Ai-based waste disposal handover management system and method and computer program for the same

Abstract

A waste discharge transfer management system comprises: an image receiving unit configured to receive operation data of a waste transport vehicle and extract hourly weight information of the waste transport vehicle in the form of an image from the operation data; a machine learning unit configured to support the use of a pre-trained image analysis model using a training image representing hourly weight information; and a loading / unloading detection unit configured to detect one or more of loading and unloading of waste on the waste transport vehicle by using an image extracted from the operation image as an input value for the image analysis model. In this case, the loading / unloading detection unit may determine whether there is an anomaly in the hourly weight information represented by the image extracted from the operation image based on whether waste is loaded or unloaded on the waste transport vehicle, using the analysis model.

Description

Technology Field

[0001] The embodiments relate to an AI (Artificial Intelligence)-based waste discharge transfer management system and method, and a computer program for the same. More specifically, the embodiments relate to a technology that receives location and weight information related to a vehicle transporting waste, such as livestock manure, from a discharger to a processor in the form of images, and accurately determines the loading and unloading of waste and identifies abnormal discharge patterns through an AI-based image classification and / or object detection algorithm. Background Technology

[0002] The proper discharge and treatment of livestock manure is a critical task in terms of environmental protection and hygiene management, and electronic transfer management systems have been introduced as a technology to manage this efficiently.

[0003] Conventional electronic livestock manure transfer management systems operate by digitizing, recording, and managing the transfer process among users consisting of dischargers, transporters, and processors. Specifically, when a discharger discharges livestock manure, they register the relevant information in the system in the form of an electronic transfer document; transporters input information regarding the transportation process; and processors register the details of the final treatment, thereby enabling the tracking of the entire discharge and treatment process.

[0004] Korean Registered Patent Publication No. 10-2404051 is one of the prior art for achieving the above purpose. Registered Patent Publication No. 10-2404051 discloses that an electronic transfer automatic management server collects location information measured by a GPS module of a manure transport vehicle and weight information measured by a weight sensor, and generates an electronic transfer document for manure by determining the loading and unloading of the manure transport vehicle based on this information.

[0005] However, the conventional system disclosed in Registered Patent Publication No. 10-2404051 determines loading and unloading based solely on the weight measured by a weight sensor mounted on the leaf spring of a manure transport vehicle. Since manure transport vehicles frequently operate on unpaved roads, leaf spring failures are common; consequently, there is a problem in that the accuracy of determining manure loading and unloading is low based only on the measurements from the weight sensor mounted on the leaf spring. Prior art literature

[0006] Registered Patent Publication No. 10-2404051 The problem to be solved

[0007] According to one aspect of the present invention for solving the problems of the aforementioned prior art, a waste discharge transfer management system and method and a computer program for the same can be provided, which can accurately determine the loading and unloading of waste and identify abnormal discharge patterns by receiving location and weight information related to a vehicle transporting waste, such as livestock manure, from a discharger to a processor in the form of an image, and by applying AI-based image classification and object detection algorithms such as a Convolutional Neural Network (CNN) and YOLO (You Only Look Once). means of solving the problem

[0008] A waste discharge transfer management system according to one aspect of the present invention comprises: an image receiving unit configured to receive operation data of a waste transport vehicle and extract hourly weight information of the waste transport vehicle in the form of an image from the operation data; a machine learning unit configured to support the use of a pre-trained image analysis model using a training image representing hourly weight information; and a loading / unloading detection unit configured to detect one or more of loading and unloading of waste for the waste transport vehicle by using an image extracted from the operation image as an input value for the image analysis model.

[0009] At this time, the loading / unloading detection unit is further configured to determine, using the analysis model, whether there is an anomaly in the hourly weight information represented by the image extracted from the operation image, based on whether waste is loaded or unloaded on the waste transport vehicle.

[0010] In one embodiment, the abnormality of the hourly weight information includes whether the buffer mechanism equipped in the waste transport vehicle is malfunctioning. For example, the buffer mechanism may be a leaf spring. In this case, the image analysis model may be pre-learned using a training image in which the malfunction of the buffer mechanism is labeled.

[0011] In one embodiment, the machine learning unit is further configured to support the use of a pre-trained object detection model to detect waste discharge transfer documents from hourly weight information corresponding to one or more waste transport vehicles. In this case, the waste discharge transfer management system further includes a discharge transfer document detection unit configured to detect discharge transfer documents corresponding to the operation of the waste transport vehicle by using an image extracted from the operation image as an input value for the object detection model.

[0012] In one embodiment, the discharge transfer document detection unit is further configured to determine whether there is a discharge abnormality corresponding to the waste transport vehicle or the transport company operating the waste transport vehicle, based on one or more of the number of discharge transfer documents detected from the operation data of the waste transport vehicle, location information where the discharge transfer documents were detected, and weather information at the time when the discharge transfer documents were detected.

[0013] A waste discharge transfer management method according to one aspect of the present invention is performed by a computing device.

[0014] In one embodiment, the waste discharge transfer management method comprises: a step in which the computing device receives operation data of a waste transport vehicle; a step in which the computing device extracts hourly weight information of the waste transport vehicle in the form of an image from the operation data; a step in which the computing device detects one or more of loading and unloading of waste for the waste transport vehicle by inputting the image extracted from the operation image into a pre-trained image analysis model using a training image representing hourly weight information; and a step in which the computing device determines, using the analysis model, whether there is an anomaly in the hourly weight information represented by the image extracted from the operation image based on whether waste is loaded or unloaded for the waste transport vehicle.

[0015] A waste discharge transfer management method according to one embodiment further includes the step of detecting a discharge transfer document corresponding to the operation of the waste transport vehicle by inputting an image extracted from the operation image into an object detection model that has been pre-trained to detect a waste discharge transfer document from hourly weight information corresponding to one or more waste transport vehicles, wherein the computing device detects a discharge transfer document corresponding to the operation of the waste transport vehicle.

[0016] In one embodiment, the step of detecting the discharge transfer document includes the step of the computing device determining whether there is a discharge abnormality corresponding to the waste transport vehicle or the transporter operating the waste transport vehicle based on one or more of the number of discharge transfer documents detected from the operation data of the waste transport vehicle, location information where the discharge transfer document was detected, and weather information at the time when the discharge transfer document was detected.

[0017] A computer program according to one aspect of the present invention may be stored on a computer-readable recording medium for implementing a waste discharge transfer management method according to embodiments of the present invention in combination with hardware. Effects of the invention

[0018] According to an AI-based waste discharge transfer management system and method in one aspect of the present invention, the problem of reduced accuracy in determining discharge status due to leaf spring failure occurring in conventional waste discharge transfer management systems is resolved, and by applying an AI-based image classification model such as CNN, vehicles with failures are excluded from the discharge status determination process, thereby providing the advantage of more reliably measuring weight changes of transport vehicles.

[0019] In addition, according to one aspect, if there are vehicles with abnormal weight measurements based on AI-based analysis results, notification is provided to the relevant agency to ensure that the vehicles operate normally, thereby improving maintenance efficiency and enhancing the accuracy of the waste emission determination model.

[0020] Furthermore, according to an AI-based waste discharge transfer management system and method according to one aspect of the present invention, by utilizing a discharge transfer document detection model based on object detection techniques such as YOLO, it is possible to effectively detect whether waste is discharged normally and / or the possibility of illegal dumping. Brief explanation of the drawing

[0021] FIG. 1 is a schematic block diagram showing the operating environment of a waste discharge transfer management system according to one embodiment. FIG. 2 is a schematic block diagram showing the hardware configuration of a waste discharge transfer management system according to one embodiment. FIG. 3 is a schematic block diagram of a waste discharge transfer management system and a waste transport vehicle communicating with it according to one embodiment. FIG. 4 is a flowchart showing each step of a waste discharge transfer management method according to one embodiment. FIG. 5 is a flowchart showing each step of the AI-based image classification process in a waste discharge transfer management method according to one embodiment. Figures 6 and 7 are diagrams showing examples of input images for AI-based image classification. FIG. 8 is a diagram illustrating the image classification results by a waste discharge transfer management method according to one embodiment. FIG. 9 is a flowchart showing each step of the process of detecting a waste discharge transfer document by a waste discharge transfer management method according to one embodiment. Figure 10 is a diagram showing an example of an input image for detecting the discharge transfer document. FIG. 11 is a diagram illustrating the detection results of a waste discharge transfer document by a waste discharge transfer management method according to one embodiment. Specific details for implementing the invention

[0022] Hereinafter, embodiments of the present invention will be examined in detail with reference to the drawings.

[0023] FIG. 1 is a schematic block diagram of an AI-based waste discharge transfer management system according to one embodiment.

[0024] As illustrated in FIG. 1, a waste discharge transfer management system (2) according to one embodiment may be configured to operate by interacting with one or more discharger terminals (110), one or more transporter terminals (120), one or more processor terminals (130), a machine learning service server (4) and / or one or more waste transport vehicles (3).

[0025] In this specification, "waste" refers to liquid and / or solid pollutants, such as livestock manure from livestock barns or liquid fertilizers, which must be discharged in accordance with the discharge transfer procedures required by administrative agencies, such as the Ministry of Environment or local governments, for environmental protection and lawful disposal, and for which a discharge transfer document must be issued and submitted upon the request of the administrative agency if necessary.

[0026] In the following specification, waste is described based on livestock manure, but the forms of waste to which the embodiments of the present invention can be applied are not limited thereto.

[0027] The discharger terminal (110) refers to a device carried by a discharger, for example, who discharges livestock manure or liquid fertilizer from a livestock barn, to communicate with a transporter terminal (120) for the discharge of livestock manure and to call a waste transport vehicle (3). For example, a discharger operating a livestock barn can use the discharger terminal (110) to call a transport vehicle (3) that matches the amount of livestock manure the discharger intends to discharge.

[0028] The transporter terminal (120) is a device corresponding to a transport company that transports livestock manure for discharge. For example, the transporter terminal (120) may be a terminal of a driver of a waste transport vehicle (3) or a terminal of a company that acts as an agent for manure transport. The transporter can call and rent a transport vehicle (600) that matches the amount of livestock manure input through the discharger terminal (110) via the transporter terminal (120).

[0029] The processor terminal (130) is a device corresponding to a processor that processes manure transported by a waste transport vehicle (3). The processor can input specific details that occur during manure processing using the processor terminal (130).

[0030] In addition, the generator, carrier, and processor can access the waste discharge transfer management system (2) using their respective terminals (110, 120, 130) to view information such as electronic transfer documents provided by the waste discharge transfer management system (2), or input necessary information into the waste discharge transfer management system (2). In addition, if there are errors in the electronic transfer documents automatically generated by the waste discharge transfer management system (2), the generator, carrier, and / or processor may directly correct them, or request corrections from the waste discharge transfer management system (2) for parts that cannot be directly corrected.

[0031] For the above operation, the waste discharge transfer management system (2) may be configured to communicate with each user terminal (110, 120, 130), machine learning service server (4) and / or transport vehicle (3), etc. through a wired and / or wireless network.

[0032] At this time, the wired and / or wireless networks include LAN (Local Area Network), MAN (Metropolitan Area Network), GSM (Global System for Mobile Network), EDGE (Enhanced Data GSM Environment), HSDPA (High Speed ​​Downlink Packet Access), W-CDMA (Wideband Code Division Multiple Access), CDMA (Code Division Multiple Access), TDMA (Time Division Multiple Access), Bluetooth, Zigbee, Wi-Fi, VoIP (Voice over Internet Protocol), LTE Advanced, IEEE802.16m, WirelessMAN-Advanced, HSPA+, 3GPP Long Term Evolution (LTE), Mobile WiMAX (IEEE 802.16e), UMB (formerly EV-DO Rev. C), Flash-OFDM, iBurst and MBWA (IEEE 802.20) systems, HIPERMAN, Utilizes Beam-Division Multiple Access (BDMA), World Interoperability for Microwave Access (Wi-MAX), and ultrasound It may refer to a communication network based on one or more communication methods selected from a group of communication methods, but is not limited thereto.

[0033] FIG. 2 is a schematic block diagram showing the hardware configuration of a waste discharge transfer management system according to one embodiment.

[0034] Referring to FIG. 2, the waste discharge transfer management system (2) according to the embodiments is implemented in the form of a computing device including hardware (200) and may include memory (210), a processor (220), a communication module (230), and an input / output unit (240).

[0035] The memory (210) is a non-transient computer-readable recording medium and may include a permanent mass storage device such as RAM (random access memory), ROM (read only memory), disk drive, SSD (solid state drive), flash memory, etc. Here, the permanent mass storage device such as ROM, SSD, flash memory, disk drive, etc. may be included in the device or server described above as a separate permanent storage device distinct from the memory (210).

[0036] Additionally, the memory (210) may store an operating system and at least one program code (e.g., code for a security module or an application installed to provide a specific service). These software components may be loaded from a computer-readable recording medium separate from the memory (210). This separate computer-readable recording medium may include a computer-readable recording medium such as a floppy drive, disk, tape, DVD / CD-ROM drive, or memory card.

[0037] In another embodiment, software components may be loaded into memory (210) via a communication module (230) rather than a computer-readable recording medium. For example, at least one program may be loaded into memory (210) based on a computer program installed by files provided over a network by developers or a file distribution system (e.g., an application store service server) that distributes installation files for applications.

[0038] The processor (220) may be configured to process instructions of a computer program by performing basic arithmetic, logic, and input / output operations. Instructions may be provided to the processor (220) by memory (210) or a communication module (230). For example, the processor (220) may be configured to execute instructions received according to program code stored in a recording device such as memory (210).

[0039] The communication module (230) can provide a function for the waste discharge transfer management system (2) to communicate with each user terminal (110, 120, 130), machine learning service server (4), and transport vehicle (3), etc., via a network. Additionally, the communication module (230) can provide a function for the waste discharge transfer management system (2) to communicate with one or more other devices via a wired and / or wireless network. That is, the communication module (230) is a part that realizes each function module described above with reference to FIG. 1, by controlling its function by a processor (220) that references memory (210).

[0040] The input / output unit (240) may be a means for interfacing with an external input / output device (not shown). For example, the external input device may include devices such as a keyboard, mouse, microphone, camera, etc., and the external output device may include devices such as a display, speaker, haptic feedback device, etc. As another example, the input / output unit (240) may be a means for interfacing with a device in which the functions for input and output are integrated into one, such as a touchscreen.

[0041] In addition, in other embodiments, the waste discharge transfer management system (2) may include more hardware components than those shown in FIG. 2 depending on the nature of the device to which it is applied. For example, when the waste discharge transfer management system (2) is applied to a user device, it may be implemented to include at least some of the input / output devices described above, or it may include additional components such as a transceiver, a GPS (Global Positioning System) module, a camera, various sensors, a DB, etc. As a more specific example, when the user device is a smartphone, it may be implemented to include additional components such as a camera module, various physical buttons, buttons using a touch panel, input / output ports, and a vibrator for vibration, which are generally included in smartphones.

[0042] However, the components and forms of the computing device described in this specification are merely exemplary, and the configuration of the computing device in which the waste discharge transfer management system (2) is implemented may differ from that described in this specification due to the adoption of other known technologies or future advancements in information and communication technology.

[0043] The waste discharge transfer management method described below may be performed by a waste discharge transfer management system (2) implemented in the form of a computing device including the hardware (200) configuration described above with reference to FIG. 2. As an example, the waste discharge transfer management method may be provided to a user in the form of a service based on at least one of an application, software, and other programs that operate based on a user device and / or a server.

[0044] FIG. 3 is a schematic block diagram of a waste discharge transfer management system and a waste transport vehicle communicating with it according to one embodiment.

[0045] Referring to FIG. 3, the waste transport vehicle (3) may include a communication module (31), a GPS module (32), a weight sensor (33), a vehicle information unit (34), and a control unit (35). At this time, each element constituting the transport vehicle (3) does not necessarily mean a form integrated with the vehicle, and for example, one or more of the communication module (31), GPS module (32), weight sensor (33), vehicle information unit (34), and control unit (35) may be provided in the form of a separate vehicle terminal and mounted on the waste transport vehicle (3).

[0046] The GPS module (32) receives signals sent from GPS satellites and can calculate the driving position including the location where the waste transport vehicle (3) loads and unloads manure.

[0047] A weight sensor (33) is mounted on a cushioning mechanism (e.g., a leaf spring) of the waste transport vehicle (3) to calculate the weight of the waste transport vehicle (3) as manure is loaded and unloaded. At this time, the weight of the transport vehicle (3) can be recorded in the form of weight information over time.

[0048] The vehicle information section (34) can be configured to store basic information such as a vehicle number that can identify the waste transport vehicle (3) and information about the company managing the vehicle with the vehicle number.

[0049] The control unit (35) can configure the operation data of the transport vehicle (3) by linking the information on the weight change of the waste transport vehicle (3) as manure is loaded and discharged with the GPS information. At this time, the operation data serves as basic data for the control unit (35) or the waste discharge transfer management system (2) to calculate the loading position, discharge position, weight before and after loading, and weight before and after discharge of the transport vehicle (3).

[0050] The communication module (31) can transmit the location information of the waste transport vehicle (3) obtained by the GPS module (32) and the weight information of the waste transport vehicle (3) obtained by the weight sensor (33) to the waste discharge transfer management system (2).

[0051] The waste discharge transfer management system (2) is configured to communicate with the communication module (31) of the transport vehicle (3) and may include an image receiving unit (21), a machine learning unit (22), and a loading / unloading detection unit (23). In one embodiment, the waste discharge transfer management system (2) may further include a discharge transfer document detection unit (24). Furthermore, in one embodiment, the waste discharge transfer management system (2) may further include a database (DB) (25) in which information received or processed by the waste discharge transfer management system (2) is stored.

[0052] Each of the above units (21-24) and DB (25) can be realized at least partially by using the hardware (200) of the waste discharge transfer management system (2).

[0053] That is, the devices described in this specification may be entirely hardware, or may have aspects that are partially hardware and partially software. For example, each unit constituting the waste discharge transfer management system (2) and each system, device, server, etc. communicating with it, and each module or unit included therein may collectively refer to devices and related software for exchanging data of a specific format and content via electronic communication. In this specification, terms such as "unit," "module," "server," "system," "platform," "device," or "terminal" are intended to refer to a combination of hardware and software driven by said hardware. For example, the hardware here may be a data processing device including a CPU or other processor. Additionally, the software driven by the hardware may refer to a running process, object, executable file, thread of execution, program, etc.

[0054] Additionally, in this specification, each component constituting the waste discharge transfer management system (2) is not intended to refer to a physically distinct component. That is, in FIG. 1, each part (21-24) and DB (25) included in the waste discharge transfer management system (2) are depicted as distinct blocks, but this is a functional distinction of the waste discharge transfer management system (2) based on the operations performed by them. Depending on the embodiment, some or all of the aforementioned parts may be integrated within a single device, or one or more parts may be implemented as separate devices physically distinct from other parts. For example, each device constituting the waste discharge transfer management system (2) may be components connected to communicate with each other in a distributed computing environment.

[0055] The image receiving unit (21) receives operation data by communicating with the communication module (31) of the waste transport vehicle (3) and can extract weight information of the transport vehicle (3) over time in the form of an image. At this time, in addition to weight information over time, the operation data may further include vehicle information such as the name of the transport company to which the vehicle belongs, the transfer number, and the vehicle number. Furthermore, in one embodiment, the operation data may include information calculated by the GPS module (32) or control unit (35) of the transport vehicle (3), such as loading time, loading location (latitude, longitude), discharge location (latitude, longitude), weight before and after loading, and / or weight information before and after discharge.

[0056] The machine learning unit (22) is configured to support learning of an artificial intelligence (AI)-based analysis or detection model and the classification and detection of information using the same. For example, the machine learning unit (22) may be configured to store an AI-based analysis or detection model itself and perform learning on it, or it may communicate with an external machine learning service server (4) so ​​that the learning and utilization of the model stored in the machine learning service server (4) is carried out through the waste discharge transfer management system (2).

[0057] In one embodiment, the AI-based model supported by the machine learning unit (22) may include an image analysis and classification model such as a Convolutional Neural Network (CNN), and / or an object detection and classification model such as YOLO (You Only Look Once). However, the types of AI-based models supported by the waste discharge transfer management system (2) may be any known or future-developed models and are not limited to those disclosed in this specification.

[0058] The loading / unloading detection unit (23) can be configured to detect the loading and / or unloading of waste on the transport vehicle (3) by using time-dependent weight information extracted from the operation data of the transport vehicle (3) as an input value for an image analysis model, such as a CNN supported by the machine learning unit (22).

[0059] Additionally, the loading / unloading detection unit (23) can derive an output value indicating whether there is an abnormality in the buffer mechanism (e.g., leaf spring) of the vehicle by utilizing the analyzed loading and / or unloading status together with weight information over time. For the above operation, an image analysis model such as a CNN may be configured to determine whether the buffer mechanism is faulty from an unknown weight information image by being pre-trained using training data in which the state of the buffer mechanism (normal or faulty, etc.) is labeled in advance. This will be described in detail later.

[0060] The discharge transfer document detection unit (24) can be configured to detect discharge transfer documents from the time-dependent weight information image of the transport vehicle (3) by inputting time-dependent weight information extracted from the operation data of the transport vehicle (3) into an object detection model such as YOLO supported by the machine learning unit (22). To this end, the object detection model can be pre-trained to identify the portion corresponding to the section where a discharge transfer document must be issued and the number thereof in the time-dependent weight information image.

[0061] At this time, the detection of a discharge transfer document means that the fact that the transport vehicle (3) loaded waste during a specific time interval and then unloaded it at a specific point in time is confirmed from the data, and it is presumed that there is a need to officially record this. That is, by analyzing the weight change pattern of the transport vehicle (3) over time, the time at which the discharge transfer document needs to be issued can be determined based on the time when the waste was loaded or unloaded.

[0062] The discharge transfer document detection unit (24) can detect discharge transfer documents by identifying key points in waste movement based on the weight change of the transport vehicle (3) over time and automatically identifying sections where a discharge transfer document is likely to be needed. In addition, in one embodiment, the discharge transfer document detection unit (24) may be further configured to determine whether there is an abnormal discharge based on the discharge transfer documents detected from the operation data of the transport vehicle (3) by further training an object detection model using operation data and discharge transfer document issuance information from one or more transporters (transport vehicles and / or transport companies) as training data. The specific operation of the discharge transfer document detection unit (24) will be described in detail later.

[0063] DB (25) can store operation data received by the waste discharge transfer management system (2) and / or information processed and generated by the waste discharge transfer management system (2) based on the operation data. For example, an electronic transfer document generated by the waste discharge transfer management system (2) based on the operation data can be stored in DB (25), so that the discharger, transporter and / or processor can access the waste discharge transfer management system (2) to view and / or modify the electronic transfer document.

[0064] FIG. 4 is a flowchart showing each step of a waste discharge transfer management method according to one embodiment.

[0065] Referring to FIG. 4, the waste discharge transfer management system (2) can use known data for one or more waste transport vehicles (3) that transport waste as training data, label the weight information over time for each vehicle with a broken leaf spring and a vehicle with a broken leaf spring, and use this to train an image analysis model such as a CNN (S101).

[0066] Although the failure of a leaf spring is described in this specification as an example, it will be readily understood by those skilled in the art that the same principle can be applied to cushioning mechanisms for other types of weight.

[0067] In FIG. 4, the waste discharge transfer management system (2) is shown to train an image analysis model by communicating with a separate machine learning service server (4), but this is exemplary, and in other embodiments, an image analysis model such as a CNN may be stored within the waste discharge transfer management system (2) and training may be performed within the waste discharge transfer management system (2).

[0068] When a discharge company (101) intends to discharge livestock manure, a waste transport vehicle (3) arrives at the discharge farm and loads the manure onto the vehicle (S102). Afterwards, the waste transport vehicle (3) transports the loaded manure to a livestock manure treatment company (104) (S103) so that the manure is unloaded at the treatment company (104) (S105).

[0069] In this specification, detailed descriptions of the loading and unloading processes performed manually by a transporter are omitted to clarify the gist of the invention.

[0070] The waste transport vehicle (3) can transmit vehicle operation data to the waste discharge transfer management system (2) at predetermined time intervals (e.g., 3 seconds) from departure, including at least the time interval between loading and unloading (S104). At this time, the operation data includes weight information of the transport vehicle (3) over time. In addition, the operation data may further include information such as the name of the transport company to which the transport vehicle (3) belongs, the transfer number, the vehicle number, the loading time, the loading location (latitude and longitude), the discharge location (latitude and longitude), the weight before and after loading, and the weight before and after discharge.

[0071] The waste discharge transfer management system (2) can extract weight information over time of the transport vehicle (3) in the form of an image from the received operation data (S106), and input the extracted image into an image analysis model such as a CNN to derive whether and when to load, and / or whether and when to unload (S107).

[0072] In addition, at this time, the waste discharge transfer management system (2) can detect whether the leaf spring of the transport vehicle (3) is broken by analyzing the image extracted from the operation information by training the image analysis model in advance with weight information according to time for each vehicle with a broken leaf spring and a vehicle without a broken leaf spring (S108).

[0073] When a failure of the leaf spring is detected from an image of the operation data, the waste discharge transfer management system (2) may provide the failure detection result to a user terminal corresponding to the waste transport vehicle (3) so that necessary measures, such as repair, may be taken (S109). For example, the failure detection result may be transmitted to the driver of the transport vehicle (3) or the manager of the transport company, but is not limited thereto.

[0074] In one embodiment, the waste discharge transfer management system (2) may generate an electronic transfer document for manure based on loading information including whether it was loaded, the time and / or location detected from the operation data, and unloading information including whether it was unloaded, the time and / or location (S110). Additionally, vehicle information and company information included in the operation data may be further utilized in generating the electronic transfer document.

[0075] At this time, vehicle information may refer to the vehicle number of the transport vehicle (3), the vehicle's loadable weight (tonnage, etc.), etc. Additionally, loading information may include the amount of weight change at the coordinates corresponding to the loading area and information on the discharge company. Additionally, unloading information may include the amount of weight change at the coordinates corresponding to the unloading area and information on the processing company.

[0076] In one embodiment, the generation of an electronic transfer document based on the above information may also be performed by a pre-trained AI-based analysis model. For example, the waste discharge transfer management system (2) may use location information, weight information, and driving information for each waste transport vehicle (3) per unit time (e.g., one day), along with the time of weight measurement, vehicle number, weight difference, time difference, location difference, result code, latitude, and longitude as features for learning, calculate the weight difference, speed difference, and location difference values ​​for each vehicle to define the loading and unloading status and train the analysis model.

[0077] The electronic transfer document generated as described above may be provided to the waste discharge transfer management system (2) to the discharge company (101), the transport vehicle (3) (referring to a terminal corresponding to the transport vehicle and / or the transport company) and / or the treatment company (104) (S111 to S113).

[0078] The AI-based image analysis process is explained in more detail below.

[0079] FIG. 5 is a flowchart showing each step of the AI-based image classification process in a waste discharge transfer management method according to one embodiment.

[0080] Referring to FIG. 5, the waste discharge transfer management system can extract weight information over time from the operation data of a waste transport vehicle, for example, in the form of an image in the form of a weight change graph (S201). FIG. 6 (a) to (c) shows such an image, which visually represents the weight change data occurring during the loading and unloading of waste on the vehicle along with the time axis.

[0081] The weight change graph shown in FIG. 6 measures the weight of a vehicle equipped with a normal leaf spring, and the weight change over time shows a pattern where the weight is close to zero when unloading and forms a peak when loading. On the other hand, FIG. 7 (a) to (c) shows the weight change of a transport vehicle with a broken leaf spring, and the graph appears in the shape of a square or rectangle, where the weight value does not decrease to zero even after unloading and remains near the peak, or where there is no part where the weight clearly decreases to zero.

[0082] The waste discharge transfer management system according to the embodiments can detect whether a leaf spring is faulty by recognizing weight information as an image and classifying it based on the difference in patterns between the weight change graphs of a vehicle equipped with a normal leaf spring and a vehicle with a faulty leaf spring.

[0083] To this end, referring again to FIG. 5, the waste discharge transfer management system can generate a graph outline to recognize the pattern of the extracted weight change graph (S202). This process is intended to allow the model to easily learn important features of the graph and serves to clarify the timing of changes, such as the loading and unloading of manure.

[0084] In one embodiment, the waste discharge transfer management system may perform image preprocessing before the step of generating the aforementioned outline, such as representing the part recognized as a weight change graph in a specific color (e.g., blue) or / or changing the color space of the image so that the representation of the said color is prominent. For example, since the HSV method is more suitable than BGR when verifying only a specific color range, the waste discharge transfer management system may convert the weight information image over time into an image in the HSV color space.

[0085] At this time, the waste discharge transfer management system may be configured to designate an area having a specific color corresponding to the graph as a region of interest within the image, generate a mask for that area, and generate an outline within the masked area.

[0086] In addition, in one embodiment, the waste discharge transfer management system may be further configured to generate an outline by excluding parts of the weight information image where the short-term peak value of the weight is less than a preset threshold value (e.g., 500 kg). For example, parts with small peak values ​​of weight may represent weight changes caused by the shaking of the leaf springs while the transport vehicle travels on unpaved roads, and these parts are unnecessary for image analysis.

[0087] The waste discharge transfer management system can train an image analysis model using labeled data in which the failure status of a leaf spring is known in advance as an image with a generated outline.

[0088] In one embodiment, the waste discharge transfer management system may expand the training data based on a labeled weight change graph to increase the amount of training data (S203). For example, the waste discharge transfer management system may generate additional training data by applying modifications such as scaling, rotation, and / or adding noise to the weight change graph data where the failure of the leaf spring is known.

[0089] In one embodiment, the waste discharge transfer management system may reconfigure the analysis model to change the color processing order read by the analysis model to match the graph data to be learned. For example, while the open-source algorithm OpenCV reads image data in the order of Blue, Green, and Red, general image data is in the RGB order; therefore, the model can be reconfigured to match the channel order of the image data to be learned.

[0090] In addition, in one embodiment, the waste discharge transfer management system may additionally perform preprocessing on the image to match the input characteristics of the analysis model before applying the training data to the image analysis model (S204). For example, by preprocessing the waste discharge transfer management system graph data into a state suitable for model training through operations such as normalizing or removing noise, the learning accuracy of the model can be increased and overfitting can be prevented.

[0091] Next, the waste discharge transfer management system can train an image analysis model, such as a CNN, using the preprocessed data (S205). Through this, a model can be constructed that can detect whether waste has been loaded or unloaded and whether the leaf spring has failed, based on the weight change data of the transport vehicle over time.

[0092] Once the image analysis model is trained, the waste discharge transfer management system uses unknown weight change graph data obtained from the operation data of a waste transport vehicle, which does not know whether the leaf spring is broken, as an input value to the analysis model (S206), and can obtain the loading information and unloading information represented by the graph data and the classification result regarding whether the leaf spring is broken (S207).

[0093] Figures 8 (a) to (c) show the results obtained by classifying weight change graph images using an AI-based analysis model according to the waste discharge transfer management method described above. In Figure 8, graphs (a), (d), (e), (f), (g), (i), and (j) are graphs in which the weight value measured by the weight sensor clearly changes to 0 or a peak value depending on loading and unloading, and these were classified by the image analysis model as image data where the leaf spring did not fail. On the other hand, in the case of graphs (b), (c), and (h), it can be seen that the weight value does not clearly decrease to 0 even after unloading, and these were classified by the image analysis model as image data where the leaf spring failed.

[0094] The inventors checked the error between the correct answer value labeled in the training data and the predicted value classified by the image analysis model using the above method, and returned the error values ​​as an average to derive an indicator of how well the image analysis model predicted (classified) the entire dataset. As a result, the validation accuracy was measured to be approximately 98%, confirming that the model according to the embodiments of the present invention can perform prediction of leaf spring failure with a high probability.

[0095] In the case of a transport vehicle classified as having a broken leaf spring by the waste discharge transfer management method according to the embodiments, it can be excluded when generating an electronic transfer document for waste discharge, and by notifying the transport company for maintenance, such as repair of the transport vehicle, accurate discharge transfer processing can be carried out.

[0096] Meanwhile, in order to determine whether the discharge transfer of waste, such as livestock manure, is being carried out accurately, in addition to assessing the operation of the transport vehicle, such as loading and unloading times and leaf spring failures, there may be cases where it is necessary to determine through the detection of discharge transfer documents whether there are any discrepancies in the quantity of the transport company's (or transport vehicle's) discharge transfer documents and whether illegal dumping has occurred.

[0097] In this specification, illegal dumping, etc., for which the discharge transfer procedure under normal administrative procedures is not performed is referred to as "discharge anomaly," and a waste discharge transfer management method according to one embodiment is configured to detect discharge anomalies based on a detected discharge transfer document using weight information over time as an input value.

[0098] FIG. 9 is a flowchart showing each step of the process of detecting a waste discharge transfer document by a waste discharge transfer management method according to one embodiment.

[0099] Referring to FIG. 9, in this embodiment, the waste discharge transfer management system extracts discharge transfer information from one or more carriers (S301), and through the extracted discharge transfer information, it can perform training of a discharge transfer document detection model based on object detection such as YOLO (S302).

[0100] In one embodiment, discharge transfer information can be extracted using a graph of weight change over time as input data, similar to the detection of leaf spring failure described above, and may include the number of discharge transfer documents extracted therefrom. FIGS. 10 (a) to (c) represent examples of such input data, and FIGS. 10 (a) to (c) respectively show weight change graphs extracted from the operation data of the waste transport vehicle when there are 7, 2, and 4 discharge transfer documents.

[0101] In one embodiment, the waste discharge transfer management system may collect discharge transfer document data (i.e., discharge transfer document issuance history) generated by each company from a transporter terminal and store it in the DB of the waste discharge transfer management system. At this time, the information stored in the DB is used to calculate statistics on discharge information for each company (e.g., average quantity of discharge transfer documents) and can function as basic data for detecting discharge anomalies in the future.

[0102] The waste discharge transfer management system can train a model by utilizing an object detection model such as YOLO to automatically identify the portion corresponding to the discharge transfer document issuance section from a graph of weight change over time. To this end, the training data in the form of a graph of weight change over time can be configured to represent various discharge transfer situations by consisting of images and labels as described above.

[0103] In one embodiment, the labels included in the training data may include the coordinates of a bounding box corresponding to the section where a discharge transfer document must be issued within a weight change graph over time. However, this is exemplary, and the method of labeling the training data is not limited thereto.

[0104] After the object detection model, such as YOLO, is trained, the waste discharge transfer management system receives the operation data of the company to be analyzed (S303), and by inputting the weight information over time extracted from the operation data (i.e., weight change graph data) into the object detection model in the form of an image, it can determine whether there is an abnormal discharge in the operation of the transport vehicle (S304).

[0105] In one embodiment, the waste discharge transfer management system further receives discharge transfer document issuance information of the relevant transport vehicle (or, equivalent transport company), and can determine whether there is an abnormal discharge by comparing the number of detected discharge transfer documents detected by the object detection model with the number of discharge transfer documents issued in the discharge transfer document issuance information.

[0106] The number of discharge transfer documents detected from operational data is based on a reduction in vehicle weight at a specific point in time; if there is a discrepancy between the number of documents detected based on this weight reduction and the actual number of issued documents, it can be presumed that there is an emission anomaly. For example, a waste discharge transfer management system can detect emission anomalies, such as illegal dumping, if the number of issued documents indicated by the issuance information is significantly lower than the number of documents detected from operational data.

[0107] As described above, in one embodiment, the waste discharge transfer management system determines whether there is a problem by comparing the issuance history of discharge transfer documents from a transport vehicle or transporter with actual waste discharge data. For example, if the issuance of a discharge transfer document is omitted or the issuance history is abnormally low, it can be determined that there is a possibility of illegal dumping or data errors.

[0108] In another embodiment, the waste discharge transfer management system may determine whether there is an abnormal discharge based on location information where the vehicle reduction occurred (e.g., GPS data included in driving data). For example, if the location of the vehicle in the section where the reduction was detected through driving data was an area where disposal was difficult, the waste discharge transfer management system may detect the reduction at that location as an abnormal discharge.

[0109] In this case, areas where disposal is difficult may refer to locations where the reduction has occurred that are (i) areas where no treatment facility is located within a certain range (e.g., 1 km) from the location, or (ii) areas where discharge is prohibited by law, such as mountainous areas, areas near rivers or the sea, water source protection zones, or riparian protection zones. Furthermore, the fact that no treatment facility is located within a certain range may include not only cases where the treatment facility itself does not exist within the range, but also cases where the treatment facility contracted with the vehicle's transporter does not exist within the range.

[0110] In one embodiment, the waste discharge transfer management system may use a method of comparing and classifying the coordinates of a reduction site and the coordinates of a treatment facility or a discharge prohibition area using a clustering algorithm, such as K-Nearest Neighbor, to determine whether the location where the vehicle reduction occurred corresponds to an area where disposal is difficult. However, this is merely exemplary, and in other embodiments, the waste discharge transfer management system may determine whether an area is difficult to dispose of using a rule-based method based on criteria such as whether the distance between coordinates is less than a preset distance.

[0111] In another embodiment, the waste discharge transfer management system may determine potential candidates for emission anomalies by incorporating weather information at the time the vehicle was reduced (e.g., the time of unloading included in the operation data) in addition to location information where the vehicle was reduced. For example, if the weather at the location was cloudy or rainy on the date or time detected as the discharge transfer document issuance section where the vehicle was reduced, the detected section is set as a candidate for an emission anomaly with a high probability of being an anomaly, and the final determination of whether an emission anomaly exists can be made through detailed analysis of additional data, such as location, for the detected cases corresponding to the candidate for an emission anomaly. At this time, the weather at the location where the vehicle was reduced may be received by the waste discharge transfer management system through a separate information provision server (not shown) that provides weather information using an API (Application Programming Interface), etc.

[0112] The above embodiments are based on the fact that illegal emissions occur more frequently when the weather is cloudy or rainy; however, the weight loss achieved under certain weather conditions may be considered as a candidate for emission excess, and is not limited by the examples described herein.

[0113] The number of issued emission transfer documents, location information where a reduction in the vehicle corresponding to the emission transfer document was detected, and weather information on the date or time where a reduction in the vehicle corresponding to the emission transfer document was detected, as described above, can be utilized for emission anomaly detection in a manner where only one of the information is used or where all three types of information are reflected in combination.

[0114] In one embodiment, the waste discharge transfer management system may be further configured to transmit a notification of the discharge abnormality to the relevant transport vehicle (or transport company) when a discharge abnormality is detected by the process described above (S305).

[0115] FIG. 11 is a diagram illustrating the detection results of a waste discharge transfer document by a waste discharge transfer management method according to one embodiment.

[0116] Each section marked with a bounding box in Fig. 11 represents a section detected by an object detection model as corresponding to a section where one discharge transfer document was issued. Based on whether the number of discharge transfer documents detected as described above is relatively higher than the actual number of waste discharge transfer documents issued by the relevant transporter, and the extent thereof, it is possible to determine whether the livestock manure discharge procedure was carried out normally based on the vehicle operation data.

[0117] The operation of the waste discharge transfer management method according to the embodiments described above may be implemented at least partially as a computer program and recorded on a computer-readable recording medium. A computer-readable recording medium on which a program for implementing the operation according to the methods of the embodiments is recorded includes all types of recording devices in which data readable by a computer is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. Additionally, the computer-readable recording medium may be distributed across networked computer systems, and computer-readable code may be stored and executed in a distributed manner. Furthermore, functional programs, codes, and code segments for implementing the present embodiment will be easily understood by a person skilled in the art to which the present embodiment belongs.

[0118] Additionally, each block or each step illustrated in the flowcharts of this specification may represent a module, segment, or part of code comprising one or more executable instructions for executing a specified logical function(s). Furthermore, in some alternative embodiments, the functions mentioned in the blocks or steps may occur out of order. For example, two blocks or steps illustrated in succession may actually be performed substantially simultaneously, or the blocks or steps may occasionally be performed in reverse order according to the corresponding function.

[0119] The present invention described above has been explained with reference to the embodiments illustrated in the drawings, but this is merely illustrative and those skilled in the art will understand that various modifications and variations of the embodiments are possible therefrom. However, such modifications should be considered to be within the technical scope of protection of the present invention. Accordingly, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.

Claims

Claim 1 As a waste discharge transfer management system, an image receiving unit configured to receive operation data of a waste transport vehicle and extract hourly weight information of the waste transport vehicle in the form of an image from the operation data; and a machine learning unit configured to support the use of a pre-trained image analysis model using a training image representing hourly weight information; and includes a loading / unloading detection unit configured to detect one or more of the loading and unloading of waste for the waste transport vehicle by using an image extracted from the above operation data as an input value for the above image analysis model, wherein each of the image extracted from the above operation data and the above learning image is an image in the form of a weight change graph including a first axis representing time and a second axis representing weight, and the learning image is labeled data that includes a graph outline generated for pattern recognition of the weight change graph and has a preset abnormality in weight information, and the machine learning unit is further configured to support the use of an object detection model that is pre-trained to detect waste discharge transfer documents from hourly weight information corresponding to one or more waste transport vehicles, and the loading / unloading detection unit is further configured to determine whether there is an abnormality in the hourly weight information represented by the image extracted from the above operation data based on whether waste is loaded or unloaded for the waste transport vehicle using the above analysis model, and the waste discharge transfer management system is further configured to receive discharge transfer document issuance information of the waste transport vehicle, and the waste transport vehicle's using an image extracted from the above operation data as an input value for the above object detection model The system further includes an emission transfer document detection unit configured to detect an emission transfer document corresponding to the operation, and to determine whether there is an emission abnormality corresponding to the waste transport vehicle or the transport company operating the waste transport vehicle based on the number of emission transfer documents detected from the operation data, location information where the emission transfer document was detected, and weather information at the time when the emission transfer document was detected.The waste discharge transfer management system is further configured such that the discharge transfer document detection unit detects that there is a discharge anomaly corresponding to the waste transport vehicle or the transport company operating the waste transport vehicle when the number of discharge transfer documents issued indicated by the issuance information is less than or equal to a preset level compared to the number of discharge transfer documents detected from the operation data, and using the object detection model, determines the section where the vehicle reduction occurred in the operation data as the discharge transfer document issuance section, and at the time corresponding to the discharge transfer document issuance section, queries the weather information of the location corresponding to the discharge transfer document issuance section from the information providing server, and if the weather information indicates cloudy or rainy weather, determines the discharge transfer document issuance section as a candidate for a discharge anomaly, and for the discharge transfer document issuance section determined as a candidate for a discharge anomaly, determines whether there is a discharge anomaly by comparing and classifying the first location information corresponding to the discharge transfer document issuance section and the second location information of a preset treatment facility or discharge prohibition area using a clustering algorithm. Claim 2 A waste discharge transfer management system according to claim 1, wherein the abnormality of the hourly weight information includes whether the buffer mechanism equipped in the waste transport vehicle is malfunctioning, and the image analysis model is a learning image in which the malfunction of the buffer mechanism is pre-learned. Claim 3 delete Claim 4 delete Claim 5 A method for managing waste discharge transfer performed by a computing device, comprising: a step in which the computing device receives operation data of a waste transport vehicle; a step in which the computing device extracts hourly weight information of the waste transport vehicle in the form of an image from the operation data; a step in which the computing device detects one or more of the loading and unloading of waste for the waste transport vehicle by inputting the image extracted from the operation data into a pre-trained image analysis model using a training image representing the hourly weight information; a step in which the computing device determines, using the analysis model, whether there is an anomaly in the hourly weight information represented by the image extracted from the operation data based on whether the waste for the waste transport vehicle is loaded or unloaded; a step in which the computing device receives discharge transfer document issuance information of the waste transport vehicle; and a step in which the computing device detects a discharge transfer document corresponding to the operation of the waste transport vehicle by inputting the image extracted from the operation data into an object detection model pre-trained to detect a discharge transfer document of waste from hourly weight information corresponding to one or more waste transport vehicles.The computing device comprises the step of determining whether there is an emission anomaly corresponding to the waste transport vehicle or the transporter operating the waste transport vehicle based on the number of emission transfer documents detected from the operation data, location information where the emission transfer documents were detected, and weather information at the time when the emission transfer documents were detected, wherein each of the image extracted from the operation data and the learning image is an image in the form of a weight change graph including a first axis representing time and a second axis representing weight, and the learning image is labeling data that includes a graph outline generated for pattern recognition of the weight change graph and has a preset abnormality in weight information, and the step of determining whether there is an emission anomaly is included in the step of the computing device detecting that there is an emission anomaly corresponding to the waste transport vehicle or the transporter operating the waste transport vehicle if the number of emission transfer documents issued indicated by the issuance information is less than a preset level compared to the number of emission transfer documents detected from the operation data; and the computing device determining the section where the vehicle's weight reduction occurred in the operation data as the emission transfer document issuance section using the object detection model. A waste discharge transfer management method comprising: a step in which the computing device queries weather information of a location corresponding to the discharge transfer document issuance section from an information providing server at a time corresponding to the discharge transfer document issuance section, and if the weather information indicates cloudy or rainy weather, determines the discharge transfer document issuance section as a candidate for discharge abnormality; and a step in which the computing device determines whether there is a discharge abnormality for the discharge transfer document issuance section determined as a candidate for discharge abnormality by comparing and classifying first location information corresponding to the discharge transfer document issuance section and second location information of a preset treatment facility or discharge prohibition area using a clustering algorithm. Claim 6 A waste discharge transfer management method according to claim 5, wherein the abnormality of the hourly weight information includes whether the buffer mechanism equipped in the waste transport vehicle is malfunctioning, and the image analysis model is pre-learned using a learning image in which the malfunction of the buffer mechanism is labeled. Claim 7 delete Claim 8 delete Claim 9 A computer program stored on a computer-readable recording medium to perform an AI-based waste discharge transfer management method in combination with hardware according to Paragraph 5 or 6.

Images