Intelligent garbage classification transportation control method, system and equipment

Abstract

The invention relates to the technical field of garbage classified transportation, and provides an intelligent garbage classified transportation control method and system, and the method comprises the steps: obtaining a coordinate region semantic tag; based on a YOLOV5 neural network model, carrying out classification identification on the junk image, and obtaining a confidence coefficient vector of a classification result; collecting the weight of garbage in the garbage can at a preset frequency, and constructing weight time sequence data; material information of garbage in the garbage can is obtained, and garbage material characteristics are obtained; obtaining multi-dimensional feature fusion information; and performing block chain chaining on the multi-dimensional feature fusion information to obtain a block chain traceability code. According to the method, the comprehensive index for tracing can be obtained, so that quick retrieval during tracing is facilitated, and the overall tracing capability of garbage classified transportation can be enhanced; the problem that in the prior art, due to the fact that garbage can positioning information, garbage materials and other multi-dimensional information are not fused, the garbage classification transportation tracking and tracing overall capacity needs to be further enhanced is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of garbage classification transportation. Specifically, it relates to an intelligent garbage classification transportation control method, system and device. Background Technique

[0002] In the garbage classification transportation system, data may involve information in different links, such as garbage types, quantities, transportation routes, etc. If the data is tampered with, it may lead to errors in garbage classification and affect the operation of the entire system. At the same time, the garbage classification transportation system may involve multiple links and systems, and it is necessary to ensure the consistency of data between different nodes. These data security issues will directly affect the realization of data transparency and traceability.

[0003] For the "Traceable Garbage Classification Transportation System and Monitoring Method" with the application number CN202410601432.1, by introducing blockchain technology, the data interaction and information recording between modules are made more secure, transparent and reliable, enhancing the traceability effect of the garbage classification transportation system. It also introduces an encrypted hash algorithm to ensure data integrity and authenticity, guaranteeing the security, immutability and transparency of data, thereby improving system efficiency and credibility.

[0004] Although the above-mentioned garbage classification transportation system and monitoring method can record the data in the garbage classification transportation process to achieve the tracking and tracing of garbage classification transportation, it does not integrate multi-dimensional information such as garbage bin positioning information and garbage material. There is still room for further enhancement in its overall ability to track and trace garbage classification transportation. Summary of the Invention

[0005] Based on this, in order to enhance the tracking and tracing ability of garbage classification transportation, the present invention provides an intelligent garbage classification transportation control method, system and device. By integrating multi-dimensional data features including coordinate region semantic tags, confidence vectors, weight time series data, and garbage material characteristics, it can obtain comprehensive indicators for tracking and tracing, facilitating rapid retrieval during tracing, and enhancing the overall traceability ability of garbage classification transportation. The specific technical solutions are as follows:

[0006] An intelligent garbage classification transportation control method includes the following steps:

[0007] Obtain the real-time position coordinate information of the garbage bin, and obtain the coordinate region semantic tag according to the real-time position coordinate information and the preset garbage bin area semantic library;

[0008] Based on the YOLOV5 neural network model, classify and identify the garbage image to obtain the confidence vector of the classification result;

[0009] Collect the garbage weight in the garbage bin at a preset frequency to construct weight time-series data;

[0010] Obtain the material information of the garbage in the garbage bin and obtain the garbage material characteristics;

[0011] Perform feature fusion on the coordinate region semantic label, confidence vector, weight time-series data, and garbage material characteristics to obtain multi-dimensional feature fusion information;

[0012] Upload the multi-dimensional feature fusion information to the blockchain to obtain a blockchain traceability code, and implement traceability tracking of garbage classification transportation based on the blockchain traceability code.

[0013] The intelligent garbage classification transportation control method can obtain comprehensive indicators for traceability for fast retrieval during traceability by obtaining the coordinate region semantic label, confidence vector, weight time-series data, and garbage material characteristics and performing feature fusion on them, which can enhance the overall traceability ability of garbage classification transportation, and solves the problem that the overall traceability ability of garbage classification transportation in the prior art needs to be further enhanced due to the lack of integration of multi-dimensional information such as garbage bin positioning information and garbage materials.

[0014] Preferably, the intelligent garbage classification transportation control method further includes the following steps:

[0015] Obtain the GPS trajectory of the garbage bin and correct the GPS trajectory according to the formula where Δp represents the optimal path correction amount, δ represents the path correction candidate adjustment amount,

[0016] represents the original navigation path data, HMM() represents the path matching algorithm based on the hidden Markov model, R represents the reference map data, and || || map represents the square of the Euclidean distance. 2 represents the square of the Euclidean distance.

[0017] Preferably, the intelligent garbage classification transportation control method further includes the following steps:

[0018] Set m RFID scanning points according to the planned path of the garbage classification transportation;

[0019] Obtain the RFID spatio-temporal verification value according to the actual scanning time, preset scanning time, preset scanning time standard deviation σ, and the number m of RFID scanning points of each RFID scanning point

[0020] Judge whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold;

[0021] Among them, t i ' represents the actual scanning time of the i-th RFID scanning point, and t i ” represents the preset scanning time of the i-th RFID scanning point, and e represents the natural constant.

[0022] Preferably, the intelligent garbage classification transportation control method further includes the following steps:

[0023] Obtain the average transportation speed v of the garbage transport vehicle;

[0024] According to the formula Obtain the preset scanning time of the i-th RFID scanning point;

[0025] Among them, t i-1 ' represents the actual scanning time of the (i - 1)-th RFID scanning point, and D i represents the distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point.

[0026] Preferably, the specific method for judging whether there is an abnormality in garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold includes:

[0027] If the RFID spatio-temporal verification value is less than the preset spatio-temporal verification threshold, it is judged that there is an abnormality in garbage transportation.

[0028] An intelligent garbage classification transportation control system for implementing the intelligent garbage classification transportation control method, which includes:

[0029] A coordinate region semantic label acquisition module, used to acquire the real-time position coordinate information of the trash can, and acquire the coordinate region semantic label according to the real-time position coordinate information and the preset trash can region semantic library;

[0030] A confidence vector acquisition module, used to classify and identify the garbage image based on the YOLOV5 neural network model, and acquire the confidence vector of the classification result;

[0031] A weight time series data construction module, used to collect the garbage weight in the trash can at a preset frequency and construct weight time series data;

[0032] A garbage material feature acquisition module, used to acquire the material information of the garbage in the trash can and acquire the garbage material feature;

[0033] A multi-dimensional feature fusion module, used to perform feature fusion on the coordinate region semantic label, confidence vector, weight time series data, and garbage material feature to obtain multi-dimensional feature fusion information;

[0034] The traceability module is used to upload the multi-dimensional feature fusion information to the blockchain, obtain the blockchain traceability code, and realize the traceability of garbage classification transportation based on the blockchain traceability code.

[0035] Preferably, the intelligent garbage classification transportation control system further includes:

[0036] The GPS trajectory correction module is used to obtain the GPS trajectory of the garbage bin and correct the GPS trajectory according to the formula where Δp represents the optimal path correction amount, δ represents the path correction candidate adjustment amount,

[0037] represents the original navigation path data, HMM() represents the path matching algorithm based on the hidden Markov model, and R represents the reference map data, and || || map represents the square of the Euclidean distance. 2 represents the square of the Euclidean distance.

[0038] Preferably, the intelligent garbage classification transportation control system further includes:

[0039] The scanning point setting module is used to set m RFID scanning points according to the planned path of the garbage classification transportation;

[0040] The spatio-temporal verification value acquisition module is used to obtain the RFID spatio-temporal verification value according to the actual scanning time, preset scanning time, preset scanning time standard deviation σ of each RFID scanning point, and the number m of RFID scanning points

[0041] The abnormality judgment module is used to judge whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold;

[0042] where t i ' represents the actual scanning time of the i-th RFID scanning point, and t i ” represents the preset scanning time of the i-th RFID scanning point, and e represents the natural constant.

[0043] Preferably, the intelligent garbage classification transportation control system further includes:

[0044] The average transportation speed acquisition module is used to obtain the average transportation speed v of the garbage transport vehicle;

[0045] The preset scanning time acquisition module is used to obtain the preset scanning time of the i-th RFID scanning point according to the formula ;

[0046] where t i-1 ' represents the actual scanning time of the (i - 1)-th RFID scanning point, and Di Indicates the travel distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point.

[0047] An intelligent garbage classification transportation control device, comprising:

[0048] A controller;

[0049] A memory storing executable instructions;

[0050] Wherein, the executable instructions can run on the controller and implement the intelligent garbage classification transportation control method. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the drawings are not necessarily drawn to scale, but the emphasis is placed on showing the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0052] Figure 1 Is an overall flow schematic diagram of an intelligent garbage classification transportation control method in an embodiment of the present invention;

[0053] Figure 2 Is a flow schematic diagram of an intelligent garbage classification transportation control method in another embodiment of the present invention Figure 1 ;

[0054] Figure 3 Is a flow schematic diagram of an intelligent garbage classification transportation control method in another embodiment of the present invention Figure 2 ;

[0055] Figure 4 Is an overall structure schematic diagram of an intelligent garbage classification transportation control system in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with its embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0057] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be an intermediate element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intermediate element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are only for the purpose of illustration and do not represent the only implementation.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs. The terms used in the specification of this invention are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0059] In the present invention, the "first" and "second" do not represent specific quantities and orders, but are only used for name distinction.

[0060] For the classified transportation and treatment of garbage, multi-dimensional characteristic information such as the weight change of garbage, material characteristics, classification quality, and corresponding regional information can be used for subsequent traceability analysis. By fusing the multi-dimensional characteristic information in the classified garbage transportation process and uploading it to the blockchain, the limitation of traditional single-dimensional data storage can be broken through, and a three-dimensional and cross-verifiable traceability evidence chain can be formed. In the existing blockchain-based traceability methods for classified garbage transportation, the blockchain uploading operation is often based on single-dimensional characteristic data, without fusing multi-dimensional information such as garbage bin positioning information and garbage materials. There is still room for further improvement in its overall ability for traceability in classified garbage transportation.

[0061] An embodiment of the present invention provides an intelligent classified garbage transportation control method, as Figure 1 shown, including the following steps:

[0062] S1. Obtain the real-time position coordinate information of the garbage bin, and obtain the coordinate area semantic label g according to the real-time position coordinate information and a preset garbage bin area semantic library loc .

[0063] Preferably, the real-time position coordinate information can be obtained through a GPS positioning module installed on the garbage bin or the garbage transport vehicle. The preset garbage bin area semantic library can be understood as a semantic library preset to include the semantic information of garbage bin areas such as kitchens, offices, meeting rooms, shopping malls, and hospitals. Specifically, through the real-time position coordinate information and the preset garbage bin area semantic library, obtain the area type where the garbage bin is located, that is, obtain the coordinate area semantic label.

[0064] By obtaining the coordinate area semantic label, the garbage transportation route can be optimized, and the scheduling strategy of the garbage transport vehicle can be dynamically adjusted.

[0065] S2. Based on the YOLOV5 neural network model, classify and identify the garbage image, and obtain the confidence vector c of the classification result cls . The confidence vector is used to reflect the probability distribution of the garbage category.

[0066] Preferably, a camera installed at the garbage bin's dropping opening captures garbage images, which are then input into a trained YOLOV5 model to output classification results in real-time and quantify the confidence level, thereby obtaining a confidence vector for the classification results. If the confidence value corresponding to the confidence vector is less than the confidence threshold, quality assessment of garbage classification can be achieved by triggering manual review or re-sorting actions, improving the quality of garbage classification.

[0067] S3. Collect the garbage weight in the garbage bin at a preset frequency to construct weight time-series data w t . Here, the garbage weight information can be collected by a high-precision pressure sensor installed in the garbage bin, and the garbage weight information is processed in combination with a filtering algorithm to eliminate environmental noise interference and improve the effectiveness of the weight time-series data. t in w t can be understood as the data collection time.

[0068] Preferably, for a certain garbage bin, obtain the garbage weight information within a preset past time period, and calculate the average change information of the garbage weight within a preset time width. For example, collect the garbage weight information within the past month or quarter at a preset frequency, and calculate the garbage weight change information of the garbage bin within a preset time width such as ten minutes or half an hour. If the garbage weight mutates within a certain set time width, that is, the garbage weight change value within a certain set time width exceeds the preset weight threshold (for example, the garbage weight in the garbage bin increases by 50 kg within ten minutes, which is greater than the preset weight threshold of 20 kg), it is determined that there is an abnormal mixed loading in the garbage bin to identify potential mixed loading behaviors.

[0069] At the same time, by obtaining the garbage weight information within a preset past time period and calculating the average change information of the garbage weight within a preset time width, it is also possible to construct a garbage weight change function curve based on the garbage weight information, predict the garbage bin capacity saturation time, and optimize the collection and transportation frequency.

[0070] S4. Obtain the material information of the garbage in the garbage bin to obtain the garbage material feature m mat .

[0071] Preferably, the garbage material feature (such as plastic type, paper fiber, etc.) can be obtained through the reflected spectrum data collected by a near-infrared sensor. For example, the garbage material feature is distinguished through spectral analysis technology to achieve refined classification of garbage of different materials and control the recycling quality of garbage. When detecting impurity materials (such as metal fragments mixed in food waste), targeted treatment can be carried out on the impurity materials to improve the resource utilization efficiency.

[0072] S5. Perform feature fusion on the coordinate region semantic label, confidence vector, weight time-series data, and garbage material feature to obtain multi-dimensional feature fusion information.

[0073] Specifically, the multi-dimensional feature fusion information F = Relu(W g ·GeoHash(g loc ) + W c ·c cls + W w ·GRU(w t ) + W m ·m mat ).

[0074] Among them, Geohash represents a spatial indexing algorithm that encodes two-dimensional longitude and latitude data into a one-dimensional string, which is used to convert GPS coordinates into 32-bit geographical hash values to achieve spatial discretized expression; GRU (Gated Recurrent Unit) represents a variant of the Recurrent Neural Network (RNN), aiming to process sequence data and capture significant temporal differences; W g , W c , W w , W m respectively represent the weight matrices of the coordinate region semantic label, the confidence vector, the significant temporal data, and the waste material characteristics.

[0075] By performing linear transformation and addition on each feature through the weight matrices of the coordinate region semantic label, the confidence vector, the significant temporal data, and the waste material characteristics, the independent contributions of each modality can be retained. Additionally, through the non-linear activation function ReLU, the non-linear expression ability of the system can be enhanced, and at the same time, negative values can be filtered out.

[0076] Preferably, for the coordinate region semantic label, the confidence vector, the significant temporal data, and the waste material characteristics, corresponding normalization processing can be performed to avoid the inconvenience of feature fusion caused by numerical differences.

[0077] Based on the formula F = Relu(W g ·GeoHash(g loc ) + W c ·c cls + W w ·GRU(w t ) + W m ·m mat ), it can perform weighted fusion of multi-modal features, taking into account the spatial, temporal, and physical attributes involved in the garbage classification process. The final comprehensive index obtained for traceability is conducive to rapid retrieval during traceability. It adopts a unified feature fusion algorithm, enabling the data in production, transportation, processing, etc. to be quickly compared through hash values, solving the problem of incompatible data formats in traditional traceability systems.

[0078] Compared with the prior art which performs blockchain uploading operations and traceability tracking for single-dimensional features, the present invention enhances the overall traceability ability of waste classification transportation by obtaining coordinate region semantic tags, confidence vectors, weight time-series data, and waste material characteristics, and fusing these features to obtain multi-dimensional feature fusion information, breaking through the limitations of traditional single-dimensional data storage and forming a three-dimensional and cross-verifiable traceability evidence chain.

[0079] S6. Upload the multi-dimensional feature fusion information to the blockchain to obtain a blockchain traceability code, and implement traceability tracking of waste classification transportation based on the blockchain traceability code.

[0080] Preferably, for the blockchain uploading structure, it can be expressed as Block j = SHA256(F j || MerkleRoot(T j-1 )) || nonce).

[0081] Among them, F j represents the current block feature vector, that is, the multi-dimensional feature fusion information of the trash can in the j-th block; MerkleRoot(T j-1 ) represents the Merkle root hash of the previous block, which can be understood as the root node hash value of the Merkle tree constructed from all transaction data in the (j - 1)-th block; nonce represents a random number; SHA256() represents a hash function.

[0082] In summary, the intelligent waste classification transportation control method can obtain comprehensive indicators for traceability tracking by obtaining coordinate region semantic tags, confidence vectors, weight time-series data, and waste material characteristics, and fusing these features to obtain multi-dimensional feature fusion information, so as to achieve fast retrieval during traceability, enhance the overall traceability ability of waste classification transportation, and solve the problem that the overall traceability ability of waste classification transportation in the prior art needs to be further enhanced due to the lack of fusion of multi-dimensional information such as trash can positioning information and waste material.

[0083] As a preferred technical solution, as Figure 2 shown, the intelligent waste classification transportation control method further includes the following steps:

[0084] S7. Obtain the GPS trajectory of the trash can and correct the GPS trajectory according to the formula where Δp represents the optimal path correction amount, δ represents the path correction candidate adjustment amount,

[0085] represents the original navigation path data, HMM() represents a path matching algorithm based on the hidden Markov model, and R map ​Represents reference map data, preferably high-precision map, || || 2 Represents the square of the Euclidean distance.

[0086] Specifically, the formula aims to correct the original GPS trajectory points with noise into the path Δp that best matches the high-precision map R map by using the state transition probability and observation probability of the Hidden Markov Model (HMM), combined with the map topology, to eliminate the GPS signal drift error.

[0087] In step S7, according to the formula the specific method for correcting the GPS trajectory includes the following steps:

[0088] S71, State space definition.

[0089] Discretize the road network of the high-precision map R map into a candidate state set {r i}, which is a section or intersection node in each road network. For example, the urban arterial road can be divided into discrete sections at 100-meter intervals as the potential states of the Hidden Markov Model (HMM).

[0090] S72, Observation probability modeling, constructing a GPS coordinate error model.

[0091] Assume that the GPS observations follow a Gaussian distribution, and the emission probability is calculated as

[0092] where μr i is the geometric center coordinate of section r i , which can be understood as the mean of the true position r i , that is, the expected GPS observation value under error-free conditions; for example, if the precise longitude and latitude of a point are known as μr i , then the observation value z t should be distributed around this mean; σ z is the GPS measurement standard deviation (usually taken as 5 - 15 meters), which can be understood as the standard deviation of the Gaussian distribution and reflects the dispersion degree of the observation error; z t represents the actually observed GPS coordinate value; r i represents the coordinate of the true position (or reference position), which, in the ideal case, corresponds to the theoretical value of the error-free GPS coordinate and is usually determined by high-precision measurement or known reference points;

[0093] S73, Transition probability optimization. Based on the road network topology and vehicle kinematic constraints, define the state transition probability at adjacent times.

[0094] S74, Path Inference and Correction.

[0095] Use the Viterbi algorithm to solve the maximum a posteriori probability path and output the corrected trajectory Δp.

[0096] Preferably, to prevent the probability value from underflowing, take the logarithm of the probability and convert it into an addition operation to optimize the data.

[0097] Based on the formula Correct the GPS trajectory. When tested with GPS data having a sampling interval of 30 seconds, the matching accuracy can reach over 85%; when the sampling time is increased to 5 seconds, the accuracy exceeds 95%. That is to say, by matching the high-precision map R through the hidden Markov model map , the GPS drift error can be eliminated.

[0098] As an optimized technical solution, as Figure 3 shown, the intelligent garbage classification transportation control method further includes the following steps:

[0099] S8, Set m RFID scanning points according to the planned path of the garbage classification transportation.

[0100] S9, Obtain the RFID spatio-temporal verification value according to the actual scanning time, preset scanning time, preset scanning time standard deviation σ, and the number m of RFID scanning points for each RFID scanning point

[0101] S10, Judge whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold.

[0102] Among them, t i ' represents the actual scanning time of the i-th RFID scanning point, t i ” represents the preset scanning time of the i-th RFID scanning point, and e represents the natural constant. For the preset scanning time standard deviation σ, it can be understood as reflecting the tolerance of the time sequence deviation caused by comprehensive factors such as road conditions fluctuations and equipment response delays during the transportation process, and it can be calculated by obtaining the statistical value of the time error variance of historical data.

[0103] Preferably, the intelligent garbage classification transportation control method further includes the following steps:

[0104] S11, Obtain the average transportation speed v of the garbage transport vehicle;

[0105] S12, Obtain the preset scanning time of the i-th RFID scanning point according to the formula ;

[0106] Among them, t i-1' represents the actual scanning time of the (i - 1)-th RFID scanning point, D i represents the distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point.

[0107] Preferably, the specific method for judging whether there is an abnormality in garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold includes: if the RFID spatio-temporal verification value is less than the preset spatio-temporal verification threshold, it is judged that there is an abnormality in garbage transportation.

[0108] When there is an abnormality in garbage transportation, if: single node |t i '-t i ”|> 3σ, it can be judged that there is an early or lag abnormality. If the consistency of the error directions of consecutive nodes exceeds 80%, it can be judged that there is a path deviation abnormality. For the consistency of the error directions of consecutive nodes, it can be obtained by calculating the positive and negative distribution of the t i '-t i ” values of multiple nodes. For example, if the number of consecutive nodes is 100, and if there are 85 consecutive nodes with the t i '-t i ” values being positive or negative at the same time, then the value of the consistency of the error directions of consecutive nodes is 85%.

[0109] An embodiment of the present invention further provides an intelligent garbage classification transportation control system for implementing the intelligent garbage classification transportation control method, as Figure 4 shown, which includes a coordinate area semantic label acquisition module, a confidence vector acquisition module, a weight time series data construction module, a garbage material feature acquisition module, a multi-dimensional feature fusion module, and a tracking and tracing module.

[0110] The coordinate area semantic label acquisition module is used to acquire the real-time position coordinate information of the garbage bin, and acquire the coordinate area semantic label according to the real-time position coordinate information and the preset garbage bin area semantic library; the confidence vector acquisition module is used to classify and identify the garbage image based on the YOLOV5 neural network model, and acquire the confidence vector of the classification result.

[0111] The weight time series data construction module is used to collect the weight of the garbage in the garbage bin at a preset frequency and construct the weight time series data; the garbage material feature acquisition module is used to acquire the material information of the garbage in the garbage bin and acquire the garbage material feature; the multi-dimensional feature fusion module is used to perform feature fusion on the coordinate area semantic label, the confidence vector, the weight time series data, and the garbage material feature to acquire multi-dimensional feature fusion information; the tracking and tracing module is used to upload the multi-dimensional feature fusion information to the blockchain to acquire the blockchain traceability code, and implement the tracking and tracing of garbage classification transportation based on the blockchain traceability code.

[0112] Preferably, the intelligent garbage classification transportation control system further includes a GPS trajectory correction module.

[0113] The GPS trajectory correction module is used to obtain the GPS trajectory of the trash can and correct the GPS trajectory according to the formula the GPS trajectory;

[0114] where, Δp represents the optimal path correction amount, which is the best adjustment value obtained by minimizing the error function. Its physical meaning is the path deviation correction amount of the inertial navigation system (or other positioning systems) and is used to align the original path with the map reference; δ represents the candidate adjustment amount for path correction, and possible correction values are traversed during the optimization process. Finally, the δ that minimizes the objective function is selected as Δp; represents the original navigation path data, which can be understood as the original path data output by the inertial navigation system (such as the INS / GPS integrated navigation system) and may contain positioning errors (such as drift, cumulative error, etc.); HMM() represents the path matching algorithm based on the hidden Markov model, which is used to perform probability modeling on the original path and generate a candidate path sequence that matches the map road network structure; R map represents the reference map data, which includes the geometric shape, topological relationship, and attribute information of the road and serves as the benchmark for path matching; || || 2 represents the square of the Euclidean distance and is used to measure the difference between the corrected path and the map reference. Minimizing this difference is the goal of the optimization.

[0115] The intelligent garbage classification transportation control system further includes a scan point setting module, a spatio-temporal verification value acquisition module, and an anomaly judgment module.

[0116] The scan point setting module is used to set m RFID scan points according to the planned path of the garbage classification transportation; the spatio-temporal verification value acquisition module is used to obtain the RFID spatio-temporal verification value according to the actual scan time, preset scan time, preset scan time standard deviation σ, and the number m of RFID scan points of each RFID scan point The anomaly judgment module is used to judge whether there is an anomaly in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold.

[0117] where, t i ' represents the actual scan time of the i-th RFID scan point, t i ” represents the preset scan time of the i-th RFID scan point, and e represents the natural constant.

[0118] The intelligent garbage classification transportation control system further includes an average transportation speed acquisition module and a preset scan time acquisition module.

[0119] The average transportation speed acquisition module is used to acquire the average transportation speed v of the garbage transport vehicle; the preset scanning time acquisition module is used to obtain the preset scanning time of the i-th RFID scanning point according to the formula where t

[0120] ' represents the actual scanning time of the (i - 1)-th RFID scanning point, and D i-1 represents the distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point. i Specifically, the multi-dimensional feature fusion information F = Relu(W

[0121] · GeoHash(g g ) + W loc · c c + W cls · GRU(w w ) + W t · m m ) mat .

[0122] Geohash represents a spatial indexing algorithm that encodes two-dimensional latitude and longitude data into a one-dimensional string, which is used to convert GPS coordinates into 32-bit geographical hash values to achieve spatial discretized expression; GRU (Gated Recurrent Unit) represents a variant of the Recurrent Neural Network (RNN), aiming to process sequence data and capture weight time series differences; W g , W c , W w , W m represent the weight matrices of the coordinate region semantic label, confidence vector, weight time series data, and garbage material characteristics respectively.

[0123] By performing linear transformation and addition on each feature through the weight matrices of the coordinate region semantic label, confidence vector, weight time series data, and garbage material characteristics, the independent contributions of each modality can be retained. In addition, through the non-linear activation function ReLU, the non-linear expression ability of the system can be enhanced, and at the same time, negative values can be filtered out.

[0124] For the coordinate region semantic label, confidence vector, weight time series data, and garbage material characteristics, corresponding normalization processing can be performed to avoid the inconvenience of feature fusion caused by numerical differences.

[0125] Based on the formula F = Relu(W g · GeoHash(g loc ) + W c · c cls + W w · GRU(w t) + W m ·m mat ) which can perform weighted fusion on multi-modal features, taking into account the spatial, temporal, and physical attributes involved in the garbage classification process. The final comprehensive index obtained for traceability is conducive to rapid retrieval during traceability.

[0126] Compared with the prior art that performs blockchain uploading operations and traceability for single-dimensional features, the present invention enhances the overall traceability ability of garbage classification transportation by obtaining coordinate region semantic tags, confidence vectors, weight time-series data, and garbage material features and performing feature fusion on them to obtain multi-dimensional feature fusion information.

[0127] In summary, the intelligent garbage classification transportation control system can obtain multi-dimensional feature fusion information by obtaining coordinate region semantic tags, confidence vectors, weight time-series data, and garbage material features and performing feature fusion on them, and can obtain a comprehensive index for traceability for rapid retrieval during traceability, which can enhance the overall traceability ability of garbage classification transportation and solve the problem that the overall traceability ability of garbage classification transportation in the prior art needs to be further enhanced due to the lack of fusion of multi-dimensional information such as garbage bin positioning information and garbage materials.

[0128] An embodiment of the present invention further provides an intelligent garbage classification transportation control device, which includes: a controller; a memory storing executable instructions; wherein, the executable instructions can run on the controller and implement the intelligent garbage classification transportation control method described above.

[0129] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0130] The above embodiments only represent several implementation manners of the present invention, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the invention patent should be subject to the appended claims.

Claims

1. An intelligent garbage classification transportation control method, characterized in that, The intelligent garbage classification transportation control method includes the following steps: Obtain the real-time position coordinate information of the garbage bin, and according to the real-time position coordinate information and the preset garbage bin area semantic library, obtain the coordinate area semantic label; Based on the YOLOV5 neural network model, classify and identify the garbage image to obtain the confidence vector of the classification result; Collect the garbage weight in the garbage bin at a preset frequency to construct weight time series data; Obtain the material information of the garbage in the garbage bin to obtain the garbage material characteristics; Perform feature fusion on the coordinate area semantic label, confidence vector, weight time series data, and garbage material characteristics to obtain multi-dimensional feature fusion information; Upload the multi-dimensional feature fusion information to the blockchain to obtain the blockchain traceability code, and realize the traceability of garbage classification transportation based on the blockchain traceability code.

2. The intelligent garbage classification transportation control method according to claim 1, characterized in that, The intelligent garbage classification transportation control method further includes the following steps: Obtain the GPS trajectory of the trash bin and correct it according to the formula correct the GPS trajectory; where Δp represents the optimal path correction amount, and δ represents the candidate adjustment amount of path correction. represents the original navigation path data, HMM() represents the path matching algorithm based on the hidden Markov model, and R map represents the reference map data, || || 2 represents the square of the Euclidean distance.

3. The intelligent garbage classification transportation control method according to claim 2, characterized in that, The intelligent garbage classification transportation control method further includes the following steps: According to the planned path of the garbage classification transportation, set m RFID scanning points; Obtain the RFID spatio-temporal verification value based on the actual scanning time, preset scanning time, preset standard deviation of the scanning time σ, and the number m of RFID scanning points for each of the said RFID scanning points Judge whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold; where t i ' represents the actual scanning time of the i-th RFID scanning point, and t i ” represents the preset scanning time of the i-th RFID scanning point, and e represents the natural constant.

4. The intelligent garbage classification transportation control method according to claim 3, characterized in that, The intelligent garbage classification transportation control method further includes the following steps: Obtain the average transportation speed v of the garbage truck; According to the formula Obtain the preset scanning time of the i-th RFID scanning point; where t i-1 ' represents the actual scanning time of the (i - 1)-th RFID scanning point, and D i represents the travel distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point.

5. The intelligent garbage classification transportation control method according to claim 4, wherein, The specific method for judging whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold includes: If the RFID spatio-temporal verification value is less than the preset spatio-temporal verification threshold, it is judged that there is an abnormality in the garbage transportation.

6. An intelligent garbage classification transportation control system for implementing the intelligent garbage classification transportation control method as described in any one of claims 1-5, characterized in that, The intelligent garbage classification transportation control system includes: A coordinate area semantic label acquisition module, which is used to obtain the real-time position coordinate information of the garbage bin, and according to the real-time position coordinate information and the preset garbage bin area semantic library, obtain the coordinate area semantic label; A confidence vector acquisition module, which is used to classify and identify the garbage image based on the YOLOV5 neural network model to obtain the confidence vector of the classification result; A weight time series data construction module, which is used to collect the garbage weight in the garbage bin at a preset frequency to construct weight time series data; A garbage material feature acquisition module, which is used to obtain the material information of the garbage in the garbage bin to obtain the garbage material characteristics; A multi-dimensional feature fusion module, which is used to perform feature fusion on the coordinate area semantic label, confidence vector, weight time series data, and garbage material characteristics to obtain multi-dimensional feature fusion information; A traceability module, which is used to upload the multi-dimensional feature fusion information to the blockchain to obtain the blockchain traceability code, and realize the traceability of garbage classification transportation based on the blockchain traceability code.

7. An intelligent garbage classification transportation control system according to claim 6, characterized in that, The intelligent garbage classification transportation control system further includes: The GPS trajectory correction module is used to obtain the GPS trajectory of the dustbin and correct it according to the formula correct the said GPS trajectory; Among them, Δp represents the optimal path correction amount, and δ represents the candidate adjustment amount of path correction. represents the original navigation path data, HMM() represents the path matching algorithm based on the hidden Markov model, and R map represents the reference map data, || || 2 represents the square of the Euclidean distance.

8. The intelligent garbage classification transportation control system according to claim 7, wherein The intelligent garbage classification transportation control system further includes: A scanning point setting module, which is used to set m RFID scanning points according to the planned path of the garbage classification transportation; Space-time verification value acquisition module, which is used to obtain the RFID space-time verification value according to the actual scanning time, preset scanning time, preset standard deviation of scanning time σ of each RFID scanning point, and the number m of RFID scanning points An abnormality judgment module, which is used to judge whether there is an abnormality in the garbage transportation according to the RFID spatio-temporal verification value and the preset spatio-temporal verification threshold; where t i ' represents the actual scanning time of the i-th RFID scanning point, and t i ” represents the preset scanning time of the i-th RFID scanning point, and e represents the natural constant.

9. The intelligent garbage classification transportation control system according to claim 8, wherein The intelligent garbage classification transportation control system further includes: An average transportation speed acquisition module, which is used to obtain the average transportation speed v of the garbage truck; A preset scanning time acquisition module, which is used to obtain the preset scanning time of the i-th RFID scanning point according to the formula ; where t i-1 ' represents the actual scanning time of the (i - 1)-th RFID scanning point, and D i represents the travel distance between the i-th RFID scanning point and the (i - 1)-th RFID scanning point.

10. An intelligent garbage classification transportation control device, characterized in that, The intelligent garbage classification transportation control device includes: A controller; A memory storing executable instructions; Among them, the executable instructions can run on the controller and implement the intelligent garbage classification transportation control method according to any one of claims 1 to 5.

Images