Odour monitoring method for a waste station

Abstract

The application discloses a method for monitoring odor of a garbage station, comprising: acquiring a real-time position and a forward direction of a garbage truck in the garbage station; controlling an odor detection module located in the forward direction of the garbage truck in the garbage station to switch to an active state according to the real-time position and the forward direction; and outputting odor monitoring information including at least an odor presence signal when the odor detection module detects odor. The trajectory pre-activation mechanism makes the sensor in the best response state before the garbage truck arrives, fundamentally solves the problem of delay from cold start to stable output of the traditional fixed-point sensor, realizes a second-level response, lays a time foundation and an accuracy foundation for timely monitoring of odor generated by illegal dumping or leakage, and the service life is not greatly reduced since the high activity state is only entered when the garbage truck is about to arrive.

Description

Technical Field

[0001] This invention relates to the field of urban sanitation management technology, and in particular to a method for odor monitoring at garbage stations. Background Technology

[0002] As a key node in the urban sanitation system, garbage stations are responsible for the transfer, sorting, and compression of household waste. However, in actual operation, garbage trucks may engage in illegal dumping (such as dumping in undesignated areas, over-dumping, or mixing in industrial waste). These behaviors are often accompanied by the instantaneous release of high-concentration odors, which seriously affect the surrounding environment and residents' lives.

[0003] While existing video surveillance systems can identify violations, they cannot directly detect odors and have blind spots. Therefore, they can only currently provide an alarm function after detecting violations.

[0004] Traditional odor monitoring systems typically employ fixed-point sensor deployments for continuous monitoring. Currently available odor detection sensors primarily utilize core detection materials based on optical, electrical, and chemical principles. These sensors require activation before operation, maximizing their activity and maintaining them at an optimal measurement state to output high-precision data signals. However, a highly active sensor is prone to irreversible damage. Therefore, higher-precision odor sensors often come with shorter lifespans and higher maintenance requirements. Consequently, odor sensors currently deployed in waste management plants exhibit a fundamental contradiction between response speed, detection accuracy, and lifespan.

[0005] This results in a situation where, when illegal dumping occurs, the odor monitoring system typically needs a response time of ten or even tens of seconds to complete the output of odor data signals. Consequently, the corresponding deodorization equipment simply cannot handle the situation in time. Therefore, current waste stations generally suffer from the problem of not being able to deal with the odors generated by illegal dumping in a timely manner.

[0006] Therefore, this technical solution proposes an odor monitoring method for waste stations that can facilitate timely response to illegal dumping of waste, ensuring a rapid response to odors generated by illegal operations. Summary of the Invention

[0007] The purpose of this invention is to address the shortcomings of existing technologies by proposing an odor monitoring method for garbage stations. Through a trajectory pre-activation mechanism, the sensor is in its optimal response state before the garbage truck arrives, fundamentally solving the problem of delay in the cold start to stable output of traditional fixed-point sensors. This achieves a second-level response, laying a temporal and accuracy foundation for timely monitoring of odors caused by illegal dumping or leakage. At the same time, it only enters a high-activity state when the garbage truck is about to arrive, so its service life is not significantly reduced.

[0008] Firstly, in order to achieve the above objectives, the present invention provides the following technical solution: Odor monitoring methods for waste disposal sites include: Obtain the real-time location and direction of travel of the garbage truck within the garbage station; Based on the real-time location and direction of travel, control the odor detection module located in the direction of the garbage truck's travel within the garbage station to switch to the active state; When the odor detection module detects an odor, it outputs odor monitoring information, including at least whether an odor is present or absent.

[0009] By adopting this implementation method and using a trajectory pre-activation mechanism, the sensor is in the optimal response state before the garbage truck arrives. This fundamentally solves the problem of delay in the transition from cold start to stable output of traditional fixed-point sensors, achieving a second-level response. This lays the time and accuracy foundation for timely monitoring of odors caused by illegal dumping or leakage. At the same time, it only enters a high-activity state when the garbage truck is about to arrive, so its service life will not be significantly reduced.

[0010] In conjunction with the first aspect, in one embodiment, before obtaining the real-time location and direction of travel of the garbage truck within the garbage station, the method further includes: Obtain the internal road information of the garbage station; The station's internal road map is divided into multiple spatial blocks, and each spatial block contains at least one odor detection module.

[0011] By adopting this implementation method, through fine spatial block division, odor monitoring can be accurate to local areas, avoiding the energy waste and sensor damage caused by uniform activation of the entire station.

[0012] In conjunction with the first aspect, in one embodiment, the method further includes acquiring the speed of the garbage truck within the garbage station; controlling the odor detection module located in the direction of the garbage truck's movement within the garbage station to switch to an active state includes: Based on the driving speed, real-time location, and direction of travel, calculate the minimum remaining time for the garbage truck to reach each detection and deodorization space; Obtain the preset activation time required for the odor detection module in each of the aforementioned spatial blocks; By comparing the minimum remaining time with the activation time, it is determined whether the odor detection module in each space block has met the activation conditions. When the odor detection module meets the activation conditions, the odor detection module is controlled to switch to the activated state.

[0013] By employing this implementation method, the matching relationship between the remaining time and the activation time is precisely calculated to ensure that the sensor has just completed preheating when the vehicle arrives. This avoids both the energy waste and lifespan loss caused by premature activation and the response delay caused by late activation.

[0014] In conjunction with the first aspect, in one embodiment, controlling the odor detection module to switch to the active state includes: preheating the detection materials of all odor sensors in the odor detection module to the rated operating state.

[0015] This implementation method ensures that odor monitoring information can be output in a very short time when odor comes into contact with the core detection material of the odor detection module.

[0016] In conjunction with the first aspect, in one embodiment, the step of outputting odor monitoring information when the odor sensor detects an odor includes: The first protective device of the odor detection module in the space block where the garbage truck is located is activated according to the real-time location of the garbage truck. The first protective device is used to cover the first odor sensor used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor, the second protective device for covering the second odor sensor is activated. The second odor sensor is used to detect odor concentration and odor type. Output the odor monitoring information, including whether there is an odor signal, odor concentration, and odor type.

[0017] This implementation method constructs a two-layer monitoring architecture capable of rapid sensing and precise verification. The first sensor acts like a sentinel, performing routine monitoring and capturing any odor within seconds. The second sensor, upon detecting an odor, performs precise concentration and type detection, providing a reliable basis for subsequent deodorization strategies. Through the coordinated control of protective devices, the entire monitoring architecture achieves a balance between rapid response, long lifespan, and high accuracy.

[0018] In conjunction with the first aspect, in one embodiment, after the garbage truck leaves the space block, after outputting the presence or absence of odor signal, odor concentration and odor type, or after reaching the preset activation time, the first protective device and the second protective device are turned off, and the first odor sensor and the odor sensor are switched back to the inactive state.

[0019] This implementation avoids the risk of increased energy consumption and lifespan reduction caused by the odor detection module remaining in standby mode for extended periods while in active state.

[0020] In conjunction with the first aspect, in one embodiment, the step of outputting odor monitoring information when the odor sensor detects an odor includes: Real-time monitoring to ensure that the garbage truck does not dump garbage in the designated location; In response to the monitoring result that the garbage truck did not dump garbage in the designated location, the first protective device of the odor detection module in the space block where the garbage truck is located is activated. The first protective device is used to cover the first odor sensor used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor, a second protective device for covering the second odor sensor is activated. The second odor sensor is used to detect the concentration and type of odor. The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, and odor type.

[0021] This implementation method, through dual verification of video and sensors, achieves accurate identification of illegal dumping behavior, preventing false alarms and covering blind spots.

[0022] In conjunction with the first aspect, in one embodiment, the first odor sensor is a PID sensor, and the second odor sensor includes at least one of an electrochemical sensor and a non-dispersive ultraviolet sensor.

[0023] In conjunction with the first aspect, in one embodiment, both the first protective device and the second protective device include: The outer casing is used to cover each of the first odor sensor and the second odor sensor; An automatic damper is provided at the air inlet and air outlet of the housing, and the automatic damper includes one of a solenoid valve, a gate valve, and a rotating baffle. The filter unit is located at the air inlet and air outlet of the housing; A fan, located inside the housing, is used to introduce air from outside the housing into the housing.

[0024] By adopting this implementation method, the response speed and service life of the odor sensor are greatly improved, making the entire odor detection unit more sensitive and unable to miss any odor, thus further enhancing the performance of this monitoring function.

[0025] In conjunction with the first aspect, in one embodiment, it further includes: In response to the odor detection result or the monitoring result that the garbage truck has not dumped garbage in the designated location, the real-time feature information of the garbage is obtained based on the pre-trained garbage feature information extraction model, including at least one of volume, composition and abnormal object conditions; The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, odor type, and waste characteristic information.

[0026] This approach provides a more comprehensive basis for decision-making in deodorization systems, enabling subsequent deodorization operations to be more precise, efficient, and balanced in resource utilization.

[0027] In conjunction with the first aspect, in one embodiment, the input features of the waste feature extraction model are: a video frame sequence of the waste dumping process after size normalization to a preset resolution and pixel value normalization; the output target of the waste feature extraction model is: at least one of the following: estimated waste volume, composition, and abnormal object conditions corresponding to the video frame sequence. The garbage feature information extraction model is constructed in the following way: Obtain a training dataset, which includes multiple sets of historical garbage dumping video samples. Each set of samples includes at least: input features composed of a sequence of historical garbage dumping video frames, and output labels composed of manually labeled results corresponding to the video frame sequence. The manually labeled results include at least one of the following: estimated garbage volume, composition, and abnormal object conditions. Model training: Using the input features as training input and the output labels as training targets, the deep learning model is supervised and trained using the cross-entropy loss function and the Adam optimizer until the loss function value of the model on the validation set no longer decreases, thus obtaining the pre-trained garbage feature information extraction model; The original feature vector output by the model is mapped into a multi-dimensional vector through a fully connected layer, where: The first dimension is mapped to a garbage volume coefficient by the Sigmoid function, and then multiplied by the preset maximum volume threshold to obtain the estimated volume; The second to Nth dimensions are mapped to the probability distribution of the components using the Softmax function, and the category corresponding to the maximum probability is taken as the component recognition result. The N+1 to Mth dimensions are judged by threshold to generate abnormal object detection results; Model Validation: After training, the model is validated using an independent test dataset, which contains labeled video samples not used in the training process. The model's recognition accuracy on the test set is no less than 90%. The garbage feature information extraction model is a deep learning model based on convolutional neural networks or visual Transformer architecture.

[0028] In conjunction with the first aspect, in one embodiment, the model of the garbage feature information extraction model is a ResNet-50 convolutional neural network or a ViT-B / 16 visual Transformer model, and the model has more than 100 million trainable parameters.

[0029] This implementation method uses a lightweight model to facilitate the rapid collection of garbage volume, composition, and abnormal object conditions during the garbage truck dumping process, thereby improving the accuracy of the data source on which the subsequent deodorization system relies for decision-making and reducing data acquisition latency.

[0030] In conjunction with the first aspect, in one embodiment, determining whether the odor detection module in each spatial block meets the activation condition includes: The deodorant residue type and deodorant residue concentration prediction value of each space block when the garbage truck arrives are obtained. When the deodorant residue concentration prediction value is lower than the first preset threshold, or the deodorant residue type does not damage the core detection material of the odor detection module, the first activation condition is met. By comparing the minimum remaining time with the activation time, if the difference between the minimum remaining time and the activation time does not exceed the preset activation time margin, the second activation condition is met. The odor detection module is switched to the active state only when both the first and second activation conditions are met.

[0031] By adopting this implementation method, the sensor is ensured to be activated only when the environment is safe and there is sufficient time through dual activation condition judgment, which effectively avoids measurement inaccuracies and shortened sensor lifespan caused by deodorant interference.

[0032] In conjunction with the first aspect, in one embodiment, obtaining the predicted concentration of deodorant residue in each space block at the time the garbage truck arrives includes: Obtain the most recent deodorization operation log and / or deodorization operation plan of the deodorization equipment associated with the space block; Combining real-time acquired meteorological parameters such as wind speed, temperature, and humidity, and based on at least one of the preset deodorant residue decay model, deodorant residue decay table, and deodorant residue decay algorithm, the predicted value of deodorant residue concentration when the garbage truck arrives in the space block is calculated.

[0033] This implementation provides a variety of flexible paths for calculating the predicted value of deodorant residual concentration.

[0034] In conjunction with the first aspect, in one embodiment, the regression model adopts an XGBoost or LightGBM gradient boosting tree model, the model contains no less than 100 decision trees, and the maximum depth of each tree does not exceed 6 layers. The input features of the deodorant residue decay model include: one-hot deodorant type coding, cumulative duration after deodorization operation, real-time wind speed, real-time temperature, real-time humidity, and initial concentration at the end of deodorization operation; the output target of the deodorant residue decay model includes: predicted value of deodorant residue concentration. The deodorant residue decay model is constructed in the following way: Obtain a historical training dataset, which includes monitoring data of actual residual concentrations after multiple deodorization operations and corresponding meteorological parameters. The monitoring data is collected by residual concentration sensors deployed in the space block or obtained through manual sampling and analysis; the training dataset contains no fewer than 1,000 records. Model training: Using the input features as training input, the output labels as training targets, mean squared error (MSE) as the loss function, and 5-fold cross-validation for hyperparameter tuning, the gradient boosting tree model is trained under supervision until the loss function value on the validation set converges. Model validation: After training, the model is validated using an independent test dataset. The coefficient of determination R² between the model's predicted values ​​and the actual monitored values ​​in the test set is not less than 0.85, and the mean absolute percentage error (MAPE) does not exceed 20%.

[0035] In conjunction with the first aspect, in one embodiment, the preset deodorant residue decay table is a multidimensional lookup table. The multidimensional lookup table uses at least one of the following as index dimensions: deodorant type, cumulative duration after deodorization operation, wind speed level, temperature level, and humidity level. It pre-stores the corresponding deodorant residue concentration value or residue concentration ratio. The predicted value of the deodorant residual concentration was obtained through the following method: Based on the type of deodorant, the cumulative duration of the garbage truck's journey, and the level matched by the real-time acquired meteorological parameters, the corresponding deodorant residual concentration value or residual concentration ratio is queried in the multidimensional lookup table. Multiply the query result by the initial concentration at the end of the deodorization operation to obtain the predicted value of the deodorant residual concentration.

[0036] This implementation provides a prediction method that requires no complex calculations and has a fast response, making it suitable for edge devices with limited computing resources.

[0037] In conjunction with the first aspect, in one embodiment, the preset deodorant residue decay algorithm is an exponential decay algorithm, whose inputs are: the initial concentration C0 at the end of the deodorization operation, the end time t0 of the deodorization operation, the expected arrival time t of the garbage truck, and the real-time wind speed v, temperature T, and humidity RH; its output is: the predicted value C of the deodorant residue concentration when the garbage truck arrives. The calculation formula for the exponential decay algorithm is as follows: ; The attenuation coefficient k is determined in the following way: ; k0 is the baseline attenuation coefficient, obtained through prior experimental calibration; f(v) is the wind speed correction function. α is the wind speed influence coefficient; g(T) is the temperature correction function. β is the temperature influence coefficient, T0 is the reference temperature; h(RH) is the humidity correction function. γ is the humidity influence coefficient, and RH0 is the reference humidity; The reference attenuation coefficient k0, wind speed influence coefficient α, temperature influence coefficient β, and humidity influence coefficient γ are calibrated separately for each type of deodorant and pre-stored.

[0038] This implementation provides a clear physical meaning and simple calculation of the attenuation model, which is convenient for engineering implementation.

[0039] In conjunction with the first aspect, in one embodiment, it further includes: If the odor detection module in the space block does not meet the first activation condition and / or the second activation condition, output status information indicating that the odor detection module cannot monitor, as well as at least one of the following: deodorant residue type, deodorant residue concentration prediction value, and activation time comparison result in the space block.

[0040] This implementation ensures that even if the odor detection module cannot be activated, the downstream system can still obtain valuable status information and make reasonable decisions, demonstrating the robustness and integrity of the system.

[0041] Secondly, in order to achieve the above objectives, the present invention also provides the following technical solution: Odor monitoring systems for waste disposal sites include: Odor sensor units are deployed in various spatial blocks within the waste station to detect the presence, type, and concentration of odors. The video surveillance unit is used to acquire video footage of the garbage truck driving and dumping. The control unit is connected to the odor sensor unit and the video monitoring unit respectively, and is used to execute the method described in the first aspect.

[0042] Thirdly, in order to achieve the above objectives, the present invention also provides the following technical solution: A computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the method described in the first aspect.

[0043] Fourthly, in order to achieve the above objectives, the present invention also provides the following technical solutions: The waste station includes the odor monitoring system described in the second aspect.

[0044] Compared with existing technologies, the beneficial effects of this invention are as follows: through the trajectory pre-activation mechanism, the sensor is in the optimal response state before the garbage truck arrives, which fundamentally solves the problem of delay in the transition from cold start to stable output of traditional fixed-point sensors, achieving second-level response. This lays the time and accuracy foundation for timely monitoring of odors caused by illegal dumping or leakage. At the same time, it only enters a high-activity state when the garbage truck is about to arrive, so its service life will not be significantly reduced. Attached Figure Description

[0045] Figure 1 This is a flowchart of the odor monitoring method for waste stations proposed in this invention; Figure 2 This is a schematic diagram of the protective device in the odor monitoring method for garbage stations proposed in this invention.

[0046] In the diagram: 1. First odor sensor; 2. Second odor sensor; 3. First protective device; 4. Second protective device; 31. Housing; 32. Automatic damper; 33. Filter unit; 34. Fan. Detailed Implementation

[0047] In the daily operation of waste collection stations, illegal dumping by garbage trucks (such as dumping in undesignated areas, over-dumping, and mixing in industrial waste) is the main cause of the instantaneous release of high-concentration odors. While existing video surveillance systems can identify violations, they cannot directly detect odors and have blind spots.

[0048] Traditional fixed-point odor sensors currently on the market typically rely on core detection materials based on optical, electrical, and chemical principles. These core detection materials require a "highly active" state to respond quickly. Operation requires activating the core detection material to enhance its activity and maintain it at an optimal measurement state before outputting high-precision detection data signals. However, high activity also means a higher risk of irreversible damage. Therefore, higher-precision odor sensors usually also mean shorter lifespans and higher maintenance requirements. Consequently, existing odor sensors generally suffer from a trade-off between response speed, detection accuracy, and lifespan. This results in odor monitoring systems typically taking ten seconds or even tens of seconds to output valid data when illegal dumping occurs, leaving corresponding deodorization equipment unable to process the data in time.

[0049] It should be understood that the following embodiments are all intended to solve the problem that existing odor monitoring technologies in waste stations cannot simultaneously achieve response speed, detection accuracy and service life, so as to provide low-latency and high-precision monitoring methods for the operation of subsequent waste station deodorization systems, and to ensure that the service life of the detection methods is not significantly reduced.

[0050] In the following embodiments, "garbage truck" should be understood as any type of garbage truck that enters the garbage station for dumping operations. "Odor detection module" refers to a device deployed in various spatial blocks of the garbage station capable of performing odor detection, including but not limited to a fast-response first odor sensor, a high-precision second odor sensor, and corresponding protective devices. "Activation state" refers to the moment when the sensors in the odor detection module, through preheating or other methods, bring their core detection materials to their rated operating state, enabling them to respond to odor signals at the fastest possible speed.

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Example 1: Please see Figure 1 The present invention provides the following technical solution: a method for odor monitoring in waste stations, comprising: Obtain the real-time location and direction of travel of the garbage truck within the garbage station; Based on the real-time location and direction of travel, control the odor detection module located in the direction of the garbage truck's travel within the garbage station to switch to the active state; When the odor detection module detects an odor, it outputs odor monitoring information, including at least whether an odor is present or absent.

[0053] As an optional implementation of this invention, when a garbage truck enters a garbage station, multi-camera systems, electronic tags, or UWB positioning base stations deployed within the station can acquire the real-time location and direction of travel of the garbage truck. Simultaneously, odor detection modules located in the direction of travel of the garbage truck are activated. The core detection materials of these odor detection modules have high activity due to factors such as increased temperature. When the garbage truck passes by, regardless of whether the garbage truck illegally dumps garbage or leaks, once the escaped odor spreads to the designed location of the odor detection module, it can immediately react with the core detection material of the odor detection module, thereby outputting a detection signal. At this time, the deodorization system deployed in the garbage station can immediately respond and start the deodorization operation, so that the odor caused by illegal dumping or leakage of garbage trucks can be eliminated in a short period of time.

[0054] Throughout the monitoring process, the odor detection module is activated before the garbage truck even reaches its location, using the truck's trajectory as the trigger for sensor pre-activation. Compared to traditional fixed-point sensors that require continuous high activity for rapid response, trajectory prediction ensures the odor detection module only enters a high-activity state when the garbage truck is about to arrive, guaranteeing both response speed and accuracy while significantly extending sensor lifespan.

[0055] Of course, the odor detection module described in this embodiment can be a single odor sensor or multiple types of odor sensors, such as PID sensors, metal oxide semiconductor sensors, non-dispersive sensors, etc., so as to cope with the detection of different odor sources in different functional waste stations. This embodiment does not limit this, but the odor detection module should at least include an odor sensor that can detect the presence or absence of odor, such as a PID sensor, otherwise the deodorization system will not be able to receive the activation command.

[0056] The term "activated state" specifically refers to the process of preheating or similar methods to bring the core detection materials of all odor sensors in the odor detection module to their rated operating state. For example, for the first odor sensor 1, which uses the PID principle, its core detection materials (such as the ultraviolet lamp and ionization chamber) need to be preheated to a stable operating temperature, and the response time can be shortened from tens of seconds during cold start to less than 2 seconds.

[0057] For example, when a garbage truck enters through the entrance, the system identifies its location at the entrance and its direction of travel towards the unloading area. The system immediately activates the odor detection module in the space containing the unloading area. When the garbage truck arrives at the unloading area, the sensor is already in optimal working condition. If illegal dumping occurs at this time, the sensor can output a signal indicating the presence or absence of odor within 1-2 seconds, while traditional solutions require more than 10 seconds.

[0058] By adopting this implementation method and using a trajectory pre-activation mechanism, the sensor is in the optimal response state before the garbage truck arrives. This fundamentally solves the problem of delay in the transition from cold start to stable output of traditional fixed-point sensors, achieving a second-level response. This lays the time and accuracy foundation for timely monitoring of odors caused by illegal dumping or leakage. At the same time, it only enters a high-activity state when the garbage truck is about to arrive, so its service life will not be significantly reduced.

[0059] Furthermore, in some possible embodiments, after the garbage truck leaves, once the odor detection module has output a signal indicating the presence or absence of odor, or after a preset activation time has been reached, it switches back to an inactive state. This avoids excessive energy consumption due to prolonged operation, or a significant reduction in lifespan due to prolonged reaction with odor molecules. Of course, existing odor sensors generally activate intermittently and automatically shut down after detecting odor or reaching a preset maximum activation time. Therefore, this embodiment does not completely contradict the monitoring methods of existing odor sensors during normal operation. Rather, this embodiment can be seen as simply adding a function that can be activated via the garbage truck's trajectory to the existing monitoring method.

[0060] Example 2: Please see Figure 1 The present invention also provides the following technical solution: a method for odor monitoring in waste stations, comprising: Obtain the internal road information of the garbage station; The station's internal road map is divided into multiple spatial blocks, and each spatial block contains at least one odor detection module. Obtain the real-time location, speed, and direction of travel of the garbage truck within the garbage station; Based on the driving speed, real-time location, and direction of travel, calculate the minimum remaining time for the garbage truck to reach each detection and deodorization space; Obtain the preset activation time required for the odor detection module in each of the aforementioned spatial blocks; By comparing the minimum remaining time with the activation time, it is determined whether the odor detection module in each space block has met the activation conditions. When the odor detection module meets the activation conditions, control the odor detection module to switch to the activated state; When the odor detection module detects an odor, it outputs odor monitoring information, including at least whether an odor is present or absent.

[0061] As an optional implementation of this invention, the garbage station is divided into finely divided spatial blocks to achieve on-demand activation rather than full-station activation, and the activation timing is precisely calculated to avoid activation too early or too late. If activation is too early, the sensor remains in a highly active state for an extended period, which, while not affecting detection, wastes its lifespan; if activation is too late, the sensor has not yet reached its optimal state before the garbage truck arrives, making second-level response impossible. This solution finds the optimal activation time by comparing the "remaining time for garbage truck arrival" with the "time required for sensor activation."

[0062] In traditional solutions, if all sensors at the entire station remain highly active, energy consumption and sensor wear will increase exponentially; if sensors are only deployed at fixed locations, it is impossible to cover all areas that garbage trucks may pass through. This solution, through spatial block division, allows the system to precisely control the activation of sensors in the "spatial block that the garbage truck is about to arrive at," while sensors in other areas remain in a low-power state.

[0063] Specifically, the system first obtains a road map of the waste station and then divides it into multiple independent spatial blocks based on road orientation and the distribution of work areas (such as unloading ports, compaction stations, waste pits, and waste truck access routes). Each spatial block is typically 5m in size. 5m to 10m 10m, ensuring coverage by a single deodorization device. At least one odor detection module is deployed in each space block, which includes a fast-response first odor sensor 1 (for detecting the presence or absence of odor) and a high-precision second odor sensor 2 (for detecting concentration and type), as well as corresponding protective devices.

[0064] For example, a garbage station divides the unloading area into space block A01, the compression station area into space block B03, and the garbage truck passage into space blocks C01, C02, C03, etc., every 10 meters. When a garbage truck is in space block C01 and is traveling towards space block C02 at a speed of 5 km / h, the system predicts that it will enter C02 in 10 seconds based on its direction of travel. Therefore, only the odor detection module in C02 is activated. Since the activation of the odor detection module in C02 requires 8 seconds, an activation command is sent 2 seconds before the garbage truck arrives (i.e., 10-8=2 seconds) to ensure that the sensor has just finished warming up when the garbage truck arrives. The sensors in C01 and other space blocks remain off or in a low-power state. If the garbage truck's speed suddenly changes (e.g., accelerates to 10 km / h), the system will recalculate and dynamically adjust the activation time. After the garbage truck enters C02, the system will then activate C03, and so on, forming a "mobile" activation zone.

[0065] By adopting this implementation method, through refined spatial block division and accurate prediction of garbage truck driving status, odor monitoring can be accurate to local areas, avoiding energy waste and sensor damage caused by uniform activation of the entire station.

[0066] Furthermore, in some possible embodiments, controlling the odor detection module to switch to the active state includes: preheating the detection materials of all odor sensors in the odor detection module to the rated operating state.

[0067] As an optional implementation of the present invention, considering that the types of odors in garbage stations are generally quite complex, including various odorous gases such as ammonia and hydrogen sulfide, and that the type of odor produced by the garbage carried in the garbage truck is unknown most of the time, in order to ensure that the deodorization system has a more comprehensive deodorization effect, all odor sensors need to be switched to the active state when activating the odor detection module to ensure that the odor monitoring information output to the deodorization system is more comprehensive and accurate.

[0068] Of course, in reality, some garbage stations may only produce a small number of types of odors. For example, if the garbage trucks that come are responsible for transferring production waste from a single industrial park, then it is possible to choose to turn on only a small number of targeted odor sensors. This is not limited in Example 1.

[0069] This implementation method ensures that the output odor monitoring information is comprehensive and reliable.

[0070] Furthermore, in some possible embodiments, such as Figure 2 As shown, the odor monitoring information output when the odor sensor detects an odor includes: The first protective device 3 of the odor detection module in the space block where the garbage truck is located is activated according to the real-time location of the garbage truck. The first protective device 3 is used to cover the first odor sensor 1 used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor 1, the second protective device 4 for covering the second odor sensor 2 is activated. The second odor sensor 2 is used to detect the odor concentration and odor type. Output the odor monitoring information, including whether there is an odor signal, odor concentration, and odor type; After the garbage truck leaves the space block, it outputs the presence or absence of odor signal, odor concentration and odor type, or reaches the preset opening time, then shuts down the first protective device 3 and the second protective device 4, and switches the first odor sensor 1 and the odor sensor back to the inactive state.

[0071] As an optional implementation of this invention, the coordinated control of a dual-layer sensor and a protective device balances response speed, detection accuracy, and sensor lifespan. Currently, the main sensors suitable for odor detection on the market are PID sensors (such as the PID-A1 sensor from Alphasense in the UK, with a T90 response time of less than 2 seconds, responding to most volatile organic compounds (VOCs) and having low power consumption). While they cannot accurately distinguish the type of odor or precisely detect its concentration, they offer fast response and relatively long lifespans. The second odor sensor (such as the 4H2S from CITY Technology for measuring hydrogen sulfide; or the Model T100 from Teledyne API in the US, a non-dispersive ultraviolet sensor for measuring ammonia) offers high accuracy but cannot be used for extended periods, otherwise its lifespan will be significantly reduced.

[0072] The first protective device 3 activates the first odor sensor 1 when the garbage truck is about to arrive, keeping it in an open state. The second odor sensor 2 is only exposed and enters the detection phase when the first odor sensor 1 detects an odor. This "sentinel + expert" division of labor means that the high-precision second odor sensor 2 does not need to be continuously exposed, greatly extending its lifespan.

[0073] For example, if a garbage truck illegally dumps waste during unloading, the first odor sensor 1 detects the odor within one second, immediately triggering the second odor sensor 2. The second odor sensor 2 outputs "hydrogen sulfide concentration 50 ppm" after three seconds, then immediately outputs odor monitoring information. Once the vehicle has left, the system deactivates the second protective device 4, and the first protective device 3 either deactivates or resumes its intermittent operation mode.

[0074] This implementation method constructs a two-layer monitoring architecture capable of rapid sensing and precise verification. The first sensor acts like a sentinel, performing routine monitoring and capturing any odor within seconds. The second sensor, upon detecting an odor, performs precise concentration and type detection, providing a reliable basis for subsequent deodorization strategies. Through the coordinated control of protective devices, the entire monitoring architecture achieves a balance between rapid response, long lifespan, and high accuracy.

[0075] Furthermore, in some possible embodiments, such as Figure 2 As shown, the odor monitoring information output when the odor sensor detects an odor includes: Real-time monitoring to ensure that the garbage truck does not dump garbage in the designated location; In response to the monitoring result that the garbage truck did not dump garbage in the designated location, the first protective device 3 of the odor detection module in the space block where the garbage truck is located is activated. The first protective device 3 is used to cover the first odor sensor 1 used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor 1, the second protective device 4 for covering the second odor sensor 2 is activated. The second odor sensor 2 is used to detect the concentration and type of odor. The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, and odor type.

[0076] As an optional implementation of this invention, accurate identification of illegal dumping is achieved through dual verification using video and sensors. Considering that relying solely on video may result in blind spots or false alarms, and relying solely on an odor sensor can only detect the presence of odor but cannot determine whether it is illegal dumping, this technical solution combines the two: the video first identifies the illegal action, then initiates sensor verification; after the sensor confirms the presence of an odor, it outputs a violation signal. This closed loop of "visual detection → olfactory confirmation" not only prevents invalid activation due to video misjudgment but also covers odor leakage in video blind spots.

[0077] Specifically, cameras deployed at key locations throughout the waste station capture real-time video streams. Using a visual big data model-based behavior recognition algorithm (such as VideoMAE or TimeSformer), the system continuously analyzes whether waste trucks are dumping outside designated legal dumping areas. When illegal dumping is detected (e.g., a vehicle lifting its compartment outside the unloading area), the system immediately triggers the odor detection module for that area. The first protective device 3 activates (if not already activated), and the first odor sensor 1 begins sampling. If the first sensor detects an odor, the second odor sensor 2 is activated for precise measurement, following the aforementioned procedure. The final output monitoring information includes, in addition to the presence, concentration, and type of odor, an "illegal dumping signal" and corresponding video evidence.

[0078] This approach, employing both video and sensor verification, enables accurate identification of illegal dumping, preventing false alarms, covering blind spots, and providing a complete chain of evidence for subsequent supervision.

[0079] Furthermore, in some possible embodiments, both the first protective device 3 and the second protective device 4 include: The outer casing 31 is used to cover each of the first odor sensor 1 and the second odor sensor 2. An automatic damper 32 is provided at the air inlet and air outlet of the housing 31. The automatic damper 32 includes one of a solenoid valve, a gate valve, and a rotating baffle. The filter unit 33 is located at the air inlet and air outlet of the housing 31; A fan 34 is disposed inside the housing 31 and is used to introduce air from outside the housing 31 into the housing 31.

[0080] As an optional implementation of the present invention, considering that the odor detection module needs to adapt to various odor diffusion conditions, and that the protective device itself will affect the diffusion direction of the odor, the response speed of the two odor sensors will be affected to a certain extent. In particular, the first odor sensor 1, which is used to detect whether there is an odor signal, needs to act as a "sentinel". Therefore, this embodiment adds a fan 34 to make the odor sensor more sensitive to odor. At the same time, a filter unit 33 (such as a dustproof net + desiccant) is added to ensure that moisture and dust in the air of the garbage station will not enter the odor sensor in a highly active state, so that the odor sensor has a more reliable service life. Of course, the existing Chinese utility model patent with announcement number CN219456104U discloses an electronic nose protection device for odor in waste treatment plants, whose structure and principle are similar to this protection device, so this embodiment will not elaborate further.

[0081] By adopting this implementation method, the response speed and service life of the odor sensor are greatly improved, making the entire odor detection unit more sensitive and unable to miss any odor, thus further enhancing the performance of this monitoring function.

[0082] Furthermore, in some possible embodiments, it also includes: In response to the odor detection result or the monitoring result that the garbage truck has not dumped garbage in the designated location, the real-time feature information of the garbage is obtained based on the pre-trained garbage feature information extraction model, including at least one of volume, composition and abnormal object conditions; The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, odor type, and waste characteristic information.

[0083] As an optional implementation of this invention, visually recognized waste features are fused with odor sensor data to provide a more comprehensive basis for decision-making in the deodorization system. Simple odor data can only tell the deodorization system "there is an odor, and at what concentration," but it cannot tell the system "what kind of waste is producing the odor." Different types of waste have different odor components, volatility, and require different types and intensities of deodorizing agents. This solution identifies waste features (volume, composition, abnormal objects) using a visual model and fuses them with odor data for output, enabling the deodorization system to "treat the root cause."

[0084] Specifically, when the system detects an odor or identifies illegal dumping, it further invokes a pre-trained waste feature extraction model to perform visual analysis of the waste during the dumping process. The model input is a sequence of dumping video frames, and the output includes the estimated volume (cubic meters) of the waste, its composition (e.g., kitchen waste, construction waste, plastic products, mixed waste, etc.), and any abnormal objects (e.g., industrial waste, hazardous materials, bulky waste). This information, along with the odor data, is then output to the deodorization system for decision-making. For example: kitchen waste + high concentration of hydrogen sulfide → select alkaline deodorant + high-intensity spray; construction waste + low concentration of ammonia → select low-intensity spray to avoid resource waste.

[0085] This approach provides a more comprehensive basis for decision-making in deodorization systems, enabling subsequent deodorization operations to be more precise, efficient, and balanced in resource utilization.

[0086] Furthermore, in some possible embodiments, the input features of the waste feature extraction model are: a video frame sequence of the waste dumping process after size normalization to a preset resolution and pixel value normalization; the output target of the waste feature extraction model is: at least one of the following: estimated waste volume, composition, and abnormal object conditions corresponding to the video frame sequence. The garbage feature information extraction model is constructed in the following way: Obtain a training dataset, which includes multiple sets of historical garbage dumping video samples. Each set of samples includes at least: input features composed of a sequence of historical garbage dumping video frames, and output labels composed of manually labeled results corresponding to the video frame sequence. The manually labeled results include at least one of the following: estimated garbage volume, composition, and abnormal object conditions. Model training: Using the input features as training input and the output labels as training targets, the deep learning model is supervised and trained using the cross-entropy loss function and the Adam optimizer until the loss function value of the model on the validation set no longer decreases, thus obtaining the pre-trained garbage feature information extraction model; The original feature vector output by the model is mapped into a multi-dimensional vector through a fully connected layer, where: The first dimension is mapped to a garbage volume coefficient by the Sigmoid function, and then multiplied by the preset maximum volume threshold to obtain the estimated volume; The second to Nth dimensions are mapped to the probability distribution of the components using the Softmax function, and the category corresponding to the maximum probability is taken as the component recognition result. The N+1 to Mth dimensions are judged by threshold to generate abnormal object detection results; Model Validation: After training, the model is validated using an independent test dataset, which contains labeled video samples not used in the training process. The model's recognition accuracy on the test set is no less than 90%. The model for extracting garbage feature information is a ResNet-50 convolutional neural network or a ViT-B / 16 visual Transformer model, and the model has more than 100 million trainable parameters.

[0087] As an optional implementation of the present invention, when extracting garbage feature information, key frames (such as the start of dumping, the middle of dumping, and the end of dumping) are extracted from the video stream, the size of each frame image is normalized, and the pixel values ​​are normalized so that the input values ​​are within a preset range (such as 0-1). Then, the garbage feature information extraction model is input, and after the original feature vector is output, it is mapped through a fully connected layer to obtain the estimated volume, composition and abnormal object detection results.

[0088] Furthermore, in some possible embodiments, the model of the garbage feature information extraction model is a ResNet-50 convolutional neural network or a ViT-B / 16 visual Transformer model, and the model has more than 100 million trainable parameters to ensure that the model has sufficient expressive power.

[0089] This implementation method uses a lightweight model to facilitate the rapid collection of garbage volume, composition, and abnormal object conditions during the garbage truck dumping process, thereby improving the accuracy of the data source on which the subsequent deodorization system relies for decision-making and reducing data acquisition latency.

[0090] Furthermore, in some possible embodiments, determining whether the odor detection module in each space block meets the activation condition includes: The deodorant residue type and deodorant residue concentration prediction value of each space block when the garbage truck arrives are obtained. When the deodorant residue concentration prediction value is lower than the first preset threshold, or the deodorant residue type does not damage the core detection material of the odor detection module, the first activation condition is met. By comparing the minimum remaining time with the activation time, if the difference between the minimum remaining time and the activation time does not exceed the preset activation time margin, the second activation condition is met. The odor detection module is switched to the active state only when both the first and second activation conditions are met.

[0091] As an optional implementation of this invention, deodorant residue interference is introduced as a precondition for sensor activation. In actual operation of a waste station, deodorization may be underway or recently completed within the space, and deodorant residue can contaminate the sensor surface, leading to inaccurate measurements or even sensor damage. Forcibly activating the sensor at this time would not only result in unreliable data but also accelerate sensor aging. This solution predicts the concentration and type of deodorant residue, activating the sensor only when the environment is safe and sufficient time is available.

[0092] Specifically, the system uses a predictive model to calculate the residual concentration of deodorant in the space when the garbage truck arrives. If the residual concentration is below a preset threshold (e.g., 0.1 mg / m³), the system will detect the deodorant residue. 3 If the residual deodorant does not damage the core material of the sensor (e.g., ion mist cannons, bio-enzymes, etc., are harmless to PID sensors), then the environment in which the odor detection module operates is safe and odor information can be collected. Simultaneously, the system calculates the difference between the remaining time until the vehicle arrives and the time required for sensor activation. If this difference does not exceed the preset activation time margin (e.g., 2 seconds), then the remaining activation time is sufficient. Activating the odor detection module is only meaningful when both conditions are met simultaneously.

[0093] For example, an alkaline spray deodorization was performed on a certain space 10 minutes ago, and the predicted residual concentration is 0.5 mg / m³. 3 If the concentration is above the threshold of 0.1, and the alkaline deodorizer corrodes the PID sensor, then the first condition is not met, and the sensor will not be activated even if there is sufficient time to avoid damage. For example, if a space block underwent bio-enzyme deodorization 30 minutes ago, and the predicted residual concentration is 0.05 mg / m³, the sensor will not be activated. 3 If the temperature is below the threshold and the bio-enzyme is harmless to the sensor, then the first condition is met; if the vehicle is about to arrive at this time and the time condition is also met, then it will be activated normally.

[0094] By adopting this implementation method, the sensor is ensured to be activated only when the environment is safe and there is sufficient time through dual activation condition judgment, which effectively avoids measurement inaccuracies and shortened sensor lifespan caused by deodorant interference.

[0095] Furthermore, in some possible embodiments, obtaining the predicted concentration of deodorant residue in each space block at the time the garbage truck arrives includes: Obtain the most recent deodorization operation log and / or deodorization operation plan of the deodorization equipment associated with the space block; Combining real-time acquired meteorological parameters such as wind speed, temperature, and humidity, and based on at least one of the preset deodorant residue decay model, deodorant residue decay table, and deodorant residue decay algorithm, the predicted value of deodorant residue concentration when the garbage truck arrives in the space block is calculated.

[0096] As an optional implementation of this invention, the method for predicting the residual concentration of deodorizer is specified. First, the log of the most recent deodorization operation (including deodorizer type, initial concentration, and operation end time) and / or the schedule of future planned deodorization operations are obtained from the deodorization system. Then, combined with real-time meteorological parameters (wind speed, temperature, humidity), the residual concentration can be calculated using various methods.

[0097] This implementation provides a variety of flexible paths for calculating the predicted value of deodorant residual concentration.

[0098] Furthermore, in some possible embodiments, the regression model adopts an XGBoost or LightGBM gradient boosting tree model, and the model contains no less than 100 decision trees, with each tree having a maximum depth of no more than 6 layers. The input features of the deodorant residue decay model include: one-hot deodorant type coding, cumulative duration after deodorization operation, real-time wind speed, real-time temperature, real-time humidity, and initial concentration at the end of deodorization operation; the output target of the deodorant residue decay model includes: predicted value of deodorant residue concentration. The deodorant residue decay model is constructed in the following way: Obtain a historical training dataset, which includes monitoring data of actual residual concentrations after multiple deodorization operations and corresponding meteorological parameters. The monitoring data is collected by residual concentration sensors deployed in the space block or obtained through manual sampling and analysis; the training dataset contains no fewer than 1,000 records. Model training: Using the input features as training input, the output labels as training targets, mean squared error (MSE) as the loss function, and 5-fold cross-validation for hyperparameter tuning, the gradient boosting tree model is trained under supervision until the loss function value on the validation set converges. Model Validation: After training, the model is validated using an independent test dataset. The coefficient of determination R between the model's predicted values ​​and the actual monitored values ​​in the test set is used. 2 Not less than 0.85, and the mean absolute percentage error (MAPE) not exceeding 20%.

[0099] Furthermore, in some possible embodiments, the preset deodorant residue decay table is a multidimensional lookup table. The multidimensional lookup table uses at least one of the following as index dimensions: deodorant type, cumulative duration after deodorization operation, wind speed level, temperature level, and humidity level, and pre-stores the corresponding deodorant residue concentration value or residue concentration ratio. The predicted value of the deodorant residual concentration was obtained through the following method: Based on the type of deodorant, the cumulative duration of the garbage truck's journey, and the level matched by the real-time acquired meteorological parameters, the corresponding deodorant residual concentration value or residual concentration ratio is queried in the multidimensional lookup table. Multiply the query result by the initial concentration at the end of the deodorization operation to obtain the predicted value of the deodorant residual concentration.

[0100] As an optional implementation of this invention, for example, for alkaline spray deodorizers, the residual concentration ratio under different conditions is calibrated experimentally and stored in a multidimensional lookup table. During a query, if the cumulative duration is 15 minutes, the wind speed is 1.2 m / s (matching the 1 m / s setting), and the temperature is 22℃ (matching the 20℃ setting), then the corresponding ratio of 0.3 is retrieved, multiplied by the initial concentration of 1 mg / m³. 3 The predicted value was 0.3 mg / m³. 3 If the parameter falls between two gears, linear interpolation is used.

[0101] This implementation provides a prediction method that requires no complex calculations and has a fast response, making it suitable for edge devices with limited computing resources.

[0102] Furthermore, in some possible embodiments, the preset deodorant residue decay algorithm is an exponential decay algorithm, whose inputs are: the initial concentration C0 at the end of the deodorization operation, the end time t0 of the deodorization operation, the expected arrival time t of the garbage truck, and the real-time wind speed v, temperature T, and humidity RH; its output is: the predicted value C of the deodorant residue concentration when the garbage truck arrives. The calculation formula for the exponential decay algorithm is as follows: ; The attenuation coefficient k is determined in the following way: ; k0 is the baseline attenuation coefficient, obtained through prior experimental calibration; f(v) is the wind speed correction function. α is the wind speed influence coefficient; g(T) is the temperature correction function. β is the temperature influence coefficient, T0 is the reference temperature; h(RH) is the humidity correction function. γ is the humidity influence coefficient, and RH0 is the reference humidity; The reference attenuation coefficient k0, wind speed influence coefficient α, temperature influence coefficient β, and humidity influence coefficient γ are calibrated separately for each type of deodorant and pre-stored.

[0103] As an optional implementation of the present invention, for example, a certain alkaline deodorant has k0 = 0.1 min. -1Given α=0.05, β=0.02, and γ=0.01, under the conditions of v=2m / s, T=30℃, and RH=70%, k=0.1×(1+0.05×2)×(1+0.02×10)×(1+0.01×20)=0.1×1.1×1.2×1.2=0.1584. If t-t0=10 minutes, C0=1mg / m 3 Therefore, C = 1 × e^{-1.584} = 0.205 mg / m³ 3 .

[0104] This implementation provides a clear physical meaning and simple calculation of the attenuation model, which is convenient for engineering implementation.

[0105] Furthermore, in some possible embodiments, it also includes: If the odor detection module in the space block does not meet the first activation condition and / or the second activation condition, output status information indicating that the odor detection module cannot monitor, as well as at least one of the following: deodorant residue type, deodorant residue concentration prediction value, and activation time comparison result in the space block.

[0106] As an optional implementation of the present invention, when the sensor cannot be activated due to interference from the deodorant or insufficient time, it is also necessary to inform the deodorization system of the reason why it cannot obtain odor monitoring information. Otherwise, the deodorization system will not be able to respond when the garbage truck leaks. Moreover, this phenomenon is not uncommon in the daily deodorization operation of garbage stations.

[0107] While the comparison results of deodorant residue type, predicted deodorant residue concentration, and activation time within the space block are not as accurate as direct odor monitoring information, they can still provide information for the deodorization system to make decisions. For example, if the residue concentration is high but the risk of leakage is low, the deodorization system can maintain the status quo; if the residue concentration is high and the risk of leakage is high, the deodorization system can increase the deodorization intensity in advance; if there is insufficient time but the risk of leakage is high, the deodorization system can take preventive short-term spraying measures.

[0108] Of course, although the above methods may result in some waste of deodorizing agents, they can at least ensure that the operation of the garbage station will not have a significant impact on the quality of the surrounding environment, making them particularly suitable for use in urban garbage stations.

[0109] This implementation ensures that even if the odor detection module cannot be activated, the downstream system can still obtain valuable status information and make reasonable decisions, demonstrating the robustness and integrity of the system.

[0110] Example 3: The present invention also provides the following technical solution: an odor monitoring system for waste stations, comprising: Odor sensor units are deployed in various spatial blocks within the waste station to detect the presence, type, and concentration of odors. The video surveillance unit is used to acquire video footage of the garbage truck driving and dumping. The control unit is connected to the odor sensor unit and the video monitoring unit respectively, and is used to execute the method described in Embodiment 1 or 2.

[0111] As an optional implementation of the present invention, the various units of the system work collaboratively to execute the methods of the aforementioned embodiments, thereby achieving efficient monitoring of odors at the waste station. Odor sensor units are deployed in each spatial block, video monitoring units are deployed at key locations, and the control unit connects to each unit via wired or wireless networks to receive data in real time and perform operations such as activation judgment and data fusion.

[0112] By adopting this implementation method, the monitoring method of Example 1 or 2 can be transformed into a practically applicable monitoring system through systematic hardware deployment.

[0113] Example 4: The present invention also provides the following technical solution: a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the method described in embodiment 1 or 2.

[0114] Example 5: The present invention also provides the following technical solution: a garbage station, including the odor monitoring system described in Example 3.

[0115] The working principle and usage process of this invention are as follows: When a garbage truck enters a garbage station, multi-camera systems, electronic tags, or UWB positioning base stations deployed within the station can acquire the real-time location and direction of travel of the garbage truck. Simultaneously, odor detection modules located in the direction of travel of the garbage truck are activated. Due to factors such as increased temperature, the core detection materials of these odor detection modules have high activity. When the garbage truck passes by, regardless of whether it illegally dumps garbage or leaks, once the escaped odor spreads to the designed location of the odor detection module, it can immediately react with the core detection material of the odor detection module, thereby outputting a detection signal. At this time, the deodorization system deployed in the garbage station can immediately respond and start the deodorization operation, so that the odor caused by illegal dumping or leakage of garbage trucks can be eliminated in a short period of time.

[0116] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An odor monitoring method for waste stations, wherein, include: Obtain the real-time location and direction of travel of the garbage truck within the garbage station; Based on the real-time location and direction of travel, control the odor detection module located in the direction of the garbage truck's travel within the garbage station to switch to the active state; When the odor detection module detects an odor, it outputs odor monitoring information, including at least whether an odor is present or absent.

2. The method according to claim 1, wherein, Before obtaining the real-time location and direction of travel of the garbage truck within the garbage station, the method further includes: Obtain the internal road information of the garbage station; The station's internal road map is divided into multiple spatial blocks, and each spatial block contains at least one odor detection module.

3. The method according to claim 2, wherein, It also includes obtaining the speed at which the garbage truck travels within the garbage station; The process of switching the odor detection module located in the direction of the garbage truck's movement within the garbage station to an active state includes: Based on the driving speed, real-time location, and direction of travel, calculate the minimum remaining time for the garbage truck to reach each detection and deodorization space; Obtain the preset activation time required for the odor detection module in each of the aforementioned spatial blocks; By comparing the minimum remaining time with the activation time, it is determined whether the odor detection module in each space block has met the activation conditions. When the odor detection module meets the activation conditions, the odor detection module is controlled to switch to the activated state.

4. The method according to claim 3, wherein, The process of switching the odor detection module to the active state includes: preheating the detection materials of all odor sensors in the odor detection module to the rated working state.

5. The method according to claim 2, wherein, The odor monitoring information output when the odor sensor detects an odor includes: The first protective device of the odor detection module in the space block where the garbage truck is located is activated according to the real-time location of the garbage truck. The first protective device is used to cover the first odor sensor used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor, the second protective device for covering the second odor sensor is activated. The second odor sensor is used to detect odor concentration and odor type. Output the odor monitoring information, including whether there is an odor signal, odor concentration, and odor type.

6. The method according to claim 3 or 5, wherein, The odor monitoring information output when the odor sensor detects an odor includes: Real-time monitoring to ensure that the garbage truck does not dump garbage in the designated location; In response to the monitoring result that the garbage truck did not dump garbage in the designated location, the first protective device of the odor detection module in the space block where the garbage truck is located is activated. The first protective device is used to cover the first odor sensor used to detect the presence or absence of odor. In response to the odor detection result of the first odor sensor, a second protective device for covering the second odor sensor is activated. The second odor sensor is used to detect the concentration and type of odor. The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, and odor type.

7. The method according to claim 6, wherein, Also includes: In response to the odor detection result or the monitoring result that the garbage truck has not dumped garbage in the designated location, the real-time feature information of the garbage is obtained based on the pre-trained garbage feature information extraction model, including at least one of volume, composition and abnormal object conditions; The output odor monitoring information includes illegal dumping signals, odor presence or absence signals, odor concentration, odor type, and waste characteristic information.

8. The method according to claim 7, wherein, The determination of whether the odor detection module in each spatial block meets the activation condition includes: The deodorant residue type and deodorant residue concentration prediction value of each space block when the garbage truck arrives are obtained. When the deodorant residue concentration prediction value is lower than the first preset threshold, or the deodorant residue type does not damage the core detection material of the odor detection module, the first activation condition is met. By comparing the minimum remaining time with the activation time, if the difference between the minimum remaining time and the activation time does not exceed the preset activation time margin, the second activation condition is met. The odor detection module is switched to the active state only when both the first and second activation conditions are met.

9. The method according to claim 8, wherein, The process of obtaining the predicted concentration of deodorant residue in each space block at the time the garbage truck arrives includes: Obtain the most recent deodorization operation log and / or deodorization operation plan of the deodorization equipment associated with the space block; Combining real-time acquired meteorological parameters such as wind speed, temperature, and humidity, and based on at least one of the preset deodorant residue decay model, deodorant residue decay table, and deodorant residue decay algorithm, the predicted value of deodorant residue concentration when the garbage truck arrives in the space block is calculated.

10. The method according to claim 8, wherein, Also includes: If the odor detection module in the space block does not meet the first activation condition and / or the second activation condition, output status information indicating that the odor detection module cannot monitor, as well as at least one of the following: deodorant residue type, deodorant residue concentration prediction value, and activation time comparison result in the space block.

Images