmit them to the feature image library, wherein compressing the image boundary line corresponding to the reorganized image into a first data packet; compressing the feature instance corresponding to the reorganized image into a second data packet; transmitting the reorganized image, the first data packet and the second data packet to the feature image library in sequence.

Images