A Smart Waste Sorting, Collection, and Transportation Monitoring Method Based on Video Edge Computing

Abstract

This invention relates to the field of waste sorting monitoring and analysis, specifically disclosing a smart waste sorting and collection monitoring method based on video edge computing. Through multiple sensors and ultrasonic detection technology, it achieves intelligent monitoring and control of the entire waste sorting and collection process. Specifically, it includes: automatically activating a supplementary lighting device based on ambient brightness to ensure image acquisition quality; identifying and matching waste bin categories through color and RFID tags; adaptively adjusting the waste bin tilting angular velocity based on waste volume and dumping rate requirements; using penetration detection to identify bagged waste and controlling the bag-breaking teeth for precise bag breaking; and combining item names and materials to statistically identify accurately classified, misclassified, and unidentifiable waste components, ultimately generating a waste disposal assessment form. This invention, through multi-sensor data fusion and edge computing technology, significantly improves the accuracy and efficiency of waste sorting identification, reduces manual intervention costs, and provides intelligent decision support for waste collection and management.

Description

Technical Field

[0001] This invention relates to the field of waste sorting monitoring and analysis, and specifically to a smart waste sorting and collection monitoring method based on video edge computing. Background Technology

[0002] Waste management is crucial for sustainable urban development. Accelerated urbanization has led to a surge in household waste, making efficient waste sorting and collection imperative. Establishing a comprehensive management system covering the entire process from waste collection and transportation to final disposal is essential for achieving effective waste sorting.

[0003] The current implementation of household waste sorting faces numerous challenges in the supervision phase. The number of residential communities implementing waste sorting has surged, making it difficult for regulatory departments to maintain continuous and comprehensive oversight. As residents develop sorting habits, there is an urgent need for precise monitoring of the accuracy of source separation, to assist service companies in improving the quality of their education and assistance in sorting, and to enhance residents' awareness of waste sorting.

[0004] However, existing waste sorting monitoring and management methods still have some limitations and shortcomings in practical applications.

[0005] For example, Chinese Patent CN115914577A discloses a lightweight automatic image acquisition system for classified waste, including a waste information collection terminal, a data transmission decision system, and a cloud server. The waste information collection terminal identifies the waste disposal behavior from infrared ranging data using a disposal detection algorithm and then captures an image of the waste. The data transmission decision system determines the appropriate transmission method based on the terminal's current network environment and transmits the waste image to the cloud. The cloud server performs image recognition processing based on the received waste image information to determine whether the disposal is satisfactory. This invention can automatically detect waste disposal behavior in waste collection stations, acquire and transmit images of the disposed waste, and the cloud server judges the disposal quality.

[0006] For example, the existing Chinese patent with publication number CN118354035A discloses an intelligent waste sorting and monitoring method, system, and product based on image recognition. This method includes determining the location information of the waste when a user disposes of waste, automatically adjusting the shooting angle according to the location information so that the image acquired at that shooting angle contains more relevant information about the waste, and then determining the shooting focal length based on the proportion of the waste area in the image at the shooting angle, so that the disposal image obtained at the determined shooting angle and shooting focal length can have a clearer and more complete image, thus improving the image quality. When performing classification and recognition based on the disposal image, it can accurately evaluate the disposal operation and provide targeted prompts, thereby improving the classification effect.

[0007] The shortcomings of the above patents are: 1. They are only applicable to well-equipped garbage stations or garbage rooms and not to outdoor garbage bins without video capture equipment; they are only applicable to situations where a single user disposes of garbage once and not to situations where garbage accumulates in the bins when garbage trucks collect garbage, thus limiting the application scenarios.

[0008] 2. The impact of changes in lighting on image quality was not taken into account, which resulted in insufficient reliability of the accuracy assessment results of image recognition-based waste sorting.

[0009] 3. It did not take into account the possibility of unopened garbage bags inside the trash can, whose contents are not visible, making it impossible to collect images of the internal garbage and perform detection and identification. Summary of the Invention

[0010] To address the above problems, this invention proposes a smart waste sorting and collection monitoring method based on video edge computing. The specific technical solution is as follows: The smart waste sorting and collection monitoring method based on video edge computing includes the following steps: S1: The supplementary lighting device is automatically activated based on the monitoring results of the ambient brightness at the waste storage point and the feeding compartment of the sorted waste truck.

[0011] S2: Based on the images of each garbage bin in the garbage storage point collected by the on-board video equipment of the sorting garbage truck, the category of each garbage bin is identified based on color and RFID tags, and then the corresponding garbage bin of the sorting garbage truck is matched.

[0012] S3: Based on the current volume of garbage in the bins and the requirements of video capture for the garbage dumping rate, analyze the angular velocity of the sorting garbage truck dumping the garbage bins and adaptively adjust the angular velocity according to the actual dumping situation.

[0013] S4: Identify bagged garbage in the trash can and control the bag-breaking teeth on the stern plate of the sorting garbage truck to break the bags.

[0014] S5: Based on the collected image data and material information of the garbage, identify and analyze the composition of the garbage in the garbage bin, and count the accurately classified, misclassified, and unidentifiable garbage.

[0015] S6: Based on the identification results of the waste composition in the trash can, generate an assessment form for waste disposal at the waste storage point.

[0016] Compared with existing technologies, the intelligent waste sorting and collection monitoring method based on video edge computing described in this invention has the following advantages: 1. Wide range of application scenarios: This invention, by installing a series of sensors, including video acquisition equipment, on sorted garbage trucks, enables the assessment of the accuracy of waste sorting during the waste collection and transportation process. It is applicable to various types of waste storage points and sorted garbage trucks as well as different environmental conditions, and has strong practicality and promotion value.

[0017] 2. Intelligent supplementary lighting and environmental adaptability: This invention monitors the ambient brightness of the garbage storage point and the garbage truck's feeding compartment, and automatically starts and adjusts the supplementary lighting device to ensure the clarity of video capture, avoid the impact of insufficient light on the accuracy of garbage classification and identification, and meet the video capture needs of different scenarios.

[0018] 3. Adaptive tipping control: This invention dynamically adjusts the tipping angular velocity of the trash can based on the volume of trash inside the can and the requirements of video capture for the tipping rate. This balances the tipping efficiency with the video capture quality, avoiding problems caused by tipping too fast or too slow, and ensuring a stable and efficient tipping process.

[0019] 4. Efficient bag breaking process: This invention uses ultrasonic detection technology to identify bagged waste and uses a bag-breaking tooth array to automatically break the bags and release the contents of the waste, which facilitates the identification and statistics of waste classification. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0022] Figure 2 This is a schematic diagram of the garbage truck used in this invention.

[0023] Figure 3 This is a schematic diagram of the feeding chamber of the present invention.

[0024] Figure 4 This is a schematic diagram of the bag-breaking tooth array on the stern plate of the present invention.

[0025] Figure 5 This is a schematic diagram of the bag-breaking teeth with different spacing according to the present invention.

[0026] Reference numerals: 1. Sorting garbage truck; 2. Lifting mechanism; 3. Feeding compartment; 4. Ship-shaped stern bottom plate. Detailed Implementation

[0027] 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.

[0028] Please see Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 As shown, the intelligent collection and monitoring method for waste sorting based on video edge computing provided by the present invention includes the following steps: S1: The supplementary lighting device is automatically activated based on the monitoring results of the ambient brightness of the waste storage point and the feeding compartment of the sorted waste truck.

[0029] As a preferred embodiment, the specific analysis process of step S1 is as follows: the ambient brightness of the garbage storage point and the garbage truck's feeding compartment is monitored by an ambient brightness detection device installed on the garbage truck body and inside the feeding compartment.

[0030] Based on the ambient brightness requirements for video capture at waste storage points and in feeding bins, lower limits for ambient brightness are set for both.

[0031] If the ambient brightness at the garbage storage point or inside the feeding compartment of the sorting garbage truck is less than the corresponding lower limit of ambient brightness, the corresponding supplementary lighting device in the sorting garbage truck will be activated, and the brightness of the supplementary lighting device will be adjusted according to the difference between the actual ambient brightness and the lower limit of ambient brightness.

[0032] It should be noted that there are multiple ambient light detection devices installed on the sorting garbage truck, and they are distributed in different locations on the sorting garbage truck according to their different detection targets. For example, the ambient light detection device used to detect the ambient light at the garbage storage point is installed on the body of the sorting garbage truck (including the front and rear of the truck), and the ambient light detection device used to detect the ambient light inside the sorting garbage truck's feeding compartment is installed inside the sorting garbage truck's feeding compartment.

[0033] It should be noted that when the sorting garbage truck is collecting garbage from the garbage storage point, it stops next to the garbage storage point. At this time, the ambient brightness detected at the garbage truck body is taken as the ambient brightness of the garbage storage point.

[0034] It should be noted that insufficient ambient light will affect the quality of the acquired video images, and consequently affect the results of video image recognition and analysis.

[0035] It should be noted that garbage storage sites are typically relatively large outdoor or semi-outdoor spaces, requiring sufficient lighting to clearly capture video footage and identify the type and storage condition of the garbage, as well as any abnormalities in the surrounding environment. Simultaneously, natural lighting conditions at different times of day, and potential obstructions, must be considered. Generally, the video capture equipment should be able to obtain clear images even under poor lighting conditions such as cloudy days, evenings, or early mornings. Lower limits can be determined by conducting on-site measurements under different lighting conditions in the area, combined with the performance parameters of the video capture equipment.

[0036] In one specific embodiment, multiple measurements revealed that when the ambient brightness was below 100 lux, noticeable noise began to appear in the video image, and object outlines became blurred, affecting the identification of garbage. Therefore, to ensure the quality of video acquisition, the lower limit of the ambient brightness at the garbage storage point was set to 150 lux, ensuring clear and accurate video acquisition in most daily situations.

[0037] It should be noted that the space inside the garbage truck's feeding compartment is relatively small and enclosed, requiring accurate video identification of the garbage during the disposal process. This can be achieved by installing brightness sensors at different locations and angles within the feeding compartment to simulate various garbage disposal scenarios and measure the minimum brightness value required for the video capture equipment to clearly identify the garbage type and the disposal action.

[0038] In one specific embodiment, testing revealed that when the ambient light level inside the bin is below 50 lux, video recognition may produce errors or omissions for some dark-colored waste or waste disposed of at complex angles. Therefore, to ensure the accuracy of video capture inside the bin of the sorting garbage truck, the lower limit of ambient light can be set to 80 lux. This ensures accurate monitoring and recording of the disposal process under various common waste disposal conditions, facilitating subsequent sorting statistics and supervision.

[0039] It should be noted that there are multiple supplementary lighting devices installed on the sorting garbage truck, and they are distributed in different locations on the sorting garbage truck according to their different areas of operation. For example, supplementary lighting devices used to supplement the environment of the garbage storage point are installed on the body of the sorting garbage truck, and supplementary lighting devices used to supplement the environment inside the feeding compartment of the sorting garbage truck are installed inside the feeding compartment of the sorting garbage truck.

[0040] It should be noted that a feasible simulation process, as shown in Tables 1 and 2, is as follows: based on the difference between the ambient brightness of the garbage storage point and its lower limit value, and combined with a preset table showing the relationship between this difference and the brightness of the corresponding supplementary lighting device at the garbage storage point, the brightness of the supplementary lighting device at the garbage storage point is matched to obtain the brightness of the supplementary lighting device at the garbage storage point.

[0041] Based on the difference between the ambient brightness inside the feeding compartment of the classified garbage truck and its lower limit value, and combined with the preset relationship table between this difference and the brightness of the corresponding supplementary lighting device in the feeding compartment, the brightness of the corresponding supplementary lighting device in the feeding compartment is matched and obtained.

[0042] Table 1. Rules for Supplemental Lighting at Garbage Storage Points

[0043]

[0044]

[0045] Table 2. Rules for Lighting the Feeding Chamber

[0046] Actual ambient brightness (lux) Brightness difference (lower limit 80 - actual brightness) Complementary light brightness (lux) 80 0 0 70 10 50 60 20 100 50 30 150 40 40 200 30 50 250 20 60 300 10 70 350 0 80 400

[0047] In this embodiment, the present invention monitors the ambient brightness of the garbage storage point and the garbage truck's feeding compartment, and automatically starts and adjusts the supplementary lighting device to ensure the clarity of video acquisition, avoid the impact of insufficient light on the accuracy of garbage classification and identification, and meet the video acquisition needs in different scenarios.

[0048] S2: Based on the images of each garbage bin in the garbage storage point collected by the on-board video equipment of the sorting garbage truck, the category of each garbage bin is identified based on color and RFID tags, and then the corresponding garbage bin of the sorting garbage truck is matched.

[0049] As a preferred embodiment, the specific analysis process of step S2 includes: extracting images of each garbage bin within the garbage storage point based on the video collected by the onboard video equipment of the garbage truck.

[0050] Image processing technology is used to obtain the grayscale value of the body of each trash can and the RFID tag in the image. The grayscale value of the body of each trash can is compared with the grayscale value range corresponding to various trash can colors stored in the database to obtain the color of each trash can.

[0051] It should be noted that there can be multiple onboard video devices for sorting garbage trucks, deployed in different locations within the truck. These onboard video devices utilize self-cleaning cameras (such as lenses with hydrophobic coatings) to reduce interference from dirt.

[0052] It should be noted that, in one specific embodiment, the grayscale value range corresponding to the various trash can colors is shown in Table 3.

[0053] Table 3. Grayscale Value Range Corresponding to Trash Can Colors

[0054] Trash can categories Trash can color grayscale range Hazardous waste red 0-10 recyclables blue 10-40 Kitchen waste (wet waste) green 41-70 Other waste (dry waste) grey 71-180

[0055] As a preferred embodiment, the specific analysis process of step S2 further includes: identifying the category of each trash can in the garbage storage point based on the color and RFID tag characteristics of various types of trash cans stored in the database, and then matching the corresponding trash can to the sorting garbage truck by combining the trash can categories applicable to the sorting garbage truck.

[0056] It should be noted that the feeding compartment of a garbage sorting truck refers to the open compartment on the top or side of the garbage sorting truck used to receive garbage. The lifting mechanism of the garbage sorting truck pours the garbage from the garbage bins into this compartment. The lifting mechanism is a device used to hang the garbage bins and automatically dump the garbage.

[0057] S3: Based on the current volume of garbage in the bins and the requirements of video capture for the garbage dumping rate, analyze the angular velocity of the sorting garbage truck dumping the garbage bins and adaptively adjust the angular velocity according to the actual dumping situation.

[0058] As a preferred embodiment, the specific analysis process of step S3 includes: analyzing the ratio between the volume of garbage inside the garbage bin and the bin's capacity based on the image of the garbage bin at the garbage storage point, to obtain the current volume of garbage in the garbage bin.

[0059] It should be noted that when the video images of the trash cans at the garbage storage point were collected, the lids of the trash cans were open.

[0060] The requirements for garbage dumping rate based on video capture from the onboard video equipment of classified garbage trucks.

[0061] It should be noted that the garbage dumping rate refers to the amount (mass or volume) of garbage dumped per unit time.

[0062] It should be noted that if the garbage dumping rate is too fast, the video equipment will not be able to capture enough details, while if the garbage dumping rate is too slow, the efficiency of garbage collection will be reduced.

[0063] It should be noted that the requirements for the garbage dumping rate of the onboard video equipment of the sorting garbage truck depend on key parameters such as the frame rate, resolution, shutter speed, and dynamic range of the video equipment. Different garbage dumping rates can be used to simulate dumping and verify the video acquisition quality and analyze errors, thereby determining the garbage dumping rate that meets the requirements. In one specific embodiment, the required garbage dumping rate for the onboard video equipment of the sorting garbage truck is 0.1–0.5 cubic meters per second.

[0064] Extract the correspondence between the volume of trash in the trash can, the tilting angular velocity, and the trash tilting rate stored in the database.

[0065] Based on the current volume of garbage in the bin and the requirements of the video capture for the garbage dumping rate, the angular velocity of the sorting garbage truck dumping the garbage bin is matched and sent to the execution terminal of the sorting garbage truck dumping the garbage bin.

[0066] It should be noted that when a garbage truck emptys its bins, the change in the angle between the bin's central axis and the horizontal line per unit time is recorded as the tilting angular velocity. The range of the angle between the bin's central axis and the horizontal line during the emptying process is as follows:

[0067] It should be noted that the correspondence between the volume of trash in the bin, the tilting angular velocity, and the trash dumping rate is based on practical or experimental testing. For example, in a controlled environment, a trash can dumping experiment is conducted, using sensors to simultaneously measure the volume of trash, the tilting angular velocity, and the trash dumping rate. A mathematical model is established using methods such as multinomial regression and machine learning, with angular velocity and volume as input variables and dumping rate as output variable, to analyze the correspondence between the volume of trash in the bin, the tilting angular velocity, and the trash dumping rate. In a specific embodiment, refer to Table 4, where some data were obtained experimentally (total trash can volume is 100L).

[0068] Table 4. Correspondence between waste volume, dumping angular velocity, and waste dumping rate

[0069] Waste volume (%) Tilting angular velocity (° / s) Waste dumping rate (L / s) 100％ 0 0.0 95％ 5 5.0 85％ 10 10.0 70％ 15 15.0 50％ 20 20.0 25％ 25 25.0 0％ 30 0.0

[0070] Data Explanation: The dumping angular velocity is positively correlated with the dumping rate; the greater the angular velocity, the faster the waste flows out.

[0071] It should be noted that the final stage of the sorting garbage truck's operation, where it emptys the garbage bins, is the lifting mechanism of the sorting garbage truck.

[0072] As a preferred embodiment, the specific analysis process of step S3 further includes: obtaining the actual garbage dumping rate of the garbage bin by real-time detection of the amount of garbage dumped during the garbage dumping process of the sorting garbage truck, determining whether the actual garbage dumping rate meets the requirements, and if it does not meet the requirements, obtaining the relative deviation of the actual garbage dumping rate and combining it with the preset relationship model between the relative deviation of the garbage dumping rate and the garbage bin dumping angular velocity correction amount, obtaining the correction amount of the garbage bin dumping angular velocity, and then performing adaptive adjustment.

[0073] It should be noted that the specific process for determining whether the actual garbage dumping rate meets the requirements is as follows: compare the actual garbage dumping rate of the garbage bin with the garbage dumping rate requirement set by the video capture, and calculate the relative deviation using the formula... The relative deviation of the actual garbage dumping rate of the garbage bin is obtained. If the relative deviation of the garbage dumping rate exceeds the set allowable range, then the actual garbage dumping rate of the garbage bin does not meet the requirements.

[0074] It should be noted that the specific method for obtaining the correction amount of the trash can's tilting angular velocity is as follows: Substitute the relative deviation of the actual trash can's tilting rate into a preset relationship model between the relative deviation of the trash can's tilting rate and the correction amount of the trash can's tilting angular velocity to obtain the correction amount of the trash can's tilting angular velocity. The relationship model includes a quantitative mapping relationship between the relative deviation of the trash can's tilting rate and the correction amount of the trash can's tilting angular velocity. In a specific embodiment, the quantitative mapping relationship between the relative deviation of the trash can's tilting rate and the correction amount of the trash can's tilting angular velocity is shown in Table 5.

[0075] Table 5. Relationship between relative deviation of garbage dumping rate and correction amount of garbage bin dumping angular velocity

[0076]

[0077] In this embodiment, the present invention dynamically adjusts the tilting angular velocity of the trash can according to the volume of trash in the trash can and the requirements of video capture for the trash dumping rate. This balances the trash can dumping efficiency with the video capture quality, avoids problems caused by dumping too fast or too slow, and ensures a stable and efficient dumping process.

[0078] S4: Identify bagged garbage in the trash can and control the bag-breaking teeth on the stern plate of the sorting garbage truck to break the bags.

[0079] As a preferred embodiment, the specific analysis process of step S4 includes: during the process of the sorting garbage truck dumping garbage bins, ultrasonic detection technology is used to perform penetration detection on the garbage inside the garbage bins, identify the bagged garbage inside the garbage bins and the volume of the bagged garbage, and if bagged garbage is detected, the bag-breaking tooth array provided on the boat-shaped stern bottom plate in the sorting garbage truck's feeding compartment is triggered.

[0080] It's important to note that the principle behind ultrasonic detection of bagged waste is primarily based on the propagation characteristics of ultrasound waves in different media, as well as the phenomena of reflection, refraction, and scattering. Bags (usually made of plastic or paper) and waste (such as various solid wastes and organic matter) are two different media, and the propagation speed and attenuation coefficient of ultrasound waves differ between them. When ultrasound waves propagate through bagged waste, these differences cause changes in the ultrasonic signal, thus providing a basis for detection.

[0081] It should be noted that the bag-breaking teeth are used to puncture the garbage bag inside the trash can to release its contents.

[0082] As a preferred embodiment, the specific analysis process of step S4 further includes: based on the volume distribution of bagged garbage in the trash can, selecting the spacing between the bag-breaking teeth, the sliding speed of the bag-breaking teeth, and the penetration depth of the bag-breaking teeth in the bag-breaking tooth array according to the set rules, generating control commands for the bag-breaking tooth array, and sending them to the execution end of the bag-breaking tooth array.

[0083] In one specific embodiment, the volume distribution of bagged waste in the trash can includes a majority of large-volume, medium-volume, and small-volume wastes, with the volume classification based on a set standard.

[0084] It should be noted that the larger the volume of bagged waste, the larger the spacing between the tearing teeth, the faster the sliding speed of the tearing teeth, and the deeper the penetration depth of the tearing teeth; the smaller the volume of bagged waste, the smaller the spacing between the tearing teeth, the slower the sliding speed of the tearing teeth, and the shallower the penetration depth of the tearing teeth.

[0085] It should be noted that a feasible simulation process generates control parameters for the bag-breaking tooth array based on the volume distribution of bagged waste inside the trash can, as detailed in Table 6.

[0086] Table 6. Control Table for Bag-Breaking Tooth Array Parameters

[0087] Waste volume distribution Bag-breaking tooth spacing Sliding speed of the bag-breaking teeth Bag-breaking tooth penetration depth Mostly large in size 120mm 80mm / s 50mm Mostly medium-sized 80mm 100mm / s 35mm Most are small in size 50mm 120mm / s 20mm

[0088] In this embodiment, the present invention uses ultrasonic detection technology to identify bagged waste, and uses a bag-breaking tooth array to automatically break the bag and release the contents of the waste, which facilitates the identification and statistics of waste classification.

[0089] S5: Based on the collected image data and material information of the garbage, identify and analyze the composition of the garbage in the garbage bin, and count the accurately classified, misclassified, and unidentifiable garbage.

[0090] As a preferred embodiment, the specific analysis process of step S5 is as follows: by collecting video of the garbage bin dumping process through the on-board video equipment of the sorting garbage truck, the actual images of each piece of garbage in the garbage bin are obtained and compared with the item sample image library stored in the database. The item name of each piece of garbage in the garbage bin is obtained by using image recognition technology. Garbage that does not match the actual image and the item sample image library is recorded as unidentifiable garbage and the images of unidentifiable garbage are stored.

[0091] It should be noted that images of unidentifiable waste will be sent to the intelligent waste sorting and collection platform for manual identification, thereby supplementing and updating the image database of sample items.

[0092] The material type of each piece of garbage in the bins is detected by the infrared sensors on the garbage truck.

[0093] Based on the item name and material type of each piece of trash in the trash can, and combined with the preset item-material-classification map, the classification category of each piece of trash in the trash can is obtained, and trash whose classification category matches the trash can category and trash whose classification category does not match the trash can category are respectively recorded as correctly classified trash and incorrectly classified trash.

[0094] It should be noted that the video equipment for collecting images of garbage during the dumping process can be installed inside the feeding compartment and at the entrance of the sorting garbage truck.

[0095] It should be noted that infrared sensors distinguish between metal, plastic, and glass materials based on infrared reflectivity.

[0096] It should be noted that one item in the constructed item-material-category map is: milk tea cup → plastic / paper → recyclable.

[0097] S6: Based on the identification results of the waste composition in the trash can, generate an assessment form for waste disposal at the waste storage point.

[0098] As a preferred option, the specific analysis process of step S6 is as follows: based on the proportion of accurately classified, incorrectly classified, and unidentifiable garbage in the trash can, the accuracy coefficient of garbage disposal is evaluated.

[0099] An assessment table for waste disposal at waste collection points is generated based on the proportion of accurately sorted, incorrectly sorted, and unidentifiable waste in the waste bins, the proportion of other types of waste among the incorrectly sorted waste, and the accuracy coefficient of waste disposal.

[0100] It should be noted that the accuracy coefficient is a comprehensive indicator used to quantify the accuracy of waste disposal. A feasible simulation process can be calculated through the following steps: Statistically calculate the percentage of waste in the bins according to the three categories: accurately classified percentage (A%), incorrectly classified percentage (E%), and unidentifiable waste percentage (U%). Calculate the accuracy coefficient using the formula: Accuracy Coefficient = A% - α × E% - β × U%, where α and β represent the penalty factors for incorrect classification and unidentifiable waste, respectively. These penalty factors can be adjusted according to actual needs. In a specific embodiment, the penalty factor for incorrect classification is 0.5, and the penalty factor for unidentifiable waste is 0.2.

[0101] It should be noted that a feasible simulation process generates an assessment form for waste disposal at waste storage points, as detailed in Table 7.

[0102] Table 7. Waste Disposal Accuracy Assessment Table

[0103]

[0104] In this embodiment, the present invention combines image recognition and infrared sensor technology to accurately identify the item name and material type of waste, thereby distinguishing between accurately classified, misclassified, and unidentifiable waste, and improving the accuracy of identification through two-factor authentication.

[0105] In this embodiment, the present invention achieves the assessment of the accuracy of garbage sorting during the garbage collection process by installing a series of sensors, including video acquisition equipment, on the sorting garbage truck. It is applicable to various garbage storage points and sorting garbage trucks as well as different environmental conditions, and has strong practicality and promotion value.

[0106] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for intelligent collection and transportation monitoring of garbage classification based on video edge computing, characterized in that, Comprise the following steps: S1: according to the garbage storage point and the monitoring result of the environment brightness in the classified garbage truck feeding cabin, the light supplement device is started automatically; S2: according to the image of each garbage can in the garbage storage point collected by the video equipment carried by the classified garbage truck, the category of each garbage can is identified based on color and RFID tag, and the corresponding garbage can of the classified garbage truck is matched; S3: according to the current garbage volume of the garbage can and the requirement of video collection on garbage dumping rate, the angular velocity of the classified garbage truck dumping the garbage can is analyzed, and the angular velocity is adaptively adjusted according to the actual dumping condition; The specific analysis process of step S3 comprises: According to the image analysis of the garbage can in the garbage storage point, the ratio between the garbage volume in the garbage can and the can volume is obtained, and the current garbage volume of the garbage can is obtained; The requirement of video collection on garbage dumping rate is obtained; The corresponding relationship between the garbage volume of the garbage can, the dumping angular velocity and the garbage dumping rate stored in the database is extracted; According to the current garbage volume of the garbage can and the requirement of video collection on garbage dumping rate, the angular velocity of the classified garbage truck dumping the garbage can is matched and sent to the execution end of the classified garbage truck dumping the garbage can; The specific analysis process of step S3 further comprises: By detecting the amount of garbage poured out during the process of the classified garbage truck dumping the garbage can, the actual garbage dumping rate of the garbage can is obtained, whether the actual garbage dumping rate meets the requirement is judged, if not, the relative deviation of the actual garbage dumping rate is obtained, and a relationship model between the preset relative deviation of the garbage dumping rate and the correction amount of the garbage can dumping angular velocity is combined to obtain the correction amount of the garbage can dumping angular velocity and then adaptively adjust it; S4: identifying the bagged garbage in the garbage can and controlling the bag breaking teeth provided on the ship type tail bottom plate in the feeding cabin of the classified garbage truck to break the bag; S5: identifying and analyzing the garbage composition in the garbage can according to the image data and material information of the collected garbage, and counting the accurately classified, misclassified and unidentifiable garbage; S6: according to the garbage composition identification result in the garbage can, an evaluation table of garbage disposal in the garbage storage point is generated. 2.The method of claim 1, wherein the method further comprises: The specific analysis process of step S1 is: The environment brightness detection device installed in the classified garbage truck body and the feeding cabin respectively monitors the environment brightness of the garbage storage point and the feeding cabin of the classified garbage truck; According to the requirement of video collection on environment brightness in the garbage storage point and the feeding cabin, the lower limit value of the environment brightness of the two is set respectively; If the environment brightness of the garbage storage point or the environment brightness in the feeding cabin of the classified garbage truck is less than the corresponding lower limit value of the environment brightness, the corresponding light supplement device in the classified garbage truck is started, and the brightness of the light supplement device is regulated according to the difference between the actual environment brightness and the lower limit of the environment brightness. 3.The method of claim 1, wherein the method further comprises: The specific analysis process of step S2 comprises: According to the video collected by the video equipment carried by the classified garbage truck, the image of each garbage can in the garbage storage point is extracted; The gray value of the barrel body part of each garbage can is obtained by using image processing technology, and the gray value of the barrel body part of each garbage can is compared with the gray value range corresponding to the color of each garbage can stored in the database, and the color of each garbage can is matched.

4. The intelligent collection and transportation monitoring method for garbage classification based on video edge computing according to claim 3, characterized in that: The specific analysis process of step S2 further includes: Based on the color and RFID tag characteristics of each type of garbage can stored in the database, the category of each garbage can in the garbage storage point is identified, and the category of the garbage can suitable for the sorting garbage truck is combined, and then the corresponding garbage can of the sorting garbage truck is matched. 5.The method of claim 1, wherein the method further comprises: The specific analysis process of step S4 includes: During the dumping process of the sorting garbage truck, the ultrasonic detection technology is used to detect the garbage in the garbage can, identify the bagged garbage in the garbage can and the volume of the bagged garbage, and if the bagged garbage is detected, the array of bag breaking teeth provided on the tail bottom plate of the ship-shaped of the feeding cabin of the sorting garbage truck is triggered. 6.The method of claim 5, wherein the method further comprises: The specific analysis process of step S4 further includes: According to the volume distribution of the bagged garbage in the garbage can, the distance between the bag breaking teeth, the sliding speed of the bag breaking teeth and the penetration depth of the bag breaking teeth in the array of bag breaking teeth are selected according to the set rules, the control instructions of the array of bag breaking teeth are generated, and are sent to the execution end of the array of bag breaking teeth. 7.The method of claim 1, wherein the method further comprises: The specific analysis process of step S5 is: Through the video of the garbage can dumping process collected by the video equipment on the sorting garbage truck, the real image of each garbage in the garbage can is obtained and compared with the sample image library stored in the database, and the image recognition technology is used to obtain the name of each garbage in the garbage can, wherein the garbage that is not matched with the sample image library is recorded as unidentifiable garbage, and the unidentifiable garbage image is stored; The material type of each garbage in the garbage can is detected by the infrared sensor on the sorting garbage truck; According to the name and material type of each garbage in the garbage can, the classification category of each garbage in the garbage can is obtained by combining the preset material-material-classification category map, and the garbage with the same and different classification categories as the garbage can is recorded as accurately classified garbage and misclassified garbage respectively. 8.The method of claim 1, wherein the method further comprises: The specific analysis process of step S6 is: According to the proportion of accurately classified, misclassified and unidentifiable garbage in the garbage can, the precision coefficient of garbage disposal is evaluated; According to the proportion of accurately classified, misclassified and unidentifiable garbage in the garbage can, the proportion of other types of garbage in the misclassified garbage, and the precision coefficient of garbage disposal, an evaluation table of garbage disposal of the garbage storage point is generated.

Images