Method, device and electronic equipment for recycling household waste

Abstract

This disclosure presents a method, apparatus, and electronic device for household waste recycling. One specific embodiment of the method includes: breaking open the bags of target waste using a bag-breaking component to obtain the broken-bag waste; collecting odor signals using an odor sensor to obtain real-time odor signals, wherein the target waste is household waste thrown in through a waste disposal port; controlling a deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on the real-time odor signals; performing coarse sorting of the broken-bag waste to obtain first waste, second waste, and third waste; performing fine sorting of the third waste to obtain fourth waste and fifth waste; drying the fourth waste; and shredding the fifth waste. This embodiment achieves waste sorting during the waste recycling stage, effectively alleviating the processing pressure of waste sorting and recycling at waste stations.

Description

Technical Field

[0001] The embodiments disclosed herein relate to the field of waste recycling, specifically to methods, apparatus, and electronic equipment for recycling household waste. Background Technology

[0002] With increasing urbanization, the amount of household waste generated has increased dramatically. Meanwhile, the conventional waste disposal method of incinerating or landfilling unsorted household waste exacerbates environmental pollution and resource waste. Currently, the standard method involves garbage trucks collecting household waste and transporting it to recycling stations for sorting and recycling. However, as recycling stations are "gathering places" for waste, this method increases the processing pressure on waste sorting and recycling facilities. Summary of the Invention

[0003] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0004] Some embodiments of this disclosure provide methods, apparatus, and electronic devices for recycling municipal solid waste to address the technical problems mentioned in the background section above.

[0005] In a first aspect, some embodiments of this disclosure provide a method for recycling household waste. The method includes: breaking open the bags of target waste using a bag-breaking component to obtain the broken-bag waste; collecting odor signals using an odor sensor to obtain real-time odor signals, wherein the target waste is household waste thrown into the waste disposal port; controlling a deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on the real-time odor signals; performing coarse sorting of the broken-bag waste to obtain first waste, second waste, and third waste, wherein the first waste is magnetic waste, the second waste is plastic waste, and the third waste is a mixture of the broken-bag waste excluding magnetic waste and plastic waste; performing fine sorting of the third waste to obtain fourth waste and fifth waste, wherein the fourth waste is kitchen waste, and the fifth waste is a mixture of the third waste excluding kitchen waste; drying the fourth waste to obtain sixth waste; and pulverizing the fifth waste to obtain seventh waste; and recycling and storing the first waste, second waste, sixth waste, and seventh waste respectively.

[0006] Secondly, some embodiments of this disclosure provide a household waste recycling device, comprising: a bag-breaking and odor-collecting unit configured to break open the target waste using a bag-breaking component to obtain the broken-open waste, and to collect odor signals using an odor sensor to obtain real-time odor signals, wherein the target waste is household waste thrown in through a waste disposal port; a control unit configured to control a deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form according to the real-time odor signals; and a waste coarse sorting unit configured to perform coarse sorting of the broken-open waste to obtain first waste, second waste, and third waste, wherein the first waste is magnetic waste, and the third waste is magnetic waste. The second type of waste is plastic waste. The third type of waste is a mixture of waste (excluding magnetic waste and plastic waste) after the bags have been broken. The waste sorting unit is configured to further sort the third type of waste to obtain the fourth and fifth types of waste, wherein the fourth type of waste is kitchen waste and the fifth type of waste is a mixture of the third type of waste (excluding kitchen waste). The waste drying and shredding unit is configured to dry the fourth type of waste to obtain the sixth type of waste and shred the fifth type of waste to obtain the seventh type of waste. The waste recycling and storage unit is configured to recycle and store the first type of waste, the second type of waste, the sixth type of waste, and the seventh type of waste, respectively.

[0007] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0008] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0009] The above-described embodiments of this disclosure have the following beneficial effects: The waste recycling method of some embodiments of this disclosure achieves waste sorting during the waste recycling stage, effectively alleviating the processing pressure of waste sorting and recycling at waste stations. Specifically, firstly, the target waste is broken open using a bag-breaking component to obtain the broken-open waste, and an odor sensor is used to collect odor signals in real time. The target waste is the household waste thrown into the waste disposal port. In practice, breaking the bag exposes the waste, facilitating subsequent waste sorting. Simultaneously, the waste disposal port often needs to be opened during waste recycling, allowing any odors inside to escape. Therefore, an odor sensor captures the odor signal at the waste disposal port. Secondly, based on the real-time odor signal, a deodorizing liquid spraying component is controlled to spray deodorizing liquid at a fixed frequency in an atomized form. In practice, conventional methods mainly involve spraying deodorizing liquid manually or at set times. However, the maximum storage capacity of deodorizing liquid is often fixed, and existing methods often result in untimely spraying or waste of deodorizing liquid. Therefore, this disclosure utilizes real-time gas signals to quantify the odor status at the garbage disposal port and employs atomization and fixed-frequency spraying to efficiently utilize the deodorizing liquid. Next, the garbage after bagging is coarsely sorted into three categories: first, second, and third waste. The first waste is magnetic waste, the second waste is plastic waste, and the third waste is a mixture of the bagged garbage excluding magnetic and plastic waste. In practice, magnetic and plastic waste are often recyclable in household waste; therefore, this disclosure uses coarse sorting to separate magnetic waste, plastic waste, and other waste. Further, the third waste is finely sorted into four and five categories: fourth waste is kitchen waste, and fifth waste is a mixture of the third waste excluding kitchen waste. In practice, kitchen waste is often non-recyclable. Furthermore, due to the action of microorganisms, kitchen waste continuously produces foul-smelling liquids and odors during transportation, necessitating further fine sorting from other waste. Next, the fourth type of waste is dried to obtain the sixth type, and the fifth type is shredded to obtain the seventh type. Drying the kitchen waste effectively reduces moisture content, lowering transportation costs, and also effectively inhibits microorganisms and kills pathogens, thereby suppressing and preventing odor generation and disease transmission. Finally, the first, second, sixth, and seventh types of waste are collected and stored separately. In summary, this method achieves waste sorting at the recycling stage, effectively alleviating the processing pressure on waste sorting and recycling stations. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0011] Figure 1 This is a flowchart of some embodiments of the municipal solid waste recycling method according to the present disclosure;

[0012] Figure 2 This is a schematic diagram showing the positional relationship between the garbage disposal port, infrared components, high-speed camera, deodorizing liquid spraying components, and bag breaking components;

[0013] Figure 3 This is a schematic diagram illustrating the process of generating object description information;

[0014] Figure 4 This is a functional layout diagram of a garbage truck;

[0015] Figure 5 These are schematic diagrams illustrating the structure of some embodiments of the household waste recycling device according to this disclosure;

[0016] Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0017] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0018] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0019] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0020] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0021] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0022] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] refer to Figure 1 The diagram illustrates a process 100 of some embodiments of a household waste recycling method according to the present disclosure. This household waste recycling method includes the following steps:

[0024] Step 101: The target waste is broken open using the bag-breaking component to obtain the broken waste, and the odor is collected by the odor sensor to obtain a real-time odor signal.

[0025] In some embodiments, the entity executing the household waste recycling method (e.g., a computing device) can break open the target waste using a bag-breaking component to obtain the broken-open waste, and collect odor signals using an odor sensor to obtain real-time odor signals.

[0026] The target waste refers to household waste thrown into the waste disposal port. The bag-breaking assembly can be a bag-breaking machine for bagged household waste. The odor sensor can be a sensor used to collect odor signals at the waste disposal port. For example, the odor sensor can be an electronic nose. The bag-breaking assembly can be located below the waste disposal port. The odor sensor can be located between the bag-breaking assembly and the waste disposal port to collect the odor signal corresponding to the gas overflowing in the direction of the waste disposal port.

[0027] In practice, once the target waste is thrown into the waste disposal port, the bag-breaking component starts working to break the bag of the waste, thus obtaining the broken-bag waste. In addition, an odor sensor can collect odor signals when the corresponding disposal port cover is opened.

[0028] It should be noted that the aforementioned computing devices can be either hardware or software. When the computing device is hardware, it can be implemented as a single server or a single terminal device. When the computing device is software, it can be installed within the hardware devices listed above. It can be implemented as a single software program or a software module. No specific limitations are made here. In practice, the computing device could be a control terminal on a garbage truck.

[0029] In some optional implementations of some embodiments, before the target waste is broken open by the bag-breaking component to obtain the broken-open waste, the method further includes:

[0030] Step S1: In response to determining that the cover of the garbage disposal port is open and the infrared component detects an object blocking the view, the first image sequence is acquired.

[0031] The aforementioned infrared component is activated in conjunction with the garbage disposal port, and the aforementioned first image sequence is acquired by a high-speed camera facing the aforementioned garbage disposal port.

[0032] As an example, see Figure 2 The diagram shows the positional relationship between the garbage disposal port, infrared component, high-speed camera, deodorizing liquid spraying component, and bag-breaking component. The garbage disposal port 1 has an isosceles trapezoidal cross-section, wider at the top and narrower at the bottom. Garbage disposal port 1 is equipped with a disposal port cover 2. The disposal port cover 2 has a split-opening structure. The disposal port cover 2 is controlled to open and close via an electric rotating shaft. When garbage needs to be disposed of, the disposal port cover 2 opens and closes via the electric rotating shaft. An odor sensor 3 is installed on the inner wall of garbage disposal port 1. The odor sensor 3 and garbage port cover 2 open in conjunction. The inner wall of garbage disposal port 1 also houses the atomizing nozzle 4 of the deodorizing liquid spraying component. A bag-breaking component 6 is located below garbage disposal port 1. An infrared component 5 and a high-speed camera 7 facing the garbage disposal port, used to acquire the first image sequence, are installed on the inner wall of the channel above the bag-breaking component 6. The high-speed camera has an elliptical housing made of transparent plastic to protect its lens.

[0033] In practice, whether the garbage disposal opening is open can be determined by detecting whether an opening command has been issued for the motorized shaft control. Simultaneously, whether the infrared component has detected an obstruction can be determined by detecting whether the infrared signal emitted by the component is blocked. Specifically, when an opening command for the motorized shaft control is detected, the corresponding disposal opening cover may fail to open effectively due to mechanical failure. Therefore, this disclosure further combines infrared sensor detection of obstruction to achieve a logical dual-judgment, thereby improving detection accuracy. When the disposal opening cover is detected to be open and the infrared component detects an obstruction, the high-speed camera used to acquire the first image sequence is activated to acquire the first image sequence. This linked control approach avoids ineffective component operation.

[0034] Step S2: Perform object recognition based on the first image sequence and the pre-trained object recognition model to obtain object description information.

[0035] The object description information mentioned above includes: object type, recognition confidence level, and object movement tendency. The object recognition model may include: a downsampling module, a feature enhancement module, a multi-scale feature fusion module, an object type classifier, an optical flow tracing module, and an object movement tendency classifier.

[0036] In practice, see Figure 3The diagram illustrates the process of generating object description information. For each image in the first image sequence, the execution entity first performs multi-scale downsampling processing on the first image using a downsampling module. This downsampling module employs a three-layer FPN (Feature Pyramid Networks) model. Assume the image size of the first image is H×W×3, where "H" represents the height, "W" represents the width, and "3" represents the number of channels. After multi-scale downsampling, the downsampling module outputs three feature maps with corresponding sizes of H×W×3, H / 4×W / 4×3, and H / 8×W / 8×3, respectively. Next, for each of these three feature maps, the execution entity enhances its features using a feature enhancement module, resulting in an enhanced feature map. The enhanced feature map has the same size as the corresponding feature map. The feature enhancement module consists of seven sequentially connected convolutional layers. A skip connection is established between the 2nd and 7th convolutional layers, and between the 3rd and 6th convolutional layers. The feature enhancement module enhances the features of each of the three feature maps sequentially, resulting in three enhanced feature maps. The dimensions of the three enhanced feature maps are H×W×3, H / 4×W / 4×3, and H / 8×W / 8×3, respectively. Next, the execution entity uses a multi-scale feature fusion module to fuse the features of the two enhanced feature maps with dimensions H×W×3 and H / 4×W / 4×3. Specifically, the multi-scale feature fusion module includes two downsampling networks at different scales (one consisting of three convolutional layers (inputting an enhanced feature map of size H×W×3), and the other consisting of two convolutional layers (inputting an enhanced feature map of size H / 4×W / 4×3)). These networks process the two enhanced feature maps and then stack them (the resulting feature map has a size of H / 8×W / 8×3). The result is then upsampled through a convolutional layer to obtain a first upsampled feature map (with a size of H×W×3). Next, the pre- and post-enhanced feature maps (with a size of H / 8×W / 8×3) are input into the upsampling convolutional layer in the multi-scale feature fusion module to obtain a second sampled feature map (with a size of H×W×3). Finally, the first and second sampled feature maps are stacked to obtain the fused feature. Furthermore, the aforementioned execution entity inputs the fused features into an object type classifier (using a multi-classifier) ​​to obtain the object type and recognition confidence corresponding to the first image. Simultaneously, by processing each first image in the first image sequence, a fused feature sequence is obtained.Next, the aforementioned execution entity inputs the fused feature sequence into the optical flow tracing module (using the LucasKanade algorithm) to generate optical flow trajectory features. Finally, the execution entity inputs the optical flow trajectory features into the object movement tendency classifier (using a multi-classifier) ​​to obtain the object movement tendency. In particular, since a corresponding object type and recognition confidence score are obtained for each first image, the object type included in the object description information can be determined through voting. Simultaneously, the recognition confidence score corresponding to the voted object type is obtained through a weighted summation of confidence scores.

[0037] The aforementioned object recognition model is a core inventive point of this disclosure. Since the high-speed camera used to acquire the first image sequence is located on the inner wall of the channel above the bag-breaking assembly, although some light enters when the dispensing port cover is open, the high-speed camera is still in a low-light environment. Therefore, how to perform object recognition in a low-light environment with low computational requirements, high detection speed, and high accuracy becomes crucial to obtaining accurate object description information. Therefore, the object recognition model of this disclosure first extracts multi-scale feature representations through a downsampling module. In practice, compared to images acquired under sufficient light, low-light images have significant differences in pixel distribution. Therefore, this disclosure captures edge and texture features under different receptive fields through multi-scale feature extraction. Next, the feature enhancement module further enhances feature representation, while skip connections are set to avoid feature forgetting. Furthermore, considering that low-resolution features contain edge and texture details, while high-resolution features contain content such as shape, the multi-scale feature fusion module of this disclosure further captures edge and texture details from two enhanced feature maps of higher resolution, and then upsamples and superimposes these details with the enhanced feature map of the lowest resolution. While enriching edge and texture details, the original image size is restored. Finally, by combining an object type classifier, an optical flow tracing module, and an object movement tendency classifier, accurate classification of object type and object movement tendency is achieved. Simultaneously, the entire network structure is decoupled and lightweight to meet the goals of low computational power requirements and high detection speed.

[0038] Step S3: In response to the above object description information meeting the preset conditions, control the bag breaking component to open.

[0039] The aforementioned preset conditions are: the object type is within the object whitelist, the recognition confidence level is greater than the preset recognition confidence level, and the object's movement tendency is characterized by moving towards the bag-breaking component. The object whitelist is a pre-constructed catalog of objects that can be processed for waste disposal.

[0040] Step S4: In response to the fact that the above object description information does not meet the above preset conditions, a foreign object entry warning is initiated.

[0041] In practice, when the object description information does not meet the above preset conditions, a foreign object entry warning can be issued to the corresponding driver of the garbage truck.

[0042] In practice, since living beings may enter the garbage disposal opening when it is open (for example, garbage collectors or garbage dumpers may illegally put their limbs into the garbage disposal opening), the rotation of the bag breaking component may cause injury to the living beings. Therefore, this disclosure combines the first image sequence to determine the object description information and combines preset conditions to avoid this tendency from occurring, thereby ensuring the safety of garbage recycling.

[0043] Step 102: Based on the real-time odor signal, control the deodorizing liquid spraying component to spray the deodorizing liquid at a fixed frequency in the form of atomization.

[0044] In some embodiments, the aforementioned executing entity can control the deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on real-time odor signals.

[0045] The deodorizing liquid spraying assembly is used for spraying deodorizing liquid. The assembly may include an atomizing nozzle, a reservoir, and a miniature DC pump. The assembly is connected to the reservoir via piping. The reservoir contains deodorizing liquid. A liquid level monitoring device is installed inside the reservoir. A miniature DC pump is installed between the assembly and the reservoir. This pump pumps the deodorizing liquid from the reservoir to the atomizing nozzle, allowing it to be sprayed at a constant frequency in an atomized form.

[0046] In practice, the aforementioned implementing entity can perform signal analysis on the odor signal to determine the odor concentration corresponding to a specific odor type, and control the deodorizing liquid spraying component to spray the deodorizing liquid according to the preset spraying amount and frequency based on the concentration range corresponding to the odor concentration.

[0047] In some optional implementations of certain embodiments, the execution entity controls the deodorizing liquid spraying assembly to perform fixed-frequency spraying of deodorizing liquid in an atomized form based on the aforementioned real-time odor signal, including:

[0048] Step S1: Determine the reference odor signal.

[0049] The baseline odor signal is the odor signal collected by the odor sensor when the garbage disposal port is open and the bag-breaking component is not activated.

[0050] In practice, assume that the opening time of the garbage disposal port cover is time T1, and the opening time of the bag-breaking component is time T2, where time T2 is longer than time T1. Since the odor sensor collects odor signals when the garbage disposal port cover is open, the odor signals collected between time T1 and time T2 can be used as the reference odor signals.

[0051] Step S2: Based on the reference odor signal, perform signal filtering on the above real-time odor signal to obtain the filtered odor signal.

[0052] In practice, firstly, the mean value of the corresponding signal can be determined based on the baseline odor signal. Then, the mean value of the real-time odor signal at each time point is subtracted from the mean value to obtain the filtered odor signal.

[0053] Specifically, when the target waste contains garbage with a strong odor, a strong odor will also be generated after the bag is broken. Therefore, this disclosure prevents odor leakage by installing a deodorizing liquid spraying component on the side of the waste disposal port. To highlight the odor generated by the target waste after the bag is broken, the real-time odor signal is filtered using a reference odor signal.

[0054] Step S3: Determine the temperature gain value based on the average temperature corresponding to the real-time odor signal acquisition, and determine the humidity gain value based on the average humidity corresponding to the real-time odor signal acquisition.

[0055] The mapping relationship between the average temperature and the temperature gain value has been pre-calibrated, and the mapping relationship between the average humidity and the humidity gain value has been pre-calibrated.

[0056] In practice, odors originate from volatile organic compound molecules. Increased temperature intensifies intermolecular motion, resulting in a stronger odor than at room temperature. Furthermore, water vapor dissolves and adsorbs some water-soluble odor molecules; higher humidity increases the propagation of these molecules. Therefore, different humidity levels and temperatures affect odor propagation. Assuming that only the waste disposal chute is connected to the external environment during waste recycling, the effects of different humidity levels and temperatures on the same odor can be determined through pre-calibration, thus obtaining corresponding temperature and humidity gain values. Specifically, average temperature and average humidity can be collected by temperature and humidity sensors, respectively. A pre-set mapping table is used to determine the corresponding temperature and humidity gain values ​​for the average temperature and humidity. Both temperature and humidity gain values ​​are coefficients between 0 and 1. For example, the higher the average temperature, the greater the corresponding temperature gain value.

[0057] Step S4: Based on the above temperature gain value and the above humidity gain value, perform signal gain on the above filtered odor signal to obtain the gained odor signal.

[0058] Wherein, the amplified signal = filtered odor signal × (first weight × temperature gain value + second weight × humidity gain value) / 2. Both the first and second weights are positive numbers, and their sum is 1. Specifically, since temperature has a greater impact on odor than humidity, the first weight can be 0.7, and the second weight can be 0.3.

[0059] Step S5: Generate odor description information based on the pre-trained odor recognition model and the above-mentioned post-gain odor signal.

[0060] The odor description information may include: odor type and odor concentration.

[0061] In practice, since steps S1 to S4 have effectively processed the real-time odor signal from three dimensions—the baseline odor signal, the temperature gain value, and the humidity gain value—and considering the limitations of computing power and the requirements for recognition speed, the odor recognition model adopts a weak classifier architecture. The odor recognition model consists of: a first 1D convolutional layer, a first 1D max pooling layer, a second 1D convolutional layer, a second 1D max pooling layer, a Dropout layer, a first linear layer, a second linear layer, a third linear layer, a first Softmax classifier, and a second Sormax layer. The first 1D convolutional layer, the first 1D max pooling layer, the second 1D convolutional layer, the second 1D max pooling layer, the Dropout layer, the first linear layer, the second linear layer, and the third linear layer are connected sequentially. The first Softmax classifier and the second Sormax layer are connected in parallel to the third linear layer. The first Softmax classifier is used to classify odor type, and the second Sormax layer is used to classify odor concentration. Specifically, the odor recognition model is trained using supervised training, allowing for manual labeling of historically collected, amplified odor signals to determine their corresponding odor type and concentration, which serves as training samples and labeling. The advantage of using a weak classifier is that it requires relatively fewer computational resources, making it suitable for applications with limited computing power. Furthermore, weak classifiers have a simple architecture, fast training speed, and facilitate rapid model deployment and subsequent model updates.

[0062] Step S5: Determine the spray control information based on the odor description information.

[0063] The spraying control information includes: single spray volume, spraying frequency, single spraying duration, and number of sprays.

[0064] In practice, spray control information tables can be preset according to the type and concentration of odor. Therefore, spray control information can be quickly determined by looking up the table.

[0065] Step S6: Based on the above spray control information, control the deodorizing liquid spraying component to spray the deodorizing liquid at a fixed frequency in the form of atomization.

[0066] In practice, conventional deodorizing liquid spraying methods often control the spraying time and amount by setting thresholds. However, different types and concentrations of waste require different amounts of deodorizing liquid for odor suppression. Setting the threshold too low results in poor deodorization, while setting it too high leads to ineffective waste due to the limited capacity of the storage tank. Therefore, this disclosure uses an odor sensor to capture real-time odor signals and adjusts the expression of odor from three dimensions: a baseline odor signal, humidity, and temperature. Finally, a weak classifier is used to quickly generate spraying control information. This method dynamically adjusts the spraying amount, number of sprays, and spraying time according to the type of waste, effectively improving both the deodorization effect and the utilization efficiency of the deodorizing liquid.

[0067] Step 103: Perform preliminary sorting of the waste after the bags are broken to obtain three categories: first waste, second waste, and third waste.

[0068] In some embodiments, the aforementioned implementing entity may perform preliminary waste sorting on the waste after the bags are broken to obtain first waste, second waste, and third waste.

[0069] The first type of waste is magnetic waste. The second type of waste is plastic waste. The third type of waste is a mixture of waste other than magnetic and plastic waste found after the bags have been broken.

[0070] In practice, magnetic waste and plastic waste are often recyclable. However, plastic is often a polymer material, and burning it along with other waste without proper separation will produce large amounts of toxic and harmful gases. Therefore, a preliminary sorting of waste into primary, secondary, and tertiary categories is necessary. Specifically, magnetic separation and air separation devices can be used to separate primary and secondary waste, with the remaining waste classified as tertiary waste.

[0071] In some optional implementations of certain embodiments, the aforementioned executing entity performs coarse waste sorting on the waste after the bags are broken, obtaining first waste, second waste, and third waste, including:

[0072] Step S1: Control the magnetic separation component to perform magnetic separation of the above-mentioned bag-broken waste to obtain the first waste.

[0073] The magnetic separation component is located below the bag-breaking component. Specifically, the magnetic separation component consists of a magnetic separator and a magnetic waste collection component. The magnetic waste collection component is equipped with an overflow alarm.

[0074] In practice, the aforementioned implementing entity can control the magnetic separation component to open, thereby separating magnetic waste and obtaining the first type of waste.

[0075] Step S2: Acquire the second image sequence.

[0076] The second image sequence was acquired by a high-speed camera positioned below the magnetic separator.

[0077] In practice, after the garbage is separated by magnetic separation, the garbage other than the first type of garbage will pass through the magnetic separator and fall further. Therefore, a second image sequence can be captured by a high-speed camera positioned below the magnetic separator. In particular, the high-speed camera positioned below the magnetic separator is turned on synchronously with the magnetic separator to capture the second image sequence.

[0078] Step S3: Perform image enhancement on the second image in the above second image sequence to obtain the enhanced image sequence.

[0079] In practice, because the ambient light is weaker at the high-speed camera used for acquiring the second image sequence, an image enhancement step is added for the second image compared to the high-speed camera used for acquiring the first image sequence. Specifically, a logarithmic transform-based image enhancement method can be used to quickly enhance the brightness of the second image, resulting in the enhanced image.

[0080] Step S4: Generate a first target information set based on the enhanced image sequence and the pre-trained first multi-target recognition model.

[0081] The first target information includes the target location and target type. The target location represents the three-dimensional coordinates transformed from the two-dimensional image coordinate system to the three-dimensional coordinate system. The first multi-target recognition model uses the NanoDet-Plus model as the backbone network. The NanoDet-Plus model was chosen because it is lightweight, highly accurate, and easy to deploy on low-computing-power platforms. In particular, the targets identified by the NanoDet-Plus model are image coordinates in the image coordinate system (two-dimensional image coordinate system). To facilitate further garbage sorting, the two-dimensional image coordinates need to be converted into three-dimensional coordinates. Considering that the position and tilt angle of the high-speed camera used to acquire the second image sequence are fixed, the rotation matrix and translation vector corresponding to the high-speed camera can be pre-calibrated. The two-dimensional image coordinates are then converted into the three-dimensional coordinates corresponding to the target location by combining the rotation matrix and translation vector. At the same time, a multi-classifier needs to be set at the end of the NanoDet-Plus model for target classification of multiple targets.

[0082] Step S5: Based on the first target information set mentioned above, the first sorting component sorts the waste in the bag-broken waste, excluding the first waste, to obtain the second waste and the third waste.

[0083] The first sorting component is an air separator. It is positioned below a high-speed camera used to acquire a second image sequence. This first sorting component receives the waste filtered by the magnetic separator and further performs air separation on the waste.

[0084] In practice, given the location and type of the target (waste), the first sorting component located below the magnetic separator can be controlled to sort the waste (excluding the first type of waste) from the bag-broken waste according to the first target information set, thus obtaining the second and third types of waste.

[0085] Step 104: Further sort the third type of waste to obtain the fourth and fifth types of waste.

[0086] In some embodiments, the aforementioned implementing entity may further classify the third type of waste to obtain the fourth and fifth types of waste.

[0087] The fourth type of waste is kitchen waste. The fifth type of waste is mixed waste other than kitchen waste, which is part of the third type of waste.

[0088] In some optional implementations of certain embodiments, the aforementioned implementing entity performs further sorting of the third type of waste to obtain the fourth and fifth types of waste, including:

[0089] Step S1: Transfer the aforementioned third type of waste to the waste sorting area by lifting the component.

[0090] The lifting component is located below the first sorting component. The third type of waste falls into the lifting component after passing through the first sorting component. The lifting component can be a hoist.

[0091] As an example, see Figure 4 The diagram shows the functional zoning of a garbage truck. The processing area for the third type of waste (detailed waste sorting area) and the processing area for the target waste (coarse waste sorting area) are arranged in parallel, and the third type of waste is transferred via a lifting assembly. Both waste processing areas are located inside the garbage truck's shell to prevent odor leakage. In practice, due to height limitations of garbage trucks, it is not possible to install further waste sorting devices below the first sorting assembly; therefore, the third type of waste needs to be transferred to the detailed waste sorting area via the lifting assembly.

[0092] Step S2: In response to the lifting component moving to the target position and starting the garbage dumping, the third image group sequence is acquired.

[0093] The aforementioned third image sequence is acquired by a high-speed camera facing the waste sorting area. The third image sequence includes infrared images and full-color images. Specifically, the high-speed camera used to acquire the third image sequence is a binocular camera, comprising an infrared camera and a full-color camera. The high-speed camera used to acquire the third image sequence is positioned facing the waste falling location corresponding to the lifting component, thereby acquiring images of the waste (third type of waste) dumped by the lifting component after it moves to the target position.

[0094] Step S4: For each third image group in the above third image group sequence, perform the following processing steps:

[0095] Step S41: Extract image features from the infrared image and the full-color image included in the third image group to obtain infrared image features and full-color image features respectively.

[0096] In practice, the aforementioned executing entity can activate model instances corresponding to the downsampling module, feature enhancement module, and multi-scale feature fusion module of the first multi-target recognition model. Specifically, the activated model instances of the first multi-target recognition model maintain the same model parameters for the first multi-target recognition model, and are used to extract image features from the full-color image included in the third image. The activated model instances of the second multi-target recognition model load model parameters specific to the infrared image, and after loading, extract image features from the infrared image to obtain both infrared image features and full-color image features.

[0097] In practice, in order to further improve the reusability of the model and avoid the problems of high model training cost, adaptation cost and deployment cost of training dedicated models, parallel feature extraction of infrared images and full-color images is achieved by sharing the downsampling module, feature enhancement module and multi-scale feature fusion module included in the first multi-target recognition model.

[0098] Step S42: Generate candidate target information groups based on the second multi-target recognition model pre-trained according to the above infrared image features, the above full-color image features, and the above fused image features.

[0099] The second multi-target recognition model consists of a feature fusion module and a multi-target localization head based on bounding box regression. The feature fusion module is used to superimpose features from the full-color image and the fused image. Specifically, the feature fusion module consists of a dual-channel convolutional network and a feature stacking layer. The dual-channel convolutional network comprises two parallel convolutional networks, each consisting of three serially connected convolutional layers. These two convolutional networks further extract and shape features from the full-color image and infrared image in parallel. The feature stacking layer is used to superimpose features from the input of the dual-channel convolutional network. The candidate target information corresponds to the identified target (garbage) and includes the target location, target type, and a 256-dimensional target description vector.

[0100] Step S5: Perform multi-target association based on the obtained candidate target information group set to obtain the second target information set.

[0101] In practice, based on the target description vectors included in the candidate target information, similarity calculations can be used to associate candidate target information within the candidate target information set, thereby obtaining a second target information set. In particular, the subsequent waste sorting process still requires the use of three-dimensional coordinates. Therefore, after target association, coordinate transformation is needed using the rotation matrix and translation vector corresponding to the high-speed camera capturing the third image sequence, converting the two-dimensional image coordinates into the three-dimensional coordinates included in the second target information. Since the position and tilt angle of the high-speed camera capturing the third image sequence are fixed, the corresponding rotation matrix and translation vector can be pre-calibrated.

[0102] Step S5: Based on the second target information set mentioned above, the third type of waste is sorted by the second sorting component to obtain the fourth and fifth types of waste.

[0103] In practice, the second sorting component can be an air separator. The second sorting component is positioned below a high-speed camera used to acquire a third image sequence. The second sorting component receives the third type of waste dumped by the lifting component and further performs air separation on the waste.

[0104] Step 105: Dry the fourth type of waste to obtain the sixth type of waste, and crush the fifth type of waste to obtain the seventh type of waste.

[0105] In some embodiments, the aforementioned executing entity may dry the fourth waste to obtain the sixth waste, and pulverize the fifth waste to obtain the seventh waste.

[0106] The fourth type of waste, after being sorted by the second sorting component, is blown into a dryer for drying kitchen waste. Once the second sorting component stops working, the aforementioned actuator controls the dryer to start drying the fourth type of waste, resulting in the sixth type of waste. Simultaneously, the fifth type of waste falls into a waste shredder below the second sorting component, resulting in the seventh type of waste. The waste shredder and the second sorting component start synchronously.

[0107] Step 106: Collect and store the first type of waste, the second type of waste, the sixth type of waste, and the seventh type of waste respectively.

[0108] In some embodiments, the aforementioned implementing entities may respectively collect and store the first type of waste, the second type of waste, the sixth type of waste, and the seventh type of waste.

[0109] In some optional implementations of certain embodiments, the aforementioned executing entity performs garbage collection and storage on the first garbage, the second garbage, the sixth garbage, and the seventh garbage, respectively, including:

[0110] Step S1: Store the first type of waste into the magnetic waste collection component.

[0111] The magnetic waste collection component is connected to the magnetic separator.

[0112] In practice, the first type of waste (magnetic waste) separated by the magnetic separator is scraped off by the upper scraper of the magnetic separator and falls into the magnetic waste collection component.

[0113] Step s2: Compress the second type of waste to obtain compressed second type of waste.

[0114] In practice, the second type of waste (plastic waste) is blown into the compressor after being sorted by the first sorting component. When the first sorting component stops working, the compressor will compress the second type of waste to obtain compressed second type of waste.

[0115] Step S3: Store the compressed second type of waste into the plastic waste collection component.

[0116] In practice, the compressed second waste is transported via conveyor belt and moved into the plastic waste collection unit.

[0117] Step S4: Store the sixth type of waste into the kitchen waste collection component.

[0118] In practice, the sixth type of waste (dried kitchen waste) is dried by a dryer and then transported by rotation to the kitchen waste collection unit.

[0119] Step S5: Store the aforementioned seventh type of waste into the mixed waste collection component.

[0120] The aforementioned magnetic waste collection component, plastic waste collection component, kitchen waste collection component, and mixed waste collection component are all equipped with overflow alarm sensors. These sensors provide an overflow warning when the collection component is full. The mixed waste collection component can be installed at the corresponding shredder outlet of the waste shredder.

[0121] In some optional implementations of some embodiments, the above method further includes:

[0122] Step S1: Determine the amount of residual liquid after spraying based on the total amount sprayed and the current liquid volume corresponding to the deodorizing liquid.

[0123] The total spray volume mentioned above is the product of the single spray volume and the number of sprays included in the spray control information. The current liquid volume of the deodorizing liquid is monitored by the liquid level monitoring device installed in the storage tank. Residual liquid volume after spraying = Current liquid volume of deodorizing liquid - Total spray volume.

[0124] Step S2: In response to the residual liquid volume after spraying being less than or equal to the preset residual liquid volume threshold, a low liquid volume warning is initiated.

[0125] In practice, a low liquid level warning can be issued to the corresponding driver of the garbage truck or the garbage collector to remind them to replenish the deodorizing liquid.

[0126] Step S3: By using the weighing component located below the plastic waste collection component, determine the weight difference of the plastic waste collection component before and after collecting the compressed second waste, and use it as the gross weight of the plastic waste.

[0127] Step S4: Determine the moisture content factor corresponding to the above-mentioned plastic waste collection components.

[0128] In practice, humidity values ​​can be collected by a humidity sensor installed inside the plastic waste collection component, which can then be used as a moisture content factor.

[0129] Step S5: Based on the above moisture content factor, correct the gross weight of the above plastic waste to obtain the corrected weight of the plastic waste.

[0130] In practice, firstly, assuming the internal volume of the plastic waste collection component is known, the moisture content of the component can be calculated using the moisture content calculation factor (water vapor content per unit space). Secondly, the difference between the gross weight of the plastic waste and its moisture content can be used as the corrected weight of the plastic waste. In particular, when the units corresponding to the gross weight and moisture content of the plastic waste are different, it is necessary to unify the units before calculating the difference.

[0131] Step S6: Generate a return value based on the weight of the plastic waste and the corresponding reward factor.

[0132] The reward factor can be the value corresponding to a unit weight of plastic waste. The return amount can represent the amount returned to the account that initiated the recycling of household waste targeting the aforementioned waste.

[0133] In practice, the product of the weight of the plastic waste and the corresponding reward factor can be used as the return value.

[0134] Step S7: Based on the above-mentioned return value, a reward will be returned to the account that initiated the recycling of household waste targeting the above-mentioned waste.

[0135] In practice, electronic transfers can be initiated to the account that initiated the collection of household waste targeting the aforementioned waste, based on the aforementioned returned values. In particular, the returned values ​​can be manually verified before the electronic transfer.

[0136] The above-described embodiments of this disclosure have the following beneficial effects: The waste recycling method of some embodiments of this disclosure achieves waste sorting during the waste recycling stage, effectively alleviating the processing pressure of waste sorting and recycling at waste stations. Specifically, firstly, the target waste is broken open using a bag-breaking component to obtain the broken-open waste, and an odor sensor is used to collect odor signals in real time. The target waste is the household waste thrown into the waste disposal port. In practice, breaking the bag exposes the waste, facilitating subsequent waste sorting. Simultaneously, the waste disposal port often needs to be opened during waste recycling, allowing any odors inside to escape. Therefore, an odor sensor captures the odor signal at the waste disposal port. Secondly, based on the real-time odor signal, a deodorizing liquid spraying component is controlled to spray deodorizing liquid at a fixed frequency in an atomized form. In practice, conventional methods mainly involve spraying deodorizing liquid manually or at set times. However, the maximum storage capacity of deodorizing liquid is often fixed, and existing methods often result in untimely spraying or waste of deodorizing liquid. Therefore, this disclosure utilizes real-time gas signals to quantify the odor status at the garbage disposal port and employs atomization and fixed-frequency spraying to efficiently utilize the deodorizing liquid. Next, the garbage after bagging is coarsely sorted into three categories: first, second, and third waste. The first waste is magnetic waste, the second waste is plastic waste, and the third waste is a mixture of the bagged garbage excluding magnetic and plastic waste. In practice, magnetic and plastic waste are often recyclable in household waste; therefore, this disclosure uses coarse sorting to separate magnetic waste, plastic waste, and other waste. Further, the third waste is finely sorted into four and five categories: fourth waste is kitchen waste, and fifth waste is a mixture of the third waste excluding kitchen waste. In practice, kitchen waste is often non-recyclable. Furthermore, due to the action of microorganisms, kitchen waste continuously produces foul-smelling liquids and odors during transportation, necessitating further fine sorting from other waste. Next, the fourth type of waste is dried to obtain the sixth type, and the fifth type is shredded to obtain the seventh type. Drying the kitchen waste effectively reduces moisture content, lowering transportation costs, and also effectively inhibits microorganisms and kills pathogens, thereby suppressing and preventing odor generation and disease transmission. Finally, the first, second, sixth, and seventh types of waste are collected and stored separately. In summary, this method achieves waste sorting at the recycling stage, effectively alleviating the processing pressure on waste sorting and recycling stations.

[0137] Further reference Figure 5As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a municipal solid waste recycling device, which are similar to... Figure 1 Corresponding to the method embodiments shown, this household waste recycling device can be specifically applied to various electronic devices.

[0138] like Figure 5 As shown, some embodiments of the household waste recycling device 500 include: a bag-breaking and odor-collecting unit 501, a control unit 502, a waste coarse sorting unit 503, a waste fine sorting unit 504, a waste drying and shredding unit 505, and a waste recycling and storage unit 506. The bag-breaking and odor-collecting unit 501 is configured to break open the target waste using a bag-breaking component to obtain the broken-open waste, and to collect odor signals using an odor sensor to obtain real-time odor signals. The target waste is household waste thrown in through the waste disposal port. The control unit 502 is configured to control a deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on the real-time odor signals. The waste coarse sorting unit 503 is configured to perform coarse sorting of the broken-open waste to obtain... The waste is categorized into three types: first waste, second waste, and third waste. First waste is magnetic waste, second waste is plastic waste, and third waste is a mixture of waste (excluding magnetic and plastic waste) after the bags have been broken. A waste sorting unit 504 is configured to further sort the third waste to obtain fourth and fifth waste. Fourth waste is kitchen waste, and fifth waste is a mixture of the third waste (excluding kitchen waste). A waste drying and shredding unit 505 is configured to dry the fourth waste to obtain sixth waste and shred the fifth waste to obtain seventh waste. A waste recycling and storage unit 506 is configured to recycle and store the first, second, sixth, and seventh wastes respectively.

[0139] It is understandable that the units described in the household waste recycling device 500 are consistent with the reference. Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method also apply to the municipal solid waste recycling device 500 and the units contained therein, and will not be repeated here.

[0140] The following is for reference. Figure 6 It shows a schematic diagram of the structure of an electronic device (e.g., a computing device) 600 suitable for implementing some embodiments of the present disclosure. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0141] like Figure 6 As shown, the electronic device 600 may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory 602 or a program loaded from a storage device 608 into a random access memory 603. The random access memory 603 also stores various programs and data required for the operation of the electronic device 600. The processing unit 601, the read-only memory 602, and the random access memory 603 are interconnected via a bus 604. An input / output interface 605 is also connected to the bus 604.

[0142] Typically, the following devices can be connected to the input / output interface 605: input devices 606 including, for example, a touchscreen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 607 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 608 including, for example, magnetic tape, hard disk, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 6 Each box shown can represent a device or multiple devices as needed.

[0143] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a read-only memory 602. When the computer program is executed by the processing device 601, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0144] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0145] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol), and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0146] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs. When the electronic device executes the aforementioned one or more programs, the electronic device causes the following actions: to break open the target waste using a bag-breaking component to obtain the broken-open waste, and to collect odor signals using an odor sensor to obtain real-time odor signals, wherein the target waste is household waste thrown in through the waste disposal port; to control a deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on the real-time odor signals; to perform coarse waste sorting on the broken-open waste to obtain first waste, second waste, and third waste, wherein the first waste is magnetic waste, the second waste is plastic waste, and the third waste is a mixture of the broken-open waste excluding magnetic waste and plastic waste; to perform fine waste sorting on the third waste to obtain fourth waste and fifth waste, wherein the fourth waste is kitchen waste, and the fifth waste is a mixture of the third waste excluding kitchen waste; to dry the fourth waste to obtain sixth waste, and to pulverize the fifth waste to obtain seventh waste; and to collect and store the first waste, the second waste, the sixth waste, and the seventh waste respectively.

[0147] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0148] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0149] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0150] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A method for recycling household waste, characterized in that, include: In response to determining that the corresponding garbage disposal port cover is open and the infrared component detects an object obstructing the opening, a first image sequence is acquired, wherein the infrared component is activated in conjunction with the garbage disposal port, and the first image sequence is acquired by a high-speed camera facing the garbage disposal port; Object recognition is performed based on the first image sequence and a pre-trained object recognition model to obtain object description information, which includes: object type, recognition confidence, and object movement tendency. The object recognition model includes: a downsampling module, a feature enhancement module, a multi-scale feature fusion module, an object type classifier, an optical flow tracing module, and an object movement tendency classifier. The downsampling module uses a three-layer FPN model. The input to the downsampling module is a first image with dimensions H×W×3, and the output is three feature maps with dimensions H×W×3, H / 4×W / 4×3, and H / 8×W / 8×3. The feature enhancement module consists of seven sequentially connected convolutional layers, with a skip connection between the second and seventh convolutional layers, and a skip connection between the third and fourth convolutional layers. A skip connection is set between the convolutional layer and the 6th convolutional layer. The output of the feature enhancement module is three feature maps with size H×W×3, H / 4×W / 4×3, and H / 8×W / 8×3. The multi-scale feature fusion module includes two downsampling networks of different scales and a convolutional layer for upsampling. One downsampling network consists of three convolutional layers and takes the enhanced feature map of size H×W×3 as input. The other downsampling network consists of two convolutional layers and takes the enhanced feature map of size H / 4×W / 4×3 as input. The convolutional layer for upsampling is used to upsample the outputs of the two downsampling networks of different scales and to upsample the enhanced feature map of size H / 8×W / 8×3. The object type classifier uses a multi-classifier, the optical flow tracing module uses the Lucas Kanade algorithm, and the object movement tendency classifier uses a multi-classifier. In response to the object description information meeting preset conditions, the bag breaking component is controlled to open. The preset conditions are: the object type is within the object whitelist, the recognition confidence is greater than the preset recognition confidence, and the object's movement tendency is characterized by moving towards the bag breaking component. In response to the object description information not meeting the preset conditions, a foreign object entry warning is initiated; The target waste is processed by breaking the bag using a bag-breaking component to obtain the waste after breaking the bag, and the odor is collected by an odor sensor to obtain a real-time odor signal. The target waste is household waste thrown in through the waste disposal port. Based on the real-time odor signal, the deodorizing liquid spraying component is controlled to spray the deodorizing liquid at a fixed frequency in the form of atomization. The waste after the bags are broken is subjected to coarse waste sorting to obtain first waste, second waste and third waste, wherein the first waste is magnetic waste, the second waste is plastic waste, and the third waste is mixed waste other than magnetic waste and plastic waste in the waste after the bags are broken; The third type of waste is further sorted to obtain the fourth and fifth types of waste. The fourth type of waste is kitchen waste, and the fifth type of waste is the mixed waste in the third type of waste, excluding kitchen waste. The fourth type of waste is dried to obtain the sixth type of waste, and the fifth type of waste is crushed to obtain the seventh type of waste; The first type of waste, the second type of waste, the sixth type of waste, and the seventh type of waste are respectively recycled and stored. The step of controlling the deodorizing liquid spraying component to spray deodorizing liquid at a fixed frequency in an atomized form based on the real-time odor signal includes: Determine the baseline odor signal, which is the odor signal collected by the odor sensor when the garbage disposal opening is open and the bag breaking component is not open; Based on the reference odor signal, the real-time odor signal is filtered to obtain the filtered odor signal; A temperature gain value is determined based on the average temperature corresponding to the real-time odor signal acquisition, and a humidity gain value is determined based on the average humidity corresponding to the real-time odor signal acquisition. The mapping relationship between the average temperature and the temperature gain value is pre-calibrated, and the mapping relationship between the average humidity and the humidity gain value is pre-calibrated. The filtered odor signal is amplified based on the temperature gain value and the humidity gain value to obtain the amplified odor signal. Based on the pre-trained odor recognition model and the post-gain odor signal, odor description information is generated; Based on the odor description information, the spray control information is determined, which includes: single spray volume, spray frequency, single spray duration, and number of sprays; According to the spray control information, the deodorizing liquid spraying component is controlled to spray the deodorizing liquid at a fixed frequency in the form of atomization.

2. The method for recycling household waste according to claim 1, characterized in that, The preliminary sorting of the waste after the bags are broken, resulting in first waste, second waste, and third waste, includes: The magnetic separation component is controlled to perform magnetic separation of the waste after the bag is broken to obtain the first waste, wherein the magnetic separation component is located below the bag breaking component; Acquire a second image sequence, wherein the second image sequence is acquired by a high-speed camera facing downwards from the magnetic separation component; Image enhancement is performed on the second image in the second image sequence to obtain an enhanced image sequence; Based on the enhanced image sequence and the pre-trained first multi-target recognition model, a first target information set is generated, wherein the first target information includes: target location and target type, and the target location represents the three-dimensional coordinates transformed from the two-dimensional image coordinate system to the three-dimensional coordinate system; Based on the first target information set, the first sorting component sorts the waste in the bag-broken waste, excluding the first waste, to obtain the second waste and the third waste.

3. The method for recycling household waste according to claim 2, characterized in that, The process of further sorting the third type of waste to obtain the fourth and fifth types of waste includes: By upgrading the components, the third type of waste is transferred to the waste sorting area; In response to the lifting component moving to the target position and starting the garbage dumping, a third image sequence is acquired, wherein the third image sequence is acquired by a high-speed camera facing the garbage sorting area, and the third image group includes: infrared images and full-color images; For each third image group in the third image group sequence, the following processing steps are performed: Image features are extracted from the infrared image and the full-color image included in the third image group respectively to obtain infrared image features and full-color image features; The infrared image features and the full-color image features are fused to generate fused image features; Based on the infrared image features, the full-color image features, and the fused image features, a second multi-target recognition model pre-trained is used to generate candidate target information groups. Based on the obtained candidate target information set, multi-target association is performed to obtain the second target information set; Based on the second target information set, the third type of waste is sorted by the second sorting component to obtain the fourth and fifth types of waste.

4. The method for recycling household waste according to claim 3, characterized in that, The process of recycling and storing the first type of waste, the second type of waste, the sixth type of waste, and the seventh type of waste includes: The first type of waste is stored in the magnetic waste collection component; The second type of waste is compressed to obtain compressed second type of waste; The compressed second waste is stored in the plastic waste collection component; The sixth type of waste is stored in the kitchen waste collection component; The seventh type of waste is stored in a mixed waste storage component. The magnetic waste storage component, the plastic waste storage component, the kitchen waste storage component, and the mixed waste storage component are all equipped with an overflow alarm sensor. The overflow alarm sensor is used to issue an overflow warning when the storage component is full.

5. The method for recycling household waste according to claim 4, characterized in that, The method further includes: The amount of residual liquid after spraying is determined based on the total amount sprayed and the current liquid volume corresponding to the deodorizing liquid, wherein the total amount sprayed is the product of the single spray volume and the number of sprays included in the spraying control information. In response to the residual liquid level being less than or equal to a preset residual liquid level threshold after spraying, a low liquid level warning is issued. The weight difference of the plastic waste storage component before and after containing the compressed second waste is determined by a weighing component located below the plastic waste storage component, and is taken as the gross weight of the plastic waste. Determine the moisture content factor corresponding to the plastic waste collection component; The gross weight of the plastic waste is corrected based on the moisture content factor to obtain the corrected weight of the plastic waste. A return value is generated based on the weight of the plastic waste and the corresponding reward factor; Based on the returned value, a reward will be returned to the account that initiated the recycling of household waste targeting the waste.

6. An electronic device, characterized in that, include: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 5.

7. A computer-readable medium, characterized in that, It stores a computer program thereon, wherein the computer program, when executed by a processor, implements the method as described in any one of claims 1 to 5.

Images