Sanitation operation supervision system

Abstract

The invention is suitable for the technical field of environmental protection and cleaning, and provides an environmental sanitation operation supervision system, which comprises a data acquisition module, a feature extraction module, a model training module, a monitoring implementation module, a public supervision module, a wireless communication module, a memory, an alarm, a processing center and a smart phone. The data acquisition module, the feature extraction module, the model training module, the monitoring implementation module, the public supervision module and the wireless communication module are respectively connected with the processing center; the smart phone is in wireless network connection with the wireless communication module in the range of a wireless network or the Internet. According to the invention, environmental sanitation vehicles and environmental sanitation workers can be automatically equipped, automatic environmental sanitation cleaning operation can be realized, the road surface environmental sanitation condition can be automatically monitored, automatic environmental sanitation operation and all-weather monitoring of the road surface environmental sanitation condition are realized, the environmental sanitation operation efficiency is improved, and the environmental sanitation operation cost is reduced.

Description

Technical Field

[0001] The present invention relates to the field of environmental protection and cleaning technology, and in particular to a sanitation operation supervision system in the field of artificial intelligence. Background Art

[0002] With the development of technologies such as the Internet of Things, big data, and artificial intelligence, the sanitation industry is gradually becoming intelligent. The application of advanced technologies such as smart sanitation systems, smart sweepers, and garbage sorting equipment has significantly improved operating efficiency and service quality. In the future, the sanitation industry will continue to develop in the direction of intelligence and greening. The application of smart sanitation systems will be more extensive, and resource allocation will be optimized through big data analysis to improve operating efficiency. Most of the existing sanitation operations still use manual dispatching, manual monitoring of operating results, and manual scoring of sanitation performance. Not only is the sanitation operation efficiency low, but the labor cost is also high. Summary of the invention

[0003] In order to address the deficiencies in the prior art, the purpose of the present invention is to provide a sanitation operation supervision system, which can automatically equip sanitation vehicles and sanitation personnel, automatically perform sanitation cleaning operations, and automatically monitor road sanitation conditions by setting up a data acquisition module, a feature extraction module, a model training module, a monitoring implementation module, and a public supervision module, thereby realizing automated sanitation operations, monitoring road sanitation conditions around the clock, improving sanitation operation efficiency, and reducing sanitation operation costs.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is:

[0005] A sanitation operation supervision system, comprising a data acquisition module, a feature extraction module, a model training module, a monitoring implementation module, a public supervision module, a wireless communication module, a memory, an alarm, a processing center, and a smart phone; the data acquisition module, the feature extraction module, the model training module, the monitoring implementation module, the public supervision module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center;

[0006] The smart phone is connected to the wireless communication module wireless network within the wireless network or the Internet respectively;

[0007] The wireless communication module is provided with a wireless network unit, which is responsible for sending and receiving wireless signals, and automatically connects to the intelligent control terminal within the effective network range, and performs effective wireless network connection through the Internet of Things system;

[0008] The alarm device compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to adjust the parameters to continue training; and the actual sanitation quality information of the road is compared with the sanitation quality standard of the road stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to rework and re-sweeping; and the actual sanitation score of the area is compared with the sanitation evaluation standard score of the area stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to continue to improve;

[0009] The memory is responsible for storing information of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, and alarm, as well as the model comprehensive evaluation standard score, road sanitation quality standard, and regional sanitation evaluation standard score;

[0010] The processing center is responsible for the information transmission of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, alarm, and memory. It is the hub center of the system and compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it will notify that it can be monitored; if it does not meet the standard, it will be passed to the alarm and notify to adjust the parameters to continue training; and compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if it meets the standard, the sanitation operation is completed; if it does not meet the standard, it will be passed to the alarm and notify to rework and re-sweep; and compares the actual sanitation score of the area with the sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will continue to maintain; if it does not meet the standard, it will be passed to the alarm and notify continuous improvement;

[0011] The data acquisition module includes an information acquisition unit, an image grayscale unit, a geometric change unit, a filtering and denoising unit, and a normalization processing unit. The image of the running track and working status of the sanitation vehicle is obtained through the sanitation monitoring equipment and converted into recognizable sanitation operation information, and then transmitted to the feature extraction module;

[0012] The feature extraction module includes a data input unit, a convolution input unit, a pooling conversion unit, a fully connected layer unit, an output conversion unit, and a feature extraction unit, which is responsible for extracting the color, texture, and shape features of the road according to the preprocessed data for subsequent model training and passing them to the model training module;

[0013] The model training module includes a training simulation unit, a model evaluation unit, an abnormality identification unit, a cost budget unit, and a parameter determination unit, which is responsible for using a convolutional neural network to perform model training based on road cleaning characteristics to learn the features that distinguish normal and abnormal images and continuously adjust model parameters to improve model accuracy, and pass them to the monitoring implementation module;

[0014] The monitoring implementation module includes a cleaning operation unit, an information collection unit, an emergency dispatch unit, and a training and education unit. Through the video monitoring system, the sanitation operations in the area are monitored in a timely and effective manner according to the trained model, and training is conducted to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations, and the information is passed to the public supervision module;

[0015] The public supervision module includes a garbage classification unit, a public participation unit, and an information aggregation unit. By establishing a sanitation complaint service center so that the public can report sanitation problems and summarize and improve them, a good atmosphere of "everyone participates and everyone supervises" is formed and the information is transmitted to the processing center.

[0016] The present invention provides a sanitation operation supervision system, which is implemented by a sanitation operation supervision method, and the method comprises the following steps:

[0017] S10, before the operation, the data acquisition module obtains the running track and working status image of the sanitation vehicle through the sanitation monitoring equipment and converts it into recognizable sanitation operation information, and passes it to the feature extraction module;

[0018] S20, the feature extraction module extracts the features of the color, texture and shape of the road according to the preprocessed data for subsequent model training, and transmits them to the model training module;

[0019] S30, the model training module uses a convolutional neural network to perform model training according to the road cleaning characteristics to learn the features of distinguishing normal and abnormal images and continuously adjusts the model parameters to improve the model accuracy, and transmits the model to the monitoring implementation module;

[0020] S40, the monitoring implementation module uses the video monitoring system to monitor the sanitation operations in the area in a timely and effective manner according to the trained model and provides training to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations, and transmits it to the public supervision module;

[0021] S50, the public supervision module establishes a sanitation complaint service center so that the public can report sanitation problems and summarize improvements, forming a good atmosphere of "everyone participates and everyone supervises", and then transmits them to the processing center;

[0022] S60. The processing center compares the actual environmental sanitation score of the area with the environmental sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will be maintained; if it does not meet the standard, it will be transmitted to the alarm and notified to continue to improve until it meets the standard.

[0023] Further, the step S10 comprises the following steps:

[0024] S11, the information collection unit installs BDS positioning and high-definition cameras on sanitation vehicles to obtain image data of the cleanliness of each road in real time, providing a basis for subsequent model training and passing it to the image grayscale unit;

[0025] S12, the image grayscale unit obtains the grayscaled road cleaning image according to the image grayscale calculation formula "f(i, j) = 0.30R(i, j) + 0.59G(i, j) + 0.11B(i, j), f(i, j) is the grayscale image, R(i, j), G(i, j), B(i, j) are the original images of three colors respectively", and transmits it to the image translation unit;

[0026] S13, the geometric change unit corrects the systematic error of image acquisition and the random error of instrument position by translation, transposition, mirroring, rotation and scaling to improve the generalization ability of the model, and transmits it to the filtering and denoising unit;

[0027] S14, the filtering and denoising unit obtains the denoised road cleaning image according to the two-dimensional median filtering calculation formula "g(x, y) = med{f(xk, yl), (k, l∈W)}, f(x, y) is the original image, g(x, y) is the filtered image, K is the template window size, and W is the neighborhood window" to suppress noise and improve the clarity of the image, and transmits it to the normalization processing unit;

[0028] S15. The normalization processing unit obtains the normalized road cleaning image data according to the image normalization conversion calculation formula "Xn = (X0-Xmin) / Xmax-Xmin), X2 is the normalized image matrix, X0 is the original image matrix, Xmax and Xmin are the maximum and minimum values ​​in the original image respectively" to facilitate subsequent feature extraction.

[0029] Further, the step S20 comprises the following steps:

[0030] S21, the data input unit calculates the formula "P" according to the filling size H =[(Ho-1)*S H +K H -H i ] / 2, P W =[(Wo-1)*S W +K W -Wi] / 2,P H , P W are the padding sizes in height and width respectively, Hi and Wi are the height and width of the input feature map respectively, and K H , K W are the height and width of the convolution kernel, S H , S W Where Ho and Wo are the values ​​of the step length in the height and width directions, respectively. Ho and Wo are the height and width of the output feature map, respectively. The padding data is obtained and input, and passed to the convolution input unit.

[0031] S22, the convolution input unit obtains the convolution input data according to the convolution input calculation formula “z(t)=∫x(m)y(tm)dm, z(t) is the output function, x(t) is the input function, y(t) is the convolution kernel function, and dm is the integral differential element of the variable m”, and passes it to the pooling conversion unit;

[0032] S23, the pooling conversion unit calculates the maximum pooling formula "Z(i, j) = max(X[i*P S (i+1)*P S ,j*P S (j+1)*PS]), Z(i, j) is a pixel value in the output feature map after pooling, i and j are the positions in the output feature map respectively, max is to find the maximum value of all pixel values ​​in the input window, X is the input feature map, PS is the window size of the pooling operation. The maximum pixel value after the maximum pooling enters the fully connected layer to retain important feature information and pass it to the fully connected layer unit;

[0033] S24, the fully connected layer unit obtains the connection layer output data according to the fully connected layer calculation formula "y = f(∑(wn*xn) + b), y is the output result, f is the activation function, wn is the weight of the nth input feature, xn is the nth input feature, n is the dimension of the input feature, and b is the bias" and enters the output layer for subsequent output calculation and is passed to the output conversion unit;

[0034] S25, the output conversion unit obtains the output value size after convolution according to the output layer conversion calculation formula "N = (P-F + 2C) / S + 1, N is the output size after convolution, P is the input size before convolution, F is the convolution kernel size, C is the number of layers of 0 added around the image, and S is the step size" for subsequent model analysis and training, and passes it to the feature extraction unit;

[0035] S26, the feature extraction unit extracts key features of road pollution, road conditions, environmental sanitation, road signs, and environmental changes of road cleanliness according to the processed image data, so as to facilitate subsequent modeling training and recognition.

[0036] Further, the step S30 comprises the following steps:

[0037] S31, the training simulation unit trains the machine learning model using the extracted feature information, identifies whether the road monitoring data is abnormal and its abnormal category, and transmits it to the model evaluation unit;

[0038] S32. The model evaluation unit obtains the comprehensive evaluation score after model training according to the model comprehensive evaluation score calculation formula "F1 = 2*(Pr*Re) / (Pr+Re, F1 is the model comprehensive evaluation score, Pr is the precision, and Re is the recall rate" to ensure the accuracy of model verification and transmit it to the processing center;

[0039] S33, the processing center compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it notifies that monitoring can be carried out; if it does not meet the standard, it is transmitted to the alarm and notifies to adjust the parameters to continue training;

[0040] S34, the abnormality recognition unit inputs the trained image into the trained convolutional neural network model to match the road cleaning monitoring image obtained in real time and confirm its abnormal category, and transmits it to the cost budget unit;

[0041] S35. The cost budget unit obtains the daily cleaning cost of the sanitation vehicle according to the daily cleaning cost calculation formula of "Jg = Lj*Cj + Lj*Xc + Lj / Dz*Dy*Gr, Jg is the daily cleaning cost of the sanitation vehicle (yuan / day), Lj is the daily cleaning mileage of the sanitation vehicle (km / day), Cj is the cleaning material fee of the sanitation vehicle per kilometer (yuan / km), Xc is the cleaning repair fee of the sanitation vehicle per kilometer (yuan / km), Dz is the daily cleaning mileage of the sanitation vehicle (km / day), Dy is the number of sanitation vehicle operators (persons), and Gr is the daily salary of the sanitation vehicle personnel (yuan / person / day)" for subsequent resource allocation, and passes it to the parameter determination unit;

[0042] S36. The parameter determination unit obtains the resource configuration of each road corresponding to the model and the automatic operation parameters of various sanitation monitoring equipment according to the road cleaning abnormality category, so as to automatically match and apply the subsequent automatic monitoring of the sanitation monitoring equipment.

[0043] Further, the step S40 comprises the following steps:

[0044] S41, the cleaning operation unit performs automatic sanitation cleaning operations according to the trained model and the target cleaning section, matching the type and number of sanitation vehicles, the number of sanitation workers and the cleaning process, and transmits the information to the information collection unit;

[0045] S42, the information collection unit obtains images of roads in the area through the set high-definition camera and converts them into environmental sanitation quality information of the road surface to ensure that the environmental sanitation quality of the road meets the standards and transmits them to the processing center;

[0046] S43, the processing center compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if the standard is met, the sanitation work is completed; if the standard is not met, the alarm is transmitted to the alarm and the work is notified to be reworked and re-sweeped;

[0047] S44, the emergency dispatch unit sends an alarm based on the abnormal feedback and uses the trained model to promptly assign the corresponding sanitation vehicle and equip sanitation workers to take special disposal measures to ensure that the difficult areas are cleaned up and passed on to the training and education unit;

[0048] S45. The training and education unit improves the professional quality and skills of sanitation workers through online training and distance education according to the requirements of the model, so that sanitation workers can save themselves and each other and work safely in emergency situations.

[0049] Further, the step S50 comprises the following steps:

[0050] S51, the garbage classification unit uses the trained model to classify and supervise garbage according to environmental protection laws and regulations, and automatically generates inspection, assessment and law enforcement work orders, and passes them to the public participation unit;

[0051] S52. The public participation unit is connected to the sanitation complaint service center by setting up mini programs, apps, and hotlines to facilitate the public to report sanitation problems so that subsequent sanitation operations can be continuously improved and passed on to the information aggregation unit;

[0052] S53. The information aggregation unit compares the sanitation maintenance information of each road obtained by the high-definition camera with the complaint information to obtain the severity of the complaint result, and summarizes it into a sanitation score for each road as a basis for subsequent sanitation improvement.

[0053] A sanitation operation supervision system provided by the present invention also includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when the functional modules execute the computer program, the steps of the sanitation operation supervision method described in any one of the above are implemented; the computer-readable storage medium stores a computer program, and when the computer program is executed by the functional modules, the steps of the sanitation operation supervision method described in any one of the above are implemented.

[0054] The present invention provides a sanitation operation supervision system, which also includes a sanitation operation supervision control device, which is implemented by the sanitation operation supervision method described above.

[0055] The beneficial effects of the present invention compared with the prior art are as follows:

[0056] By setting up data acquisition module, feature extraction module, model training module, monitoring implementation module and public supervision module, sanitation vehicles and their sanitation personnel can be automatically equipped, sanitation cleaning operations can be automatically performed, and road sanitation conditions can be automatically monitored, thus realizing automated sanitation operations, monitoring road sanitation conditions around the clock, improving sanitation operation efficiency and reducing sanitation operation costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or exemplary technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0058] Figure 1 It is a schematic diagram of the system module of the present invention;

[0059] Figure 2 It is a schematic diagram of a data acquisition module of the present invention;

[0060] Figure 3 It is a schematic diagram of a feature extraction module of the present invention;

[0061] Figure 4 This is a schematic diagram of a model training module of the present invention;

[0062] Figure 5 It is a schematic diagram of the monitoring implementation module of the present invention;

[0063] Figure 6 It is a schematic diagram of the public supervision module of the present invention;

[0064] Figure 7 It is a schematic diagram of the method flow control program of the present invention;

[0065] Figure 8 This is a schematic diagram of step S10 in the method flow of the present invention;

[0066] Fig. 9 This is a schematic diagram of step S20 in the method flow of the present invention;

[0067] Fig.10 This is a schematic diagram of step S30 in the method flow of the present invention;

[0068] Fig.11 This is a schematic diagram of step S40 in the method flow of the present invention;

[0069] Fig.12 It is a program schematic diagram of step S50 in the method flow of the present invention. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0071] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments:

[0072] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0073] It should be noted that when a module is referred to as being "disposed on" another module, it can be directly on the other module or indirectly on the other module. When a module is referred to as being "connected to" another module, it can be directly connected to the other module or indirectly connected to the other module.

[0074] In the description of the present application, "multiple" means two or more, unless otherwise clearly and specifically defined; "several" means one or more, unless otherwise clearly and specifically defined.

[0075] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0076] See also Figure 1 As shown, the present invention provides a sanitation operation supervision system, including a data acquisition module, a feature extraction module, a model training module, a monitoring implementation module, a public supervision module, a wireless communication module, a memory, an alarm, a processing center, and a smart phone; the data acquisition module, the feature extraction module, the model training module, the monitoring implementation module, the public supervision module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center.

[0077] The smart phone is respectively connected to the wireless communication module wireless network within the range of the wireless network or the Internet, wherein the wireless network includes the Internet of Things and Wi-Fi, and the Internet of Things includes but is not limited to Z-wave, ZigBee, AMQP, XMPP, websocket, LwM2M, MQTT, CoAP, HTTP, REST API, LoRaWAN, NB-IoT, Bluetooth and other protocol networks.

[0078] The wireless communication module is provided with a wireless network unit, which is responsible for sending and receiving wireless signals, and is automatically networked with the intelligent control terminal within an effective network range, and an effective wireless network connection is performed through the Internet of Things system.

[0079] The alarm compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory. If the standard is not met, it will automatically sound an alarm and notify to adjust the parameters to continue training; and compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory. If the standard is not met, it will automatically sound an alarm and notify to rework and re-sweep; and compares the actual sanitation score of the area with the sanitation evaluation standard score of the area stored in the memory. If the standard is not met, it will automatically sound an alarm and notify to continue to improve.

[0080] The memory is responsible for the information storage of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, and alarm, as well as the storage of model comprehensive evaluation standard points, road sanitation quality standards, and regional sanitation evaluation standard points.

[0081] The processing center is responsible for the information transmission of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, alarm, and memory. It is the hub center of the system and compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it will notify that monitoring can be carried out; if it does not meet the standard, it will be passed to the alarm and notified to adjust the parameters to continue training; and compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if it meets the standard, the sanitation operation is completed; if it does not meet the standard, it will be passed to the alarm and notified to rework and re-sweep; and compares the actual sanitation score of the area with the sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will be maintained continuously; if it does not meet the standard, it will be passed to the alarm and notified to continue to improve.

[0082] See also Figure 2 As shown, the data acquisition module includes an information acquisition unit, an image grayscale unit, a geometric change unit, a filtering and denoising unit, and a normalization processing unit. The images of the running trajectory and working status of the sanitation vehicle are obtained through the sanitation monitoring equipment and converted into recognizable sanitation operation information, and passed to the feature extraction module.

[0083] Furthermore, the information acquisition unit obtains image data of the cleanliness of each road surface in real time by installing BDS positioning and high-definition cameras on sanitation vehicles, providing a basis for subsequent model training and passing it to the image grayscale unit; the image grayscale unit obtains the grayscale road cleaning image according to the image grayscale calculation formula "f(I,j)=0.30R(I,j)+0.59G(I,j)+0.11B(I,j)" and passes it to the image translation unit; the geometric change unit obtains the road cleaning image after grayscale by translation, transposition, mirroring, rotation , scaling to correct the systematic error of image acquisition and the random error of instrument position, and pass it to the filtering and denoising unit; the filtering and denoising unit obtains the denoised road cleaning image according to the two-dimensional median filtering calculation formula "g(x,y)=med{f(xk,yl),(k,l∈W)}", and passes it to the normalization processing unit; the normalization processing unit obtains the normalized road cleaning image data according to the image normalization conversion calculation formula "Xn=(X0-Xmin) / Xmax-Xmin)" for subsequent feature extraction.

[0084] See also Figure 3 As shown, the feature extraction module includes a data input unit, a convolution input unit, a pooling conversion unit, a fully connected layer unit, an output conversion unit, and a feature extraction unit, which is responsible for extracting the color, texture, and shape features of the road based on the preprocessed data for subsequent model training and passing them to the model training module.

[0085] Further, the data input unit calculates the filling size according to the formula "P H =[(Ho-1)*S H +K H -H i] / 2, PW = [(Wo-1)*SW + KW-Wi] / 2" to obtain the padding data and input it, and pass it to the convolution input unit; the convolution input unit obtains the convolution input data according to the convolution input calculation formula "z(t) = ∫x(m)y(tm)dm" and passes it to the pooling conversion unit; the pooling conversion unit obtains the maximum pixel value after the maximum pooling according to the maximum pooling calculation formula "Z(i, j) = max(X[i*PS(i+1)*PS, j*PS(j+1)*PS])" to enter the fully connected layer and pass it to the fully connected The connection layer unit; the fully connected layer unit obtains the connection layer output data according to the fully connected layer calculation formula "y=f(∑(wn*xn)+b)" and enters the output layer, and passes it to the output conversion unit; the output conversion unit obtains the output value size after convolution according to the output layer conversion calculation formula "N=(P-F+2C) / S+1", and passes it to the feature extraction unit; the feature extraction unit extracts the key features of road pollution, road conditions, environmental sanitation, road signs, and environmental changes of road cleaning according to the processed image data, so as to facilitate subsequent modeling training and recognition.

[0086] See also Figure 4 As shown, the model training module includes a training simulation unit, a model evaluation unit, an abnormality identification unit, a cost budget unit, and a parameter determination unit. A convolutional neural network is used to perform model training according to road cleaning characteristics to learn to distinguish features between normal and abnormal images and continuously adjust model parameters to improve model accuracy, which are then passed to the monitoring implementation module.

[0087] Further, the training simulation unit trains the machine learning model using the extracted feature information, identifies whether the road monitoring data is abnormal and its abnormal category, and transmits it to the model evaluation unit; the model evaluation unit obtains the comprehensive evaluation score after model training according to the model comprehensive evaluation score calculation formula "F1=2*(Pr*Re) / (Pr+Re)", and transmits it to the processing center; the abnormality recognition unit inputs the trained image into the trained convolutional neural network model and matches it with the road cleaning monitoring image obtained in real time and confirms its abnormal category, and transmits it to the cost budget unit; the cost budget unit calculates the daily cleaning cost of the sanitation vehicle according to the formula "J g =Lj*C j +L j *Xc+L j / D z *D y *G r "The daily cleaning cost of the sanitation vehicle is obtained and passed to the parameter determination unit; the parameter determination unit obtains the resource configuration corresponding to the model and the automatic operation parameters of the sanitation monitoring equipment according to the road cleaning abnormality category.

[0088] See also Figure 5As shown, the monitoring implementation module includes a cleaning operation unit, an information collection unit, an emergency dispatch unit, and a training and education unit. Through the video monitoring system, the sanitation operations in the area are timely and effectively monitored according to the trained model, and training is carried out to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations and pass them on to the public supervision module.

[0089] Furthermore, the cleaning operation unit performs automatic sanitation cleaning operations by matching the type and number of sanitation vehicles, the number of sanitation workers and the cleaning process with the target cleaning section according to the trained model, and transmits the results to the information collection unit; the information collection unit obtains images of roads in the area through the set high-definition camera and converts them into sanitation quality information of the road surface to ensure that the sanitation quality of the roads meets the standards, and transmits the results to the processing center; the emergency dispatch unit issues an alarm based on abnormal feedback and promptly assigns the corresponding sanitation vehicles and equips sanitation workers to take special measures through the trained model to ensure that difficult areas are cleaned up, and transmits the results to the training and education unit; the training and education unit improves the professional quality and skills of sanitation workers through online training and distance education according to the model requirements, so that sanitation workers can save themselves and each other and work safely in an emergency.

[0090] See also Figure 6 As shown, the public supervision module includes a garbage classification unit, a public participation unit, and an information aggregation unit. By establishing a sanitation complaint service center so that the public can report sanitation problems and summarize and improve them, a good atmosphere of "everyone participates and everyone supervises" is formed and the information is transmitted to the processing center.

[0091] Furthermore, the garbage classification unit uses the trained model to classify and supervise garbage according to environmental protection laws and regulations, and automatically generates inspection, assessment and law enforcement work orders, and passes them to the public participation unit; the public participation unit is connected to the sanitation complaint service center by setting up mini-programs, apps, and hotlines to facilitate the public to report sanitation problems so that subsequent sanitation operations can be continuously improved, and passed to the information aggregation unit; the information aggregation unit compares the sanitation maintenance information of each road obtained by high-definition cameras with the complaint information to obtain the severity of the complaint results, and summarizes them into sanitation scores for each road as a basis for subsequent sanitation improvements.

[0092] System operation working principle:

[0093] Before the operation, the data acquisition module obtains the running track and working status images of the sanitation vehicle through the sanitation monitoring equipment and converts them into recognizable sanitation operation information, and passes them to the feature extraction module; then the feature extraction module extracts the color, texture, and shape features of the road according to the preprocessed data for subsequent model training, and passes them to the model training module; the model training module uses the convolutional neural network to train the model according to the road cleaning characteristics to learn the features of distinguishing normal and abnormal images and continuously adjust the model parameters to improve the model accuracy, and passes it to the monitoring implementation module; the control monitoring implementation module uses the video monitoring system to monitor the sanitation operations of the area in a timely and effective manner according to the trained model and trains to improve the operation efficiency to ensure the effectiveness of the sanitation operations, and passes it to the public supervision module; then the public supervision module establishes a sanitation complaint service center so that the public can reflect sanitation problems and summarize and improve them, forming a good atmosphere of "everyone participates and everyone supervises", and passes it to the processing center; the processing center compares the actual sanitation score of the area with the sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will continue to maintain it; if it does not meet the standard, it will be passed to the alarm and notify continuous improvement until it meets the standard.

[0094] When sanitation managers are outdoors or out of town, they can use their smartphones to automatically connect to wireless networks or the Internet through wireless communication modules, thereby combining smartphones with the Internet and the Internet of Things. Through mobile phone APP, they can remotely monitor the operation of each sanitation station in real time, realize real-time monitoring of sanitation operations, and promptly dispatch sanitation vehicles to support operations on other roads, thereby improving sanitation management efficiency and reducing sanitation management costs.

[0095] See also Figure 7 As shown, a sanitation operation supervision system provided by the present invention is implemented by a sanitation operation supervision method, and the method comprises the following steps:

[0096] S10, before the operation, the data acquisition module obtains the running track and working status image of the sanitation vehicle through the sanitation monitoring equipment and converts it into recognizable sanitation operation information, and passes it to the feature extraction module;

[0097] See also Figure 8 As shown, the step S10 comprises the following steps:

[0098] S11, the information collection unit installs BDS positioning and high-definition cameras on sanitation vehicles to obtain image data of the cleanliness of each road in real time, providing a basis for subsequent model training and passing it to the image grayscale unit;

[0099] It is further explained that a regional road cleanliness status database is established through the historical accumulation of cleanliness conditions of various road surfaces, including an internal database of various cleanliness conditions accumulated from daily inspections of various roads within the unit or historical inspection data of cleanliness conditions of the same road sections obtained from other platforms such as public security monitoring and community monitoring, as well as a statistical sharing database of road cleanliness conditions of public media in the area, and third-party data platforms such as urban management departments and municipal departments. The sanitation and cleanliness status image data of the same road is automatically obtained through data collectors to ensure data monitoring and integrity, comply with relevant data protection regulations, and check data monitoring and integrity to ensure that the data has no duplication, anomalies or missing values.

[0100] S12, the image grayscale unit obtains the grayscaled road cleaning image according to the image grayscale calculation formula "f(i, j) = 0.30R(i, j) + 0.59G(i, j) + 0.11B(i, j), f(i, j) is the grayscale image, R(i, j), G(i, j), B(i, j) are the original images of three colors respectively", and transmits it to the image translation unit;

[0101] It is further explained that the grayscale image is obtained through the grayscale calculation formula, the sampling value in the input signal is checked to determine whether it represents the signal itself, and the values ​​in the window are sorted by using an observation window composed of an odd number of samples, the middle value is taken as the output, the earliest value is discarded, and a new sample is obtained. The above calculation process is repeated to remove noise in the image or other signals, so as to subsequently extract the key features of the road cleaning image, including organizational morphology such as ferrite, pearlite, martensite, appearance monitoring such as surface roughness, cracks, inclusions, internal anomalies such as pores, shrinkage holes, cracks, grain size, etc.

[0102] S13, the geometric change unit corrects the systematic error of image acquisition and the random error of instrument position by translation, transposition, mirroring, rotation and scaling to improve the generalization ability of the model, and transmits it to the filtering and denoising unit;

[0103] To further illustrate, the image translation conversion calculation formula "f(x,y)=f(i+△x,j+△y)" is used to move all the pixels on the original image according to the given offset to change the image position, but the image content is not changed to obtain the translated image, that is, x=i+△x, y=j+△y, where f(x,y) is the translated image, f(i,j) is the grayscale image, △x and △y are the translation amounts on the x and y axes respectively; the image transposition calculation formula "f(x1,y1)=f(x0,y0)" is used to transpose the x and y coordinates of the image pixels, where (x0,y0) is the point on the original image, (x1,y1) is the grayscale image, and (x1,y1) is the grayscale image. is the corresponding new coordinate point after transposition, so the width and length of the image will be interchanged; the image after mirror transposition is obtained through the image mirror transformation calculation formula "f1(x',y')=f(M-1-x1,y1),f2(x',y')=f(x1,N-1-y1),f3(x',y')=f(M-1-x1,N-1-y1)", where f1, f1, f1 are horizontal mirror, vertical mirror, and diagonal mirror respectively, (x',y') are the corresponding coordinates in the mirrored image, and (x1,y1) are the pixel coordinates in the original image; the image rotation calculation formula "i'=i*cos(θ)-j sin(θ), j'=i*sin(θ)+j cos(θ)” to obtain the rotated image, where (i, j) is the pixel coordinate in the original image f(i, j), and (i', j') is the coordinate of the corresponding pixel in the image g(i', j') after the rotation transformation; the rotated image is obtained by the image rotation calculation formula "x'=ax+by+c, y'=dx+ey+f", where (x', y') is the coordinate in the rotated image, (x, y) is the coordinate in the original image, a and d are the scaling factor transformation coefficients, b and e are the rotation factor transformation coefficients, and c and f are the translation factor transformation coefficients; the original image is subjected to one or two or more of the above-mentioned geometric transformations of translation, transposition, mirroring, rotation and scaling according to respective needs.

[0104] S14, the filtering and denoising unit obtains the denoised road cleaning image according to the two-dimensional median filtering calculation formula "g(x, y) = med{f(xk, yl), (k, l∈W)}, f(x, y) is the original image, g(x, y) is the filtered image, K is the template window size, and W is the neighborhood window" to suppress noise and improve the clarity of the image, and transmits it to the normalization processing unit;

[0105] To further illustrate, the above-mentioned two-dimensional median filtering calculation formula replaces the value of a point in a digital image or digital sequence with the median value of each point in a neighborhood of the point, so that the surrounding pixel values ​​are close to the true value, thereby eliminating isolated noise points. That is, a two-dimensional sliding template with a certain structure is used to sort the pixels in the board according to the size of the pixel value to generate a monotonically rising or falling two-dimensional data sequence. The filtering output is usually a 3*3, 5*5 area, and can also be a linear, circular, cross, annular and other different shapes. k is half the width of the template window, usually an odd number, for example: a 3*3 template window has k=1; a 5*5 template window has k=2.

[0106] S15. The normalization processing unit obtains the normalized road cleaning image data according to the image normalization conversion calculation formula "Xn = (X0-Xmin) / Xmax-Xmin), X2 is the normalized image matrix, X0 is the original image matrix, Xmax and Xmin are the maximum and minimum values ​​in the original image respectively" to facilitate subsequent feature extraction.

[0107] Further explanation: the pixel value of the road cleaning image is normalized from the range of 0-255 to between 0-1 by the image normalization conversion calculation formula "Xn = (X0-Xmin) / Xmax-Xmin)", the distribution distance of the pixel value is reduced, and the invariant moment of the road cleaning image is used to obtain a set of parameters so that it can eliminate the influence of other transformation functions on the transformation of the road cleaning image, that is, it is converted into a unique standard form to resist affine transformation, so that the road cleaning image can resist the attack of geometric transformation, and the invariant in the road cleaning image can be found, so that these road cleaning images are the same or the same series, which can reduce the dynamic range of the road cleaning image, which is helpful for subsequent further processing and analysis; the normalized road cleaning image is smoothed to suppress the grayscale jump of the image, and sharpened to enhance the edge and details of the image.

[0108] S20, the feature extraction module extracts the features of the color, texture and shape of the road according to the preprocessed data for subsequent model training, and transmits them to the model training module;

[0109] See also Fig. 9 As shown, the step S20 comprises the following steps:

[0110] S21, the data input unit calculates the formula "P" according to the filling size H =[(Ho-1)*S H +K H -Hi] / 2,P W =[(Wo-1)*S W +K W -Wi] / 2,P H, P W are the padding sizes in height and width respectively, Hi and Wi are the height and width of the input feature map respectively, and K H , KW are the height and width of the convolution kernel, SH and SW are the values ​​of the step size in the height and width directions, Ho and Wo are the height and width of the output feature map, respectively. The padding data is obtained and input, and passed to the convolution input unit;

[0111] It is further explained that after the image is normalized, the data is used to establish a convolutional neural network model containing an input layer, a convolution layer, a pooling layer, a fully connected layer, and an output layer, and then the calculated padding data is input. Since the convolution kernel in the convolution layer contains a weight coefficient, and the pooling layer does not contain a weight coefficient, the pooling layer is not an independent layer. The convolution layer is used to extract local features of the image, the pooling layer is used to reduce the dimension of the feature map, and the fully connected layer is used to integrate features and classify them; the amount of padding is determined by the size of the convolution kernel, the step size, and the padding size. In order to keep the height and width of the input and output images the same, or to reduce the size quickly in continuous convolution operations, the amount is usually set to the convolution kernel size minus one.

[0112] S22, the convolution input unit obtains the convolution input data according to the convolution input calculation formula “z(t)=∫x(m)y(tm)dm, z(t) is the output function, x(t) is the input function, y(t) is the convolution kernel function, and dm is the integral differential element of the variable m”, and passes it to the pooling conversion unit;

[0113] Further explanation: the integral differential element dm is the "width" or "increment" when integrating each infinitesimal interval of the variable m. By accumulating the function values ​​on these infinitesimal intervals (i.e. x(m)y(tm)), the cumulative effect on the entire definition domain, i.e. z(t), is obtained. dm is the integral differential element of the variable m, i.e. the infinitesimal quantity in the integration process; then the convolution calculation is performed to obtain the convolution value, which enters the convolution layer and is activated as the feature value output by the convolution layer or the fully connected layer; the function activation is calculated by the ReLU function. The formula "f(x)=max(0,x+Y)" is used to obtain the activation function value and enter the pooling layer. In the formula, f(x) is the activated function, x is the characteristic value of the output of the convolutional layer or the fully connected layer, and Y is a random variable. When the input x is greater than 0, the output is x, and when the input x is less than or equal to 0, the output is 0. When x=0, it is not differentiable, that is, when the input is a positive number, the original value is maintained, and when the input is a negative number, the output is 0; but the ReLU activation function is used after the convolutional layer to increase the nonlinearity of the network and promote the model to learn complex patterns.

[0114] S23, the pooling conversion unit calculates the maximum pooling formula "Z(i, j) = max(X[i*P S (i+1)*P S,j*P S (j+1)*P S ]), Z(i, j) is a pixel value in the output feature map after pooling, i and j are the positions in the output feature map respectively, max is to find the maximum value of all pixel values ​​in the input window, X is the input feature map, PS is the window size of the pooling operation. The maximum pixel value after the maximum pooling enters the fully connected layer to retain important feature information and pass it to the fully connected layer unit;

[0115] Further explanation: According to the convolution input data, the average pooling calculation formula "Z(i, j) = mean(X[i*P S (i+1)*P S , j*PS(j+1)*PS])" obtains the average value of all values ​​in each area, which can retain more information in the feature map, help improve the accuracy of the model, retain more detail information, and is used in the deeper layers of the network to ensure the integrity of the information. In the formula, Z(i, j) is a pixel value in the output feature map after pooling, and mean is the average value of all pixel values ​​in the input window; then the maximum pooling calculation formula is used to select the largest value from each area of ​​the input feature map as the output, retaining the most significant features in the feature map, reducing the size of the feature map, and losing some useful information. Since the maximum value of each area is focused on, the grain size characteristics are retained, and features with better classification recognition are selected.

[0116] S24, the fully connected layer unit obtains the connection layer output data according to the fully connected layer calculation formula "y = f(∑(wn*xn) + b), y is the output result, f is the activation function, wn is the weight of the nth input feature, xn is the nth input feature, n is the dimension of the input feature, and b is the bias" and enters the output layer for subsequent output calculation and is passed to the output conversion unit;

[0117] Further explanation: after the pooled data is subjected to the fully connected layer, the output value after convolution is calculated according to the fully connected layer calculation formula "y=f(∑(wn*xn)+b)", where y is the output result, f is the activation function, wn is the weight of the nth input feature, xn is the nth input feature, n is the dimension of the input feature, and b is the bias; the "y=f(∑(wn*xn)+b)" is calculated from "y=f(w*x+b)" as "y=f(w1*x1+w2*x2+…+w n *x n +b)", where w is the weight and xn is the input feature; the features extracted by the convolution layer and the pooling layer are integrated and classified through the fully connected layer, so that each neuron is connected to all the neurons in the previous layer, and the output is generated through weighted summation and nonlinear activation function.

[0118] S25, the output conversion unit obtains the output value size after convolution according to the output layer conversion calculation formula "N = (P-F + 2C) / S + 1, N is the output size after convolution, P is the input size before convolution, F is the convolution kernel size, C is the number of layers of 0 added around the image, and S is the step size" for subsequent model analysis and training, and passes it to the feature extraction unit;

[0119] Further explanation, the output size N is the size of the feature map after the convolution operation, the input size P is the width or height of the input image or feature map, and the padding size P in the height direction is obtained by calculating the above step S21. H , the padding size P in the width direction W The convolution kernel size F is obtained by calculating the above step S22. The number of layers C of adding 0 around the image is the number of layers of adding 0 around the input image or feature map to control the size of the output feature map, which is also obtained by the above network construction. The step size S is the distance that the convolution kernel slides on the input image or feature map, which is obtained by the above calculation.

[0120] S26, the feature extraction unit extracts key features of road pollution, road conditions, environmental sanitation, road signs, and environmental changes of road cleanliness according to the processed image data, so as to facilitate subsequent modeling training and recognition.

[0121] To further explain, various road sanitation anomalies are classified as follows: "poor sanitation" is category 1, "messy greenery" is category 2, "garbage chaos" is category 3, "dilapidated facilities" is category 4, and "littering" is category 5; when category 1 is a positive sample, the remaining categories are negative samples, that is, the labeled data is (1, 0, 0, 0, 0), when category 2 is a positive sample, the remaining categories are negative samples, that is, the labeled data is (0, 1, 0, 0, 0), when category 3 is a positive sample, the remaining categories are negative samples, that is, the labeled data is (0, 0, 1, 0, 0), when category 4 is a positive sample, the remaining categories are negative samples, that is, the labeled data is (0, 0, 0, 1, 0), when category 5 is a positive sample, the remaining categories are negative samples, that is, the labeled data is (0, 0, 0, 0, 1).

[0122] S30, the model training module uses a convolutional neural network to perform model training according to the road cleaning characteristics to learn the features of distinguishing normal and abnormal images and continuously adjusts the model parameters to improve the model accuracy, and transmits the model to the monitoring implementation module;

[0123] See also Fig.10 As shown, the step S30 comprises the following steps:

[0124] S31, the training simulation unit trains the machine learning model using the extracted feature information, identifies whether the road monitoring data is abnormal and its abnormal category, and transmits it to the model evaluation unit;

[0125] It is further explained that the processed image data is divided into a training set and a test set. According to the characteristics of road cleaning monitoring anomalies, a 10-layer deep convolutional neural network structure is constructed to train the road cleaning image data set that has been labeled with categories to obtain the probability of the corresponding category; and a labeled data set containing road cleaning samples is established, including the bounding box coordinates and category labels of the road cleaning area. These data are used to train the YOLOv5 model, and the hyperparameters are adjusted to optimize the detection performance; the convolutional neural network is used to train the labeled category road cleaning image data set to obtain the probability of the corresponding category, and the back propagation principle is used to continuously train the data until the total loss function of all samples is minimized, so that the probability of the predicted sample is close to 0 in the remaining categories and close to 1 in this category.

[0126] S32. The model evaluation unit obtains the comprehensive evaluation score after model training according to the model comprehensive evaluation score calculation formula "F1 = 2*(Pr*Re) / (Pr+Re, F1 is the model comprehensive evaluation score, Pr is the precision, and Re is the recall rate" to ensure the accuracy of model verification and transmit it to the processing center;

[0127] Further explanation: the precision rate Pr and recall rate Re are respectively calculated by the formula “Pr=∑T P / ∑(T P +F P ), Re = ∑T P / ∑T P +∑F N "Get, where F1 is the comprehensive evaluation score of the model, T P is the number of samples correctly predicted as positive, F P is the number of samples incorrectly predicted as positive, F N is the number of samples that are incorrectly predicted as negative; the number of samples that are correctly predicted as positive is T P Also called true positive, true example, the number of samples that are incorrectly predicted to be positive F P Also called false positive, false positive example, the number of samples that are incorrectly predicted to be negative class F N Also called false negatives, false negatives, to verify the generalization ability and accuracy of the model; the generalization ability and accuracy of the model are verified by cross-validation to obtain the model evaluation score, ensuring that the model accurately predicts the nature of road cleaning under different conditions; during the training process, the weights and biases in the network are updated through the back-propagation algorithm to train the neural network so that it gradually learns the abnormal characteristics of road cleaning in the image.

[0128] S33, the processing center compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it notifies that monitoring can be carried out; if it does not meet the standard, it is transmitted to the alarm and notifies to adjust the parameters to continue training;

[0129] It is further explained that the comprehensive evaluation standard of the model is divided into a comprehensive evaluation score based on the training of similar road cleaning anomaly models in the history of the sanitation unit itself and a comprehensive evaluation score based on the training of similar road cleaning anomaly models in the road cleaning condition database of the area to obtain the optimal evaluation score; because the size of the model recall data is affected by factors such as data monitoring, model design, amount of training data, and contextual information; when the data in the database is incomplete, inaccurate or redundant, it is difficult for the system to accurately find matching data, resulting in false positives and false negatives; because in deep learning, deeper network structures and more training data can enhance the expressive power of the model, enabling it to better distinguish between targets and backgrounds.

[0130] S34, the abnormality recognition unit inputs the trained image into the trained convolutional neural network model to match the road cleaning monitoring image obtained in real time and confirm its abnormal category, and transmits it to the cost budget unit;

[0131] To further illustrate, the image is input into the model and the trained model is used to predict the new monitoring image. The type of road cleaning anomaly is judged based on the output result. If the model's judgment on the road cleaning anomaly reaches a certain threshold, the early warning system is triggered and corresponding solutions are taken in time. The road cleaning anomaly types include: dirty and poor sanitation, messy greening, chaotic garbage, dilapidated facilities, and random garbage dumping. These trained models are applied to subsequent road cleaning detection. By inputting new sample data, the trained model and its corresponding detection process parameters can quickly obtain detection results, replacing manual judgment of the quality of road cleaning.

[0132] S35. The cost budget unit obtains the daily cleaning cost of the sanitation vehicle according to the daily cleaning cost calculation formula of "Jg = Lj*Cj + Lj*Xc + Lj / Dz*Dy*Gr, Jg is the daily cleaning cost of the sanitation vehicle (yuan / day), Lj is the daily cleaning mileage of the sanitation vehicle (km / day), Cj is the cleaning material fee of the sanitation vehicle per kilometer (yuan / km), Xc is the cleaning repair fee of the sanitation vehicle per kilometer (yuan / km), Dz is the daily cleaning mileage of the sanitation vehicle (km / day), Dy is the number of sanitation vehicle operators (persons), and Gr is the daily salary of the sanitation vehicle personnel (yuan / person / day)" for subsequent resource allocation, and passes it to the parameter determination unit;

[0133] To further explain, the daily cleaning mileage Lj of the sanitation vehicle, the cleaning material fee Cj per kilometer of the sanitation vehicle, the cleaning repair fee Xc per kilometer of the sanitation vehicle, the daily cleaning mileage Dz of the sanitation vehicle, the number of sanitation vehicle operators Dy, and the daily salary Gr of the sanitation vehicle personnel are the historical average daily cleaning costs of the sanitation vehicle on a fixed road, excluding the costs of repeated cleaning when there are difficult problems on the road, or the costs of external support cleaning; the daily cleaning costs for the same sanitation vehicle on each road may be the same or different; the daily cleaning costs of different sanitation vehicles on the same road may be the same or different; the cleaning costs of the same sanitation vehicle may be the same or different if different personnel are equipped with it.

[0134] S36. The parameter determination unit obtains the resource configuration of each road corresponding to the model and the automatic operation parameters of various sanitation monitoring equipment according to the road cleaning abnormality category, so as to automatically match and apply the subsequent automatic monitoring of the sanitation monitoring equipment.

[0135] It is further explained that the resource allocation of each road includes the number and type of sanitation vehicles and the number of sanitation personnel on each road, which is obtained by referring to the historical data of the configuration of sanitation vehicles on each road, and the spare sanitation vehicles and sanitation vehicles that have completed sanitation tasks are called up in emergency situations; the automatic operation parameters of various sanitation monitoring equipment include: wide-angle camera: shutter speed, imaging mode, light source, storage format, etc.; BSD pedestrian detection system: detection threshold, processing speed, etc.; video monitoring system: frame rate, resolution, bit rate and three-stream, etc.; GPS positioning system: signal frequency, propagation time, operating frequency, etc.; on-board hard disk recorder: video resolution, network mode, etc., to ensure that the model matches the operating parameters of the sanitation monitoring equipment during subsequent actual automatic detection, and automatically controls the sanitation monitoring equipment for automatic monitoring through program control instructions to ensure the accuracy of automatic monitoring of road sanitation cleaning.

[0136] S40, the monitoring implementation module uses the video monitoring system to monitor the sanitation operations in the area in a timely and effective manner according to the trained model and provides training to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations, and transmits it to the public supervision module;

[0137] See also Fig.11 As shown, the step S40 comprises the following steps:

[0138] S41, the cleaning operation unit performs automatic sanitation cleaning operations according to the trained model and the target cleaning section, matching the type and number of sanitation vehicles, the number of sanitation workers and the cleaning process, and transmits the information to the information collection unit;

[0139] To further explain, according to the system's sanitation cleaning dispatch orders for each road, the corresponding types and numbers of sanitation vehicles and the number of sanitation personnel and the coordinated operations are arranged; for flat sections where operations can be automated, sanitation vehicles assigned by sprinkler trucks, washing and sweeping trucks, small sweepers, etc., follow the steps of "one sprinkling, one sweeping, one washing, one flushing, and one spraying" to complete the three processes of "coarse, fine, and lowering" to improve the level of refined road cleaning operations; for sections that cannot be automated or sections such as sidewalks or flower beds, manual cleaning operations are implemented, cleaning from the sidewalks to the motor vehicle road surface, and then the sanitation vehicles clean and collect the dirt to ensure that the road surface is free of garbage bags, paper scraps, plastic bags, bricks and stones, human and animal feces, loose soil on the road surface, leaves and other debris.

[0140] S42, the information collection unit obtains images of roads in the area through the set high-definition camera and converts them into environmental sanitation quality information of the road surface to ensure that the environmental sanitation quality of the road meets the standards and transmits them to the processing center;

[0141] It is further explained that the high-definition camera is an ultra-wide-angle clarity camera that can take 360-degree photos, and can automatically focus, automatically adjust shutter speed, light source, automatically select imaging mode, storage format, etc.; by setting up a number of high-definition cameras at key locations in public areas such as roads and squares in economically prosperous areas, and installing high-definition cameras at the four corners of sanitation vehicles such as sprinkler trucks, tow trucks, sweepers, garbage trucks, and sanitation patrol cars, the roads in prosperous areas are monitored by set cameras every day and combined with irregular patrols of various sanitation vehicles every day. Among them, roads in remote areas are monitored by regular patrols of sanitation vehicles. Some economically developed cities can also use drones for low-altitude inspections, and timely monitor abnormal events such as daily sanitation cleaning conditions on roads, garbage generation on roads after cleaning, damage to sanitation facilities, and damage to road surfaces, including automatic detection of littering, overflowing garbage cans, damaged sanitation facilities, abnormal sanitation vehicle operation sites, and generation of warning images and violation information, thereby realizing real-time supervision and automated management, greatly improving the efficiency of sanitation operations.

[0142] S43, the processing center compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if the standard is met, the sanitation work is completed; if the standard is not met, the alarm is transmitted to the alarm and the work is notified to be reworked and re-sweeped;

[0143] It is further explained that the road sanitation quality standards include that there is no exposed garbage, no fruit peels and paper scraps, no sand and gravel, no mud and dirty water, no spilled food, no human and animal feces, and complete sanitation facilities in the target cleaning areas around the road surface, such as the road surface, leaking holes, wall roots, tree holes, guardrails, and fruit peel bins; for places where oil stains are difficult to remove, sanitation vehicles are used to spray oil cleaning fluid for cleaning, and road sanitation quality information is obtained by high-definition cameras installed on the sanitation vehicles or on the road surface. If there are still debris on the road surface after cleaning, the sanitation vehicle is notified to clean it again. If there are leaking holes, wall roots, tree holes, guardrails and other areas that cannot be cleaned automatically or areas that cannot be cleaned completely by automatic cleaning, the system will automatically transmit information to the accompanying cleaning personnel or nearby cleaning personnel to perform manual cleaning operations until the sanitation quality of the target cleaning areas such as leaking eyes, wall roots, tree holes, guardrails and other target cleaning areas within the cleaning range of the cleaning area and its surrounding areas meets the standards; if difficult and complicated problems are encountered, an alarm will be issued to dispatch resources for emergency treatment, and it will be used as emergency training content.

[0144] S44, the emergency dispatch unit sends an alarm based on the abnormal feedback and uses the trained model to promptly assign the corresponding sanitation vehicle and equip sanitation workers to take special disposal measures to ensure that the difficult areas are cleaned up and passed on to the training and education unit;

[0145] It is further explained that by installing GPS positioning systems on patrol cars, garbage trucks, sprinkler trucks and other sanitation vehicles, the system can track and record the geographical location of the vehicle in a timely manner, and monitor and record the vehicle's operating status and the work of the staff throughout the process, so that managers can understand the vehicle's geographical location and operating status in real time, and record images inside and outside the vehicle in the event of an accident, and send them to the command center in a timely manner, providing strong evidence for handling accidents and tracing responsibility for accidents, and helping managers to reasonably dispatch vehicles to handle various sanitation anomalies on various sections of the road within the jurisdiction, assist in sanitation cleaning on other sections of the road, and improve work efficiency; according to weather forecasts, traffic conditions and other data, timely guide sanitation vehicles and sanitation workers to adjust countermeasures and take corresponding protective measures: in hot weather, the system can automatically adjust the operating time to find a more suitable cool spot for sanitation workers; in rainy and snowy weather, the system can automatically remind sanitation workers to wear non-slip shoes and pay attention to driving safety.

[0146] S45. The training and education unit improves the professional quality and skills of sanitation workers through online training and distance education according to the requirements of the model, so that sanitation workers can save themselves and each other and work safely in emergency situations.

[0147] Further explanation: According to the daily accumulated road frequent abnormal areas and their model management requirements, existing sanitation personnel are trained through online training, distance education, etc., and sanitation workers are trained online through mobile phones to improve their professional quality and skill level. The latest operating specifications, safety knowledge, first aid skills and other training are provided to sanitation workers, so that sanitation workers can save themselves and each other in emergencies; when the dispatch order is received from the management center, sanitation vehicles and sanitation workers are equipped to go to the corresponding areas for emergency sanitation operations. Sprinkler trucks, tow trucks, and sweepers can respectively spray water, remove obstacles on the road, and remove garbage and foreign objects. These sanitation vehicles can be multi-functional sanitation vehicles such as sprinkling water, clearing obstacles, and sweeping the floor to ensure that urban roads are clean and tidy 24 hours a day, so as to realize the new model of AI+urban governance.

[0148] S50, the public supervision module establishes a sanitation complaint service center so that the public can report sanitation problems and summarize improvements, forming a good atmosphere of "everyone participates and everyone supervises", and then transmits them to the processing center;

[0149] See also Fig.12 As shown, the step S50 comprises the following steps:

[0150] S51, the garbage classification unit uses the trained model to classify and supervise garbage according to environmental protection laws and regulations, and automatically generates inspection, assessment and law enforcement work orders, and passes them to the public participation unit;

[0151] Further explanation: According to the trained model, the garbage of urban road sanitation is divided into: 1) Domestic garbage: food residues and organic waste such as chicken shells, bread, fruit peels, and vegetable leaves, which can be recycled through composting and other methods; 2) Non-recyclable garbage: waste tissues, plastic bags, disposable tableware and other non-recyclable waste, which are properly treated and disposed of to reduce the impact on the environment; 3) Hazardous garbage: syringes, thermometers, batteries and other garbage that are harmful to human health or the natural environment, which are specially treated to prevent harm to the environment and human health; 4) Recyclable garbage: plastic bottles, cartons, glass bottles and other garbage that can be recycled and reused, which can reduce the consumption of natural resources; 5) Special garbage: expired medicines, pesticides, fluorescent pens and other garbage that need special treatment to prevent harm to the environment and human health. In this way, AI technology is used to supervise garbage classification, reduce the difficulty of managing delayed delivery, and transform the supervision of garbage classification delivery from "human defense" to "technical defense".

[0152] S52. The public participation unit is connected to the sanitation complaint service center by setting up mini programs, apps, and hotlines to facilitate the public to report sanitation problems so that subsequent sanitation operations can be continuously improved and passed on to the information aggregation unit;

[0153] It is further explained that small programs, apps, and hotlines are posted at corresponding locations on the roads and promoted on platforms such as WeChat and Alipay, so that citizens can easily obtain small programs, apps, and hotlines to make complaints when they discover sanitation problems. This can make up for problems that cannot be discovered in sanitation operations, or difficult sanitation problems in some areas that cannot be solved. It can promote the sanitation department to gather strength to solve difficult sanitation problems, including: insufficient or backward sanitation facilities in newly developed areas and urban fringes, resulting in littering of domestic garbage; unclear responsibilities between urban management and gardening, water affairs, etc., resulting in no maintenance of public facilities such as parks and rivers; due to the lack of environmental awareness of some citizens, they casually discard packaging bags, cigarette butts, fruit peels and cores and other garbage; garbage dumping and exposed loess in green belts are frequent and difficult to maintain; and so on. Many difficult problems that cannot be discovered by the sanitation system can be easily discovered through public complaints, forming a good atmosphere of "everyone participates and everyone supervises."

[0154] S53. The information aggregation unit compares the sanitation maintenance information of each road obtained by the high-definition camera with the complaint information to obtain the severity of the complaint result, and summarizes it into a sanitation score for each road as a basis for subsequent sanitation improvement.

[0155] It is further explained that each road uses high-definition cameras installed at key garbage locations to obtain the sanitation maintenance conditions on the road, including whether the garbage is disposed in designated areas, whether it is disposed according to categories, whether the sanitation facilities are not deliberately damaged, etc., and summarizes sanitation maintenance information based on district level, streets, roads, and responsible units. It also classifies and summarizes sanitation information such as exposed garbage, fruit peels and paper scraps, sand and gravel, loose soil, mud and dirty water, spilled food, human and animal feces, and the integrity of sanitation facilities in the area, and subdivides it into sanitation target areas such as road surfaces, leaking eyes, wall roots, tree holes, guardrails, and fruit peel bins, so that corresponding measures can be taken to deal with them, and a five-level responsible sanitation monitoring and management system of "district level + street + road + responsible entity" is achieved, among which the responsible entities include enterprises or shops in the area and the corresponding sanitation personnel.

[0156] S60. The processing center compares the actual environmental sanitation score of the area with the environmental sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will be maintained; if it does not meet the standard, it will be transmitted to the alarm and notified to continue to improve until it meets the standard.

[0157] Further explanation: the sanitation scoring standards for the area include: 1) Cleaning and maintenance time: the number of times the road is cleaned and maintained every day, and the number of times the sidewalk is washed every week; 2) Cleanliness of the road surface: there is no waste, dirt, or other debris on the road surface, ditches, tree pits, wall roots, guardrails, flower beds, rainwater inlets, etc., and there are no large-scale stains and accumulated sewage on the road surface, and no ice in winter; 3) Cleanliness of collection facilities: the area around the trash can is clean and tidy, with no exposed garbage, scattered garbage, or overflowing garbage, and no stagnant water or mosquitoes and flies; 4) Cleanliness of waste bins: the waste bins and their surroundings are clean and tidy, without random postings, graffiti, and are cleaned in time, and there is no overflowing garbage; 5) Exposed Garbage and dust control: roads are in their original colors, with no large areas of exposed garbage, large garbage and construction waste, and cleaning operations do not generate dust; 6) Sanitation maintenance: garbage is placed in designated areas and by category, and sanitation facilities are fully maintained; etc. This information is obtained through high-definition cameras at key locations on the road, high-definition cameras on major garbage bins, and high-definition cameras on various sanitation vehicles to obtain information on the road sanitation maintenance of the responsible entity before sanitation cleaning, as well as the road sanitation cleaning situation after sanitation operations. This information is used as the scoring standard for each road. The sanitation requirements of each road are different, so its sanitation scoring standards are also different.

[0158] To further explain, the above steps are displayed in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows, unless clearly stated in this document. There is no strict order restriction for the execution of these steps, and these steps can also be executed in other orders, and some steps can include multiple sub-steps or multiple stages, which are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps; if the actual units of the parameters of the above formulas are inconsistent with those in the formulas, the machine control system will perform unit recognition and automatic conversion to the units in the formulas to ensure unit consistency.

[0159] Further explanation, the present invention is described according to the content implemented by the software program of a sanitation operation supervision and control system. Each sanitation operation supervision method is divided into several modules or units to implement the software program instructions generated by each step. The software program instructions include the above-mentioned sanitation operation supervision method.

[0160] A sanitation operation supervision system provided by the present invention also includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when the functional modules execute the computer program, the steps of the sanitation operation supervision method described in any one of the above are implemented; the computer-readable storage medium stores a computer program, and when the computer program is executed by the functional modules, the steps of the sanitation operation supervision method described in any one of the above are implemented.

[0161] It is further explained that the computer-readable storage medium includes a memory, which can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as the program and its instructions corresponding to a sanitation operation supervision method in the present invention, including information transmission instructions for each module; the memory executes various functional applications and data processing of each module by running the stored non-volatile software programs and instructions, that is, implements a sanitation operation supervision method in the above-mentioned method embodiment; one or more units are stored in the memory, and when executed by the one or more modules, a sanitation operation supervision method in any of the above-mentioned method embodiments is executed; the computer-readable storage medium stores computer executable instructions, which are executed by one or more modules, and can also be a sanitation operation supervision method in any of the above-mentioned method embodiments.

[0162] To further illustrate, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two, and may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device, including but not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above, wherein the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, carrying a computer-readable program code, and the propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above, and may also be any computer-readable medium other than a computer-readable storage medium, and the computer-readable signal medium may send, propagate or transmit a program for use by or in combination with an instruction execution system, apparatus or device, and the program code that may be contained may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0163] It is further explained that the computer program is divided into multiple modules / units, and the multiple modules / units are stored in the memory and executed by each module / unit to complete the present invention; the multiple modules / units can be a series of computer program instructions that can complete specific functions, and the instructions are used to describe the execution process of the computer program in each module / unit.

[0164] It is further explained that the contents described above in the present invention are written with reference to the contents implemented by the software copyrights such as "Shikelang Environmental Protection Mini Program Software (Registration Number: 2021SR0811335)", "Shikelang Garbage Classification Information Maintenance Service Management System (Registration Number: 2024SR0006817)", "Shikelang Garbage Classification and Recycling Management System (Registration Number: 2024SR0009429)", "Sanitation Vehicle Remote Visual Monitoring System (Registration Number: 2024SR1271467)", and "Sanitation Garbage Treatment Management Comprehensive Service Platform (Registration Number: 2024SR1277575)" applied for by our company on June 1, 2021.

[0165] The present invention provides a sanitation operation supervision system, which also includes a sanitation operation supervision control device, which is implemented by the sanitation operation supervision method described above.

[0166] To further explain, the control device is composed of the above-mentioned related equipment or devices. According to local sanitation needs, the camera can be fixedly installed on the sanitation vehicle, or it can be temporarily installed when the sanitation vehicle is working outside, and then removed after the work is completed. The cameras at each road monitoring position can be fixedly installed or temporarily installed, and removed when not needed. The camera on the trash can can be installed on the trash can or on a commanding point near the can, and can monitor the placement of garbage into the trash can and the details of its classification. Each camera transmits via wired or wireless. These equipment structures are not described in detail here, and there are no structural restrictions. The above only describes the software automatic control function.

[0167] It is obvious to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above, and that the present application can be implemented in other specific forms without departing from the spirit or essential features of the present application. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and the scope of the present application is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present application.

[0168] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical solution of the present application can be modified or replaced by equivalents, and all should be included in the scope of protection of the present application.

Claims

1. A sanitation operation supervision system, characterized in that: The system comprises a data acquisition module, a feature extraction module, a model training module, a monitoring implementation module, a public supervision module, a wireless communication module, a memory, an alarm, a processing center, and a smart phone; the data acquisition module, the feature extraction module, the model training module, the monitoring implementation module, the public supervision module, the wireless communication module, the memory, and the alarm are respectively connected to the processing center; the smart phone is respectively connected to the wireless network of the wireless communication module within the range of a wireless network or the Internet; the system is implemented by a sanitation operation supervision method, and the method comprises the following steps: S10, before the operation, the data acquisition module obtains the running track and working status image of the sanitation vehicle through the sanitation monitoring equipment and converts it into recognizable sanitation operation information, and passes it to the feature extraction module; S20, the feature extraction module extracts the features of the color, texture and shape of the road according to the preprocessed data for subsequent model training, and transmits them to the model training module; S30, the model training module uses a convolutional neural network to perform model training according to the road cleaning characteristics to learn the features of distinguishing normal and abnormal images and continuously adjust the model parameters, and pass them to the monitoring implementation module; S40, the monitoring implementation module uses the video monitoring system to monitor the sanitation operations in the area in a timely and effective manner according to the trained model and provides training to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations, and transmits it to the public supervision module; S50, the public supervision module establishes a sanitation complaint service center so that the public can report sanitation problems and summarize them for improvement, thus forming a good atmosphere of "everyone participates and everyone supervises", and then transmits them to the processing center; S60. The processing center compares the actual environmental sanitation score of the area with the environmental sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will be maintained; if it does not meet the standard, it will be transmitted to the alarm and notified to continue to improve until it meets the standard.

2. A sanitation operation supervision system as claimed in claim 1, characterized in that: Also includes: The smart phone is connected to the wireless network of the wireless communication module within the range of the wireless network or the Internet; The wireless communication module is provided with a wireless network unit, which is responsible for sending and receiving wireless signals, and automatically connects to the intelligent control terminal within the effective network range, and performs effective wireless network connection through the Internet of Things system; The alarm device compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to adjust the parameters to continue training; and the actual sanitation quality information of the road is compared with the sanitation quality standard of the road stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to rework and re-sweeping; and the actual sanitation score of the area is compared with the sanitation evaluation standard score of the area stored in the memory. If the standard is not met, a sound alarm is automatically issued and a notification is given to continue to improve; The memory is responsible for storing information of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, and alarm, as well as the model comprehensive evaluation standard score, road sanitation quality standard, and regional sanitation evaluation standard score; The processing center is responsible for the information transmission of the data acquisition module, feature extraction module, model training module, monitoring implementation module, public supervision module, wireless communication module, alarm, and memory. It is the hub center of the system and compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it will notify that it can be monitored; if it does not meet the standard, it will be passed to the alarm and notify to adjust the parameters to continue training; and compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if it meets the standard, the sanitation operation is completed; if it does not meet the standard, it will be passed to the alarm and notify to rework and re-sweep; and compares the actual sanitation score of the area with the sanitation assessment standard score of the area stored in the memory: if it meets the standard, it will continue to maintain; if it does not meet the standard, it will be passed to the alarm and notify continuous improvement; The data acquisition module includes an information acquisition unit, an image grayscale unit, a geometric change unit, a filtering and denoising unit, and a normalization processing unit. The image of the running track and working status of the sanitation vehicle is obtained through the sanitation monitoring equipment and converted into recognizable sanitation operation information, and then transmitted to the feature extraction module; The feature extraction module includes a data input unit, a convolution input unit, a pooling conversion unit, a fully connected layer unit, an output conversion unit, and a feature extraction unit, which is responsible for extracting the color, texture, and shape features of the road according to the preprocessed data for subsequent model training and passing them to the model training module; The model training module includes a training simulation unit, a model evaluation unit, an abnormality identification unit, a cost budget unit, and a parameter determination unit, which is responsible for using a convolutional neural network to perform model training based on road cleaning characteristics to learn the features that distinguish normal and abnormal images and continuously adjust model parameters to improve model accuracy, and pass them to the monitoring implementation module; The monitoring implementation module includes a cleaning operation unit, an information collection unit, an emergency dispatch unit, and a training and education unit. Through the video monitoring system, the sanitation operations in the area are monitored in a timely and effective manner according to the trained model, and training is conducted to improve the operation efficiency, so as to ensure the effectiveness of the sanitation operations, and the information is passed to the public supervision module; The public supervision module includes a garbage classification unit, a public participation unit, and an information aggregation unit. By establishing a sanitation complaint service center so that the public can report sanitation problems and summarize and improve them, a good atmosphere of "everyone participates and everyone supervises" is formed and the information is transmitted to the processing center.

3. A sanitation operation supervision system as claimed in claim 1, characterized in that: The method further comprises: Described step S10, comprises the following steps: S11, the information collection unit installs BDS positioning and high-definition cameras on sanitation vehicles to obtain image data of the cleanliness of each road in real time, providing a basis for subsequent model training and passing it to the image grayscale unit; S12, the image grayscale unit obtains the grayscaled road cleaning image according to the image grayscale calculation formula "f(i, j) = 0.30R(i, j) + 0.59G(i, j) + 0.11B(i, j), f(i, j) is the grayscale image, R(i, j), G(i, j), B(i, j) are the original images of three colors respectively", and transmits it to the image translation unit; S13, the geometric change unit corrects the systematic error of image acquisition and the random error of instrument position by translation, transposition, mirroring, rotation and scaling to improve the generalization ability of the model, and transmits it to the filtering and denoising unit; S14, the filtering and denoising unit obtains the denoised road cleaning image according to the two-dimensional median filtering calculation formula "g(x, y) = med{f(xk, yl), (k, l∈W)}, f(x, y) is the original image, g(x, y) is the filtered image, K is the template window size, and W is the neighborhood window" to suppress noise and improve the clarity of the image, and transmits it to the normalization processing unit; S15, the normalization processing unit obtains the normalized road cleaning image data according to the image normalization conversion calculation formula "Xn=(X0-Xmin) / Xmax-Xmin), X2 is the normalized image matrix, X0 is the original image matrix, Xmax and Xmin are the maximum value and minimum value in the original image respectively", so as to extract the subsequent features; Described step S20, comprises the following steps: S21, the data input unit calculates the filling size according to the formula "P H =[(Ho-1)*S H +K H -H i ] / 2, P W =[(Wo-1)*S W +K W -Wi] / 2,P H , P W are the padding sizes in height and width respectively, Hi and Wi are the height and width of the input feature map respectively, and K H , K W are the height and width of the convolution kernel, S H , S W Where Ho and Wo are the values ​​of the step length in the height and width directions, respectively. Ho and Wo are the height and width of the output feature map, respectively. The padding data is obtained and input, and passed to the convolution input unit. S22, the convolution input unit obtains the convolution input data according to the convolution input calculation formula "z(t) = ∫x(m)y(tm)dm, z(t) is the output function, x(t) is the input function, y(t) is the convolution kernel function, and dm is the integral differential element of the variable m", and passes it to the pooling conversion unit; S23, the pooling conversion unit calculates the maximum pooling formula "Z(i, j) = max(X[i*P S (i+1)*P S ,j*P S (j+1)*P S ]), Z(i, j) is a pixel value in the output feature map after pooling, i and j are the positions in the output feature map respectively, max is the maximum value of all pixel values ​​in the input window, X is the input feature map, P S The maximum pixel value after the maximum pooling is obtained as the window size of the pooling operation and enters the fully connected layer to retain important feature information and pass it to the fully connected layer unit; S24, the fully connected layer unit is calculated according to the fully connected layer formula "y=f(∑(w n *x n )+b), y is the output result, f is the activation function, w n is the weight of the nth input feature, x n is n input features, n is the dimension of the input features, and b is the bias. The output data of the connection layer is obtained and enters the output layer for subsequent output calculation and is passed to the output conversion unit. S25, the output conversion unit obtains the output value size after convolution according to the output layer conversion calculation formula "N = (P-F + 2C) / S + 1, N is the output size after convolution, P is the input size before convolution, F is the convolution kernel size, C is the number of layers of 0 added around the image, and S is the step size" for subsequent model analysis and training, and passes it to the feature extraction unit; S26, a feature extraction unit extracts key features of road pollution, road conditions, environmental sanitation, road signs, and environmental changes of road cleanliness according to the processed image data, so as to facilitate subsequent modeling training and recognition; Described step S30, comprises the following steps: S31, the training simulation unit trains the machine learning model using the extracted feature information, identifies whether the road monitoring data is abnormal and its abnormal category, and transmits it to the model evaluation unit; S32. The model evaluation unit obtains the comprehensive evaluation score after model training according to the model comprehensive evaluation score calculation formula "F1 = 2*(Pr*Re) / (Pr+Re), F1 is the model comprehensive evaluation score, Pr is the precision rate, and Re is the recall rate" to ensure the accuracy of model verification and transmit it to the processing center; S33, the processing center compares the actual comprehensive evaluation score of the model with the comprehensive evaluation standard score of the model stored in the memory: if it meets the standard, it notifies that monitoring can be carried out; if it does not meet the standard, it is transmitted to the alarm and notifies to adjust the parameters to continue training; S34, the abnormality recognition unit inputs the trained image into the trained convolutional neural network model to match the road cleaning monitoring image obtained in real time and confirm its abnormal category, and transmits it to the cost budget unit; S35, the cost budget unit is based on the daily cleaning cost calculation formula of the sanitation vehicle"J g =L j *C j +L j *Xc+L j / D z *D y *G r , J g is the daily cleaning cost of the sanitation vehicle (yuan / day), L j is the daily cleaning mileage of the sanitation vehicle (km / day), C j is the cleaning material cost of the sanitation vehicle per kilometer (yuan / km), Xc is the cleaning repair cost of the sanitation vehicle per kilometer (yuan / km), Dz is the daily cleaning mileage of the sanitation vehicle (km / day), y is the number of sanitation vehicle operators (persons), Gr is the daily salary of sanitation vehicle personnel (yuan / person / day)” to obtain the daily cleaning cost of the sanitation vehicle for subsequent resource allocation and pass it to the parameter determination unit; S36. The parameter determination unit obtains the resource configuration of each road corresponding to the model and the automatic operation parameters of various sanitation monitoring equipment according to the road cleaning abnormality category, so as to automatically match and apply the subsequent automatic monitoring of the sanitation monitoring equipment.

4. A sanitation operation supervision system as claimed in claim 1, characterized in that: The method further comprises: The step S40 comprises the following steps: S41, the cleaning operation unit performs automatic sanitation cleaning operations according to the trained model and the target cleaning section, matching the type and number of sanitation vehicles, the number of sanitation workers and the cleaning process, and transmits the information to the information collection unit; S42, the information collection unit obtains images of roads in the area through the set high-definition camera and converts them into environmental sanitation quality information of the road surface to ensure that the environmental sanitation quality of the road meets the standards and transmits them to the processing center; S43, the processing center compares the actual sanitation quality information of the road with the sanitation quality standard of the road stored in the memory: if the standard is met, the sanitation work is completed; if the standard is not met, the alarm is transmitted to the alarm and the work is notified to be reworked and re-sweeped; S44, the emergency dispatch unit sends an alarm based on the abnormal feedback and uses the trained model to promptly assign the corresponding sanitation vehicle and equip sanitation workers to take special disposal measures to ensure that the difficult areas are cleaned up and passed on to the training and education unit; S45. The training and education unit improves the professional quality and skills of sanitation workers through online training and distance education according to the requirements of the model, so that sanitation workers can save themselves and each other and work safely in emergency situations.

5. A sanitation operation supervision system as claimed in claim 1, characterized in that: The method further comprises: Described step S50 comprises the following steps: S51, the garbage classification unit uses the trained model to classify and supervise garbage according to environmental protection laws and regulations, and automatically generates inspection, assessment and law enforcement work orders, and passes them to the public participation unit; S52. The public participation unit is connected to the sanitation complaint service center by setting up mini programs, apps, and hotlines to facilitate the public to report sanitation problems so that subsequent sanitation operations can be continuously improved and passed on to the information aggregation unit; S53. The information aggregation unit compares the sanitation maintenance information of each road obtained by the high-definition camera with the complaint information to obtain the severity of the complaint result, and summarizes it into a sanitation score for each road as a basis for subsequent sanitation improvement.

6. A sanitation operation supervision system as described in claims 1-5, characterized in that: It also includes a computer-readable storage medium containing a memory; the memory stores a computer program, and when the functional modules execute the computer program, the steps of a sanitation operation supervision method described in any one of claims 1 to 5 are implemented; the computer-readable storage medium stores a computer program, and when the computer program is executed by the functional modules, the steps of a sanitation operation supervision method described in any one of claims 1 to 5 are implemented.

7. A sanitation operation supervision system as described in claims 1-5, characterized in that: It also includes a sanitation operation supervision and control device, which is implemented by the sanitation operation supervision method described in claims 1-5 above.

Images