Method for determining state of palm fruit, and system therefor

Abstract

The present invention relates to a system for determining a state of palm fruit, the system comprising: a server (100); a supplier terminal (200); and a factory device (300), wherein the supplier terminal (200) receives supplier information (210) from a supplier, generates first image data by photographing fruits at a supply site, generates GPS information (220) of the supply site at a fruit-photographing time point, and generates, on the basis of the first image data, first state-quantity information (230) by using a first artificial intelligence model, the factory device (300) acquires, from the supplier terminal (200), the supplier information (210), the GPS information (220) and the first state-quantity information (230), generates second image data by photographing the fruits transported by a conveyor belt, and generates, on the basis of the second image data, second state-quantity information (310) by using a second artificial intelligence model, and the server (100) acquires the supplier information (210), the GPS information (220) at the photographing time point, the first state-quantity information (230), and the second state-quantity information (310) and stores same in a DB.

Description

Method and system for determining the condition of palm fruit

[0001] The present invention belongs to the field of agricultural product processing, distribution, and processing technology, and in particular relates to a system and method for automatically determining and managing the condition of palm fruits. More specifically, the invention relates to a system that captures images of palm fruits using supplier terminals and factory equipment, evaluates the condition and quantity of the fruits using an artificial intelligence model, and manages these data in an integrated manner through a server. This can contribute to improving the efficiency of quality control and supply chain management for palm fruits.

[0002]

[0003] Conventional agricultural systems primarily utilize various equipment and software for quality control and data management during the production, harvesting, and distribution processes of palm fruit. For example, manually inspecting and recording the condition of fruit harvested at the source is time-consuming and susceptible to errors caused by human subjectivity. Furthermore, using separate systems or equipment for each stage makes integrated information management difficult and limits the efficient analysis and utilization of data across the entire supply chain.

[0004] Furthermore, existing palm fruit condition assessment systems rely primarily on data collection and analysis at a single point, limiting their ability to effectively integrate and utilize diverse data generated throughout the entire supply chain. Consequently, quality control of palm fruits is inconsistent, and it is difficult to monitor and respond to their status in real time. Additionally, existing systems lack user-friendly interfaces or automated report generation capabilities, requiring agricultural workers to possess the specialized knowledge or know-how necessary to effectively utilize the data.

[0005] Therefore, there has recently been active development of systems that utilize artificial intelligence (AI) technology to automatically determine the condition of palm fruits and manage integrated data across the entire supply chain. Despite these efforts, there is still a need for an integrated system that effectively integrates data collected from various supply points, analyzes conditions in real time to generate automated reports, and provides a user-friendly interface. Accordingly, the present invention aims to resolve conventional problems and improve the efficiency of palm fruit quality management by providing an integrated system and method for automatically determining and managing the condition of palm fruits.

[0006] Meanwhile, although the present invention was derived based at least on the technical background examined above, the technical problem or objective of the present disclosure is not limited to solving the problems or disadvantages examined above. That is, in addition to the technical issues examined above, the present disclosure can cover various technical issues related to the contents described below.

[0007]

[0008] The present invention aims to provide a system capable of automatically and accurately determining the condition of palm fruits. Furthermore, it seeks to streamline quality management throughout the entire supply chain through integrated data management between suppliers and factories.

[0009]

[0010] A system for determining the condition of palm fruit to solve the aforementioned problem, wherein the system comprises a server (100); a supplier terminal (200); and a factory device (300); wherein the supplier terminal (200) receives supplier information (210) from a supplier, photographs the fruit at the supply site to generate first image data, generates GPS information (220) of the time of photographing the fruit at the supply site, and generates first state-quantity information (230) using a first artificial intelligence model based on the first image data; wherein the factory device (300) obtains the supplier information (210); GPS information (220); and first state-quantity information (230) from the supplier terminal (200), photographs the fruit being transported by a conveyor belt to generate second image data, and generates second state-quantity information (310) using a second artificial intelligence model based on the second image data; and wherein the server (100) includes the supplier information (210); The above shooting time GPS information (220); the above first state-quantity information (230); and the second state-quantity information (310) can be obtained and stored in a DB.

[0011]

[0012] The present invention collects images of palm fruits from supply sites and factories and analyzes them using an artificial intelligence model, thereby accurately determining the condition of palm fruits and centrally managing the information.

[0013] In addition, by comparing the status-quantity information of the supplier and the factory and performing an ensemble when the difference is small, reports can be generated based on accurate data.

[0014] In addition, by including a salary report that reflects the proportion of suitable quantities, fair salary weights can be determined.

[0015] In addition, by enabling data creation, modification, and retrieval through the provider terminal, data accuracy can be improved and reports can be easily accessed.

[0016] In addition, by classifying the condition of palm fruits using an image analysis-based artificial intelligence model, the automation and accuracy of fruit condition assessment can be improved.

[0017] In addition, if there are significant data discrepancies, data reliability can be ensured by performing manual verification through an inspector.

[0018] In addition, the accuracy of the model can be continuously improved by further training the artificial intelligence model using verified data.

[0019] In addition, model performance can be efficiently optimized by selectively additionally training AI models with significant differences.

[0020]

[0021] FIG. 1 is a diagram showing the components of a server according to one embodiment of the present invention.

[0022] FIG. 2 is a drawing illustrating a system according to one embodiment of the present invention.

[0023] FIG. 3 is a diagram illustrating a method for identifying each supplier location when there are multiple suppliers according to an embodiment of the present invention.

[0024] FIG. 4 is a diagram showing a situation in which a plurality of suppliers or a plurality of RFID cards are delivered to a factory device according to one embodiment of the present invention.

[0025] FIG. 5 is a diagram illustrating a method for generating reports and salary information according to one embodiment of the present invention.

[0026] FIG. 6 is a diagram showing a program configuration that can be output to a supplier terminal according to an embodiment of the present invention.

[0027] FIG. 7 is a drawing for explaining state-quantity according to one embodiment of the present invention.

[0028]

[0029] The present invention will be described in detail below.

[0030] Prior to the description, some embodiments of the present invention are described in detail with reference to the exemplary drawings. In assigning reference numerals to the components of each drawing, the same reference numerals are used for identical components even if they are shown in different drawings. Furthermore, in describing the embodiments of the present invention, if it is determined that a detailed description would hinder understanding of the embodiments of the present invention, such detailed description may be omitted.

[0031] In addition, terms such as first, second, A, B, (a), (b), etc., may be used when describing the components of the embodiments of the present invention. These terms are intended merely to distinguish the components from other components, and the essence, order, or sequence of the components is not limited by these terms. Where it is stated that a component is "connected," "combined," or "joined" to another component, it should be understood that the component may be directly connected or joined to the other component, but that components may also be "connected," "combined," or "joined" between each component.

[0032] Before the explanation, the terms used in the present invention are defined.

[0033] The supplier information (210) of the present invention may include information such as the supplier's name, ID, and means of transportation. The first image data refers to an image of a palm fruit taken at the supply location.

[0034] In addition, GPS information (220) indicates location information (e.g., coordinates) at the time of photographing the fruit.

[0035] In addition, the first artificial intelligence model refers to a model that analyzes the condition and quantity of palm fruits based on the first image data.

[0036] In addition, the first state-quantity information (230) refers to the quantity information of the fruit by state obtained through the first artificial intelligence model.

[0037] Additionally, the factory device (300) includes a device for processing and analyzing palm fruits in a factory that produces products including palm oil from palm fruits. It may include a conveyor belt, an inspection device, etc.

[0038] In addition, the second image data refers to an image of fruit being transported via a conveyor belt in a factory.

[0039] In addition, the second artificial intelligence model refers to a model that analyzes the condition and quantity of the fruit based on the second image data.

[0040] Additionally, the second state-quantity information (310) refers to the quantity information of the fruit by state obtained through the second artificial intelligence model. DB refers to a database and stores and manages various information.

[0041] In addition, "when the difference is less than or equal to a predetermined value" means when the difference between the first state-quantity information and the second state-quantity information is within a pre-set allowable range. "Ensembling to generate third state-quantity information" refers to combining the two state-quantity information to generate more accurate information.

[0042] In addition, the pay report includes details regarding the pay to be paid to the supplier. "Payment information determined in proportion to the proportion of suitable quantity" means that the pay is determined based on the proportion of fruit of suitable quality among the total fruit supplied.

[0043] In addition, data creation buttons, data modification buttons, and report viewing buttons refer to interactive elements in an application or web interface that users can select to perform functions. A bounding box refers to a rectangular area used to mark a region of interest within an image. "Predicting" refers to the process in which an artificial intelligence model produces results based on input data.

[0044] Additionally, spine size refers to the length of the sharp spines formed on the externally visible part of the palm fruit. "Determined multiple" refers to a multiple value set in advance to serve as a standard of comparison. Overripe, suitable, and rotten states refer to conditions classified according to the maturity and quality of the palm fruit. Semi-ripe, unripe, and poor moisture states indicate that the fruit has not yet fully ripened or is of inferior quality.

[0045] In addition, "additional training" refers to retraining a model using new data to improve the performance of an artificial intelligence model.

[0046]

[0047] From now on, embodiments will be described based on each drawing of the present invention.

[0048] FIG. 1 is a diagram showing the components of a server according to one embodiment of the present invention.

[0049] The configuration of the server (100) illustrated in FIG. 1 is merely a simplified example. In one embodiment of the present disclosure, the server (100) may include other configurations for performing the computing environment of the server (100), and only some of the disclosed configurations may constitute the server (100).

[0050] The server (100) may include a processor (110), memory (120), and network (130).

[0051] The processor (110) may be composed of one or more cores and may include processors for data analysis and deep learning, such as a central processing unit (CPU) of a server, a general purpose graphics processing unit (GPGPU), and a tensor processing unit (TPU). The processor (110) may read a computer program stored in memory (120) and perform data processing for machine learning according to one embodiment of the present disclosure. According to one embodiment of the present disclosure, the processor (110) may perform operations for training a neural network model. The processor (110) may perform calculations for training a neural network model, such as processing input data for training in deep learning (DL), extracting features from input data, calculating errors, and updating parameters of the neural network model using backpropagation. At least one of the CPU, GPGPU, and TPU of the processor (110) may process the training of the neural network model. For example, a CPU and a GPGPU can work together to process the training of a neural network model and the classification of data using the neural network model. Additionally, in one embodiment of the present disclosure, processors of a plurality of servers can be used together to process the training of a neural network model and the classification of data using the neural network model. Furthermore, a computer program executed on a server according to one embodiment of the present disclosure may be a CPU, GPGPU, or TPU executable program.

[0052] According to one embodiment of the present disclosure, the memory (120) can store any form of information generated or determined by the processor (110) and any form of information received by the network (130).

[0053] According to one embodiment of the present disclosure, the memory (120) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory), RAM (Random Access Memory), SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, a magnetic disk, and an optical disk. The server (100) may operate in connection with web storage that performs the storage function of the memory (120) on the internet. The description of the memory described above is merely an example and the present disclosure is not limited thereto.

[0054] A network (130) according to one embodiment of the present disclosure may use various wired communication systems such as a Public Switched Telephone Network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL (Very High Speed ​​DSL), UADSL (Universal Asymmetric DSL), HDSL (High Bit Rate DSL), and a Local Area Network (LAN).

[0055] Additionally, the network (130) presented in this disclosure may use various wireless communication systems such as CDMA (Code Division Multi Access), TDMA (Time Division Multi Access), FDMA (Frequency Division Multi Access), OFDMA (Orthogonal Frequency Division Multi Access), SC-FDMA (Single Carrier-FDMA), and other systems.

[0056] In the present disclosure, the network (130) can be configured regardless of the mode of communication, such as wired and wireless, and can be configured as various communication networks, such as a Personal Area Network (PAN) or a Wide Area Network (WAN). In addition, the network may be the known World Wide Web (WWW) and may utilize wireless transmission technologies used for short-range communication, such as Infrared Data Association (IrDA) or Bluetooth. The technologies described in the present disclosure may also be used in other networks mentioned above.

[0057] The server of the present invention may include SAP. SAP stands for "Systems, Applications, and Products in Data Processing" and is an integrated software system designed to manage various business processes of an enterprise and efficiently perform data processing. SAP has particular strengths in large-scale data management and analysis and is widely used in various fields such as supply chain management, human resource management, and financial management.

[0058] By introducing SAP to the server of the present invention, various data generated during the palm fruit supply process can be managed and analyzed in an integrated manner. For example, data such as supplier information, GPS data at the time of shooting, and status-quantity information can be stored in SAP's database system, enabling the efficient generation of reports and payroll information based on this data. Furthermore, by utilizing SAP's advanced data analysis capabilities, the process of calculating payroll based on palm fruit quality and quantity can be further automated, and real-time data analysis can be performed to enhance operational efficiency throughout the supply chain.

[0059] SAP supports multiple users in simultaneously accessing and processing data, and also provides security features that allow access permissions to be set for each department or user. This enables various stakeholders, such as suppliers, managers, and factory staff, to quickly retrieve necessary information and process tasks accordingly. Servers utilizing SAP centrally manage data generated throughout the entire supply chain, thereby supporting the maintenance of data consistency and efficient decision-making.

[0060] Therefore, by including SAP in the system of the present invention, complex data related to the supply management of palm fruits can be more easily integrated and managed, thereby improving the overall efficiency of the supply chain.

[0061]

[0062] FIG. 2 is a drawing illustrating a system according to one embodiment of the present invention.

[0063] FIG. 2 is a diagram illustrating a system according to an embodiment of the present invention. As can be seen in FIG. 2, supplier information (210), GPS information at the time of shooting (220), and first status-quantity information (230) are generated along with a fruit image captured at the supply location through a supplier terminal (200). The information is transmitted to a factory device (300) via a loading location, a means of transport, and an unloading location. At the factory device, the condition of the transported fruit is evaluated to generate second status-quantity information (310), and all of the information is transmitted to a server (100). The server integrates, stores, and manages the information collected from the supplier terminal and the factory device, thereby enabling real-time monitoring of the palm fruit's condition and the generation of a report.

[0064] For example, a supplier terminal (200) photographs the fruit at the palm fruit supply site and generates first status-quantity information (230) along with GPS information (220) at the time of photography. Subsequently, the supplied fruit is loaded onto a transport vehicle at the loading site and moved to the factory's unloading site via the transport vehicle, and the factory device (300) analyzes the condition of the fruit again to generate second status-quantity information (310). The server (100) integrates all of this information to evaluate the quality and quantity of the fruit, and based on this, generates an appropriate report for the supplier and factory manager, and can provide real-time status monitoring and analysis results.

[0065] The process of a terminal and a factory device according to the present invention photographing and analyzing a fruit includes a procedure of generating a bounding box containing the fruit using an artificial intelligence model and classifying objects within the bounding box. First, a first image data is generated by photographing a fruit at a supplier terminal. At this time, an artificial intelligence model is used to automatically detect the fruit in the image and form a bounding box surrounding each fruit. For example, if there are multiple fruits in a single image, the artificial intelligence model distinguishes each fruit individually and marks them as bounding boxes.

[0066] When a bounding box is generated, the condition of the fruit is classified by analyzing the size of the fruit and the size of the spines contained within it. By comparing the size of the palm fruit and the size of the spines, it is determined whether the fruit is overripe, suitable, or overgrown; conversely, if the size is smaller than a preset threshold, it is classified as semi-ripe, unripe, or poorly moist. Through this process, first condition-quantity information collected from the supply source is generated.

[0067] The same analysis is performed when the fruit is photographed again at the factory to generate second image data. The factory equipment photographs the transported fruit to create a bounding box and re-analyzes the condition of the fruit contained within it. At this stage as well, an artificial intelligence model accurately determines the condition and quantity of the fruit to generate second condition-quantity information. This enables the collection of more precise data regarding the condition of the palm fruit.

[0068] Examples of artificial intelligence models that can be used in the present invention include Convolutional Neural Networks (CNN), which are excellent for image classification, or models such as YOLO (You Only Look Once) and Faster R-CNN, which are widely used for object detection. These models can be used to quickly detect the location of fruits within an image, form bounding boxes, and efficiently classify the state of each fruit. Additionally, to analyze the state of the fruits, a pre-trained neural network such as ResNet or EfficientNet can be applied to derive more accurate results.

[0069] The reason for identifying GPS information at the time of shooting is to accurately record the location of the supply site where the fruit was photographed, thereby automatically matching this information with the means of transport and the supplier's terminal. This allows for the automatic assignment of the fruit supply site location or ID, enhancing the accuracy and efficiency of management. In other words, based on the GPS information contained in the image data captured by the supplier's terminal, the system can automatically recognize which supply site the fruit was harvested from without any manual intervention.

[0070] In conventional methods, this task was primarily performed manually, requiring the input of supplier information using RFID card readers and RFID cards. The method of recording supplier information by having the supplier insert or tag an RFID card required additional equipment and processing time, and also carried the risk of error. However, in this invention, by utilizing GPS information at the time of capture, supplier information can be automatically matched without a separate RFID system, making supply chain management much simpler and more efficient.

[0071] Meanwhile, as an alternative embodiment, the present invention may provide a method for transmitting status and quantity data of palm fruits by considering two situations depending on the data communication environment.

[0072] First, if communication is impossible, the first image data, GPS information (220), and first status-quantity information (230) generated by the supplier terminal (200) are temporarily stored in a storage device. Subsequently, when the fruit is loaded onto a transport vehicle and moved to the factory, the data can be transmitted to the factory device (300). If communication is also impossible at the factory, the data transported with the fruit is transmitted in an office environment where a wired or wireless network is available after arriving at the factory. This allows data to be safely collected and transmitted even at a supply location without a network connection.

[0073] Secondly, if communication is possible, the supplier terminal (200) can generate first image data, GPS information (220), and first status-quantity information (230), and then transmit the data in real time to a designated data server or cloud via a wireless network. In this case, the status and quantity information of the fruit can be managed in an integrated manner in real time at a central server rather than at the supplier, and the supplier and the factory manager can immediately check and analyze the data.

[0074] That is, according to embodiments of the present invention, a flexible system capable of stably collecting and transmitting data regardless of the communication environment can be provided.

[0075]

[0076] FIG. 3 is a diagram illustrating a method for identifying each supplier location when there are multiple suppliers according to an embodiment of the present invention.

[0077] Referring to FIG. 3, in the present invention, when a supplier terminal photographs a fruit, GPS information at the time of photography is automatically recorded to identify the location of each supply location. Through this, even when multiple supply locations exist, the location information of the fruit harvested at each supply location is automatically matched. This allows the supply locations of the fruit to be clearly distinguished without the intervention of the supplier, thereby increasing the accuracy of information processing.

[0078] For example, fruit harvested at the first supply site is photographed by the first supplier's terminal, and the first GPS information is automatically recorded. Similarly, fruit is photographed at the second, third, and fourth supply sites through their respective supplier terminals, and the second, third, and fourth GPS information is generated. At this time, each terminal automatically checks the GPS at the time of shooting and transmits the location or supply site ID of the corresponding supply site to the server without any separate operation.

[0079] Conventionally, this task had to be handled manually, and suppliers entered supply location information through RFID card readers. Each supplier matched supply location information by tagging an RFID card installed at the supply location to a reader; however, this method required additional equipment and time, and there was a possibility of errors. In contrast, in the present invention, since GPS information is automatically matched when the supplier photographs the fruit, a separate RFID card or reader is not required, and supply location information can be managed more quickly and accurately.

[0080] This invention can be particularly useful in countries where high-performance smartphones equipped with RFID reader functions are not commonly used. Conventional systems required the use of RFID cards and RFID readers to record supplier information, which limited their use in general smartphones without RFID reader capabilities. However, this invention can replace RFID systems by automatically recording GPS information at the time of shooting when the supplier terminal photographs the fruit.

[0081] This method can be easily applied even to standard smartphones, not just high-performance ones, and provides both cost reduction and ease of management because it does not require separate RFID equipment or readers. For example, if a supplier simply photographs a fruit with a terminal, GPS information is automatically recorded, allowing the supplier's location information to be accurately transmitted to a server. Therefore, even in countries or regions without RFID capabilities, the system of this invention enables effective management of location information.

[0082]

[0083] FIG. 4 is a diagram showing a situation in which a plurality of suppliers or a plurality of RFID cards are delivered to a factory device according to one embodiment of the present invention.

[0084] Referring to FIG. 4, of course, the supplier terminals of the present invention can transmit information to the factory device using a wireless network. However, in environments or countries where wireless networks are not readily available, information can be transmitted by the suppliers directly delivering the terminals themselves to the factory device instead of wireless transmission. This is a method that can ensure efficient data transmission even in areas where network infrastructure is limited.

[0085] For example, assume that multiple suppliers bring supplier terminals or RFID cards containing fruit images taken at their respective supply sites, GPS data, and status-quantity information to the factory. Each supplier inserts or connects their terminal or RFID card to a factory device to transmit the data stored on the terminal or card to the factory device. The factory device collects the data transmitted from the multiple supplier terminals or RFID cards and processes the data for each supplier.

[0086] In a specific embodiment, the first supplier uploads data by connecting information stored in its terminal to the factory device, and the second supplier can perform the same operation using an RFID card. The factory device recognizes the supplier information, GPS information at the time of shooting, and first status-quantity information stored in each terminal or RFID card, respectively, and processes them by integrating them. Through this, information on the fruit harvested at each supply site can be accurately transmitted to the factory regardless of the network environment.

[0087] This method can be effectively used even in countries or regions with unstable wireless networks, and allows for the secure integration of data into factory systems by directly delivering terminals or RFID cards.

[0088]

[0089] FIG. 5 is a diagram illustrating a method for generating reports and payroll information according to an embodiment of the present invention. In the system of the present invention, the server (100) collects various information from the supplier terminal and the factory device, comprehensively analyzes it, and then automatically generates reports and payroll information according to the supplier's fruit status and quantity.

[0090] The server first receives supplier information (210), GPS information at the time of shooting (220), and first state-quantity information (230) from the supplier terminal (200). These information includes image data generated by the supplier photographing palm fruit at the supply site and the results of artificial intelligence analysis thereon. Along with supplier information including the supplier's name, ID, and location information where the fruit was harvested, the first state-quantity information is data generated by an artificial intelligence model analyzing the state of the fruit (e.g., suitable state, unripe state, overripe state, etc.). This information is transmitted to the server and combined with additional data from the factory device.

[0091] The second state-quantity information (310) collected from the factory device (300) includes the results of analyzing the state and quantity of palm fruit captured in the factory. Like the first state-quantity information, the second state-quantity information is also the result of an artificial intelligence model analyzing image data captured while the fruit is being transported via a conveyor belt in the factory. The server can compare these two state-quantity informations to check for a match, and if they match, ensemble them to generate a third state-quantity information. This ensemble data provides a more accurate evaluation of the fruit's state and quantity, allowing for a more reliable evaluation of the fruit's quality.

[0092] The server generates a report based on the collected status-quantity information. This report includes supplier information as well as detailed evaluations based on the condition of the fruit. For example, the report specifies the percentage of fruit deemed suitable out of the total quantity supplied by the supplier. This information plays a crucial role for suppliers and managers in evaluating the quality of palm fruit and reviewing the efficiency of the supply process.

[0093] In addition, the server plays an important role in determining the supplier's pay information. In the pay report according to the present invention, the pay is determined in proportion to the quality of the supplied fruit, particularly the proportion of fruit judged to be in a suitable state. The higher the quantity of suitable fruit relative to the total quantity of supplied fruit, the more pay the supplier receives. For example, if a first supplier supplies 1,000 fruits and 800 of them are classified as suitable, the ratio of suitable quantity becomes 80%. Based on this ratio, the server calculates the pay of the corresponding supplier, and the supplier who supplied high-quality fruit receives a corresponding reward.

[0094] This method provides an incentive that encourages suppliers to prioritize fruit quality management. In other words, because supplying high-quality fruit can lead to higher pay than simply supplying a large quantity, suppliers focus more on quality control. The server automatically performs these pay calculations, and the pay report includes not only the amount the supplier will receive but also a detailed explanation of how the pay was determined. For example, the report may specify the ratio of the appropriate quantity based on the first state-quantity information and the second state-quantity information, as well as the pay calculation process based on that ratio.

[0095] Meanwhile, even if palm fruits are harvested normally and loaded at the supply site, they may naturally ripen during the supply process and turn into overripe fruits that are not easy to process. This is a case that can occur in just a few hours, so the feed calculation process may not be reasonable by ensembling first-state-yield information with second-state-yield information.

[0096] Accordingly, alternative to one embodiment related to the aforementioned ensemble, the system can generate an appropriate pay report related to the 'harvester' based on information from the supply source (first state-quantity information) or information from the factory (second state-quantity information).

[0097] For example, after a harvester harvests palm fruit, they use a supplier terminal (200) to capture an image of the fruit and generate first image data. This first image data is analyzed through a first artificial intelligence model, and the quantity of palm fruit in a suitable state can be calculated as 95% of the total harvest quantity. This first state-quantity information (230) can be transmitted to a server (100) and stored in a DB.

[0098] Furthermore, the system can evaluate the harvester's work performance based on this information and generate a corresponding pay report. For example, a higher percentage of suitable palm fruits indicates superior harvesting skills and quality control capabilities, so incentives can be provided when calculating the pay. Conversely, if the percentage of suitable fruits is low, additional training or feedback can be provided to improve the harvester's capabilities.

[0099] In addition, if both the first state-quantity information and the second state-quantity information exist, an appropriate payroll report related to the 'carrier' can be generated based on the error between the first state-quantity information and the second state-quantity information.

[0100] For example, if the percentage of suitable fruit measured at the supply point was 95% but the percentage measured at the factory was 85%, the error is 10%. This suggests that the quality of the fruit deteriorated during transportation. The system described above can calculate wages by reflecting this error in the transporter's performance evaluation. Since a smaller error indicates that the transporter contributed to maintaining quality by transporting the fruit quickly and safely, incentives can be provided accordingly. Conversely, if the error is large, problems in the transportation process can be analyzed and improvement measures can be suggested.

[0101] For example, since the quality of the fruit may deteriorate due to exposure to high temperatures during transport, the system can take measures to educate the transporter on appropriate storage and transportation methods or provide a transport vehicle equipped with refrigeration facilities. Additionally, by utilizing GPS information to track the transporter's route and time, it can identify whether there were any unnecessary delays or detours and improve transportation efficiency.

[0102] Consequently, the present invention significantly improves the efficiency of supply chain management through a system that automatically evaluates the quality and quantity of palm fruits and provides fair compensation to suppliers based on this evaluation. This system eliminates the inconvenience of manual calculation or recording and enables suppliers and managers to make decisions based on reliable data.

[0103]

[0104] FIG. 6 is a diagram showing a program configuration that can be output to a supplier terminal according to one embodiment of the present invention.

[0105] In FIG. 6, various information input through the supplier terminal (200) and data processing and management functions performed based on said information are clearly explained.

[0106] The supplier terminal receives and processes various information from the supplier, and for this purpose, provides an application or web-based program interface. This interface includes functions such as supplier information input, conditional branching, data generation, data modification, and report viewing. The user can input information such as the supplier name, supplier ID, and means of transport through the terminal. This supplier information (210) is important data necessary for recording and managing the location and status of the heat supply, and is used to track the movement and quality of the heat supply in the server and factory device.

[0107] The program interface of the terminal enables user interaction in the form of a data generation button, a data modification button, and a report inquiry button. When the data generation button is pressed, the supplier photographs the fruit at the supply location, and the captured image is stored as the first image data. At this time, GPS information (220) regarding the photographed fruit is also generated simultaneously and stored as data associated with the fruit's supply location. Subsequently, the first artificial intelligence model analyzes this image data to generate a bounding box containing the fruit, and analyzes the condition and quantity of the fruit within the bounding box to generate the first condition-quantity information (230).

[0108] If a user performs a modification operation via the data modification button, new information can be entered or updated based on the already stored first state-quantity information. During this process, the system provides the functionality to retrieve previously entered data or modify necessary parts, and allows for the modification of selected image data to save the updated information.

[0109] Additionally, through the report inquiry button, a report generated on the server based on the first status-quantity information (230) collected from multiple suppliers can be viewed. This report includes information on the overall status and quantity of the supplied fruit and comprehensively shows the quality evaluation and quantity of the fruit supplied by each supplier. This plays an important role in managing the quality of the fruit supplied by the suppliers and reviewing performance during the supply process.

[0110] Consequently, the supplier terminal of the present invention provides various data processing and management functions, enabling users to generate, modify, and view data in real time, thereby allowing for the efficient management of the quality and quantity of palm fruits throughout the supply chain. In particular, this system can increase supplier productivity and contribute to fair quality assessment and payroll calculation.

[0111]

[0112] FIG. 7 is a drawing for explaining state-quantity according to one embodiment of the present invention.

[0113] Referring to FIG. 7, the present invention analyzes images of palm fruits to automatically classify their condition into overripe, suitable, semi-ripe, unripe, rotten, poor moisture, etc., and calculates the yield according to each condition. In this process, a first artificial intelligence model receives captured fruit image data as input, detects the fruits, and performs analysis by generating a bounding box around the detected fruits. The bounding box displays not only the condition of each fruit but also the predicted probability of that condition. As a result, the user can clearly confirm the probability that the fruit belongs to a specific condition, and the overall quality evaluation is performed more reliably.

[0114] For example, let's assume that a first AI model detects 10 fruits in an image of palm fruits. A bounding box is generated around each fruit, and the fruit's state and probability are displayed inside it. The state displayed inside the bounding box represents the state of the fruit predicted by the AI ​​model and may appear in a format such as, for example, "Fit state (0.82)" or "Overripe state (0.76)". Here, "Fit state (0.82)" means that there is an 82% probability that the fruit is in a fit state, and "Overripe state (0.76)" indicates that there is a 76% probability that it is in an overripe state.

[0115] The first artificial intelligence model analyzes the condition based on fruit size and spine size. Spine size refers to the length and density of the fibers visible on the outside of the fruit and serves as an important indicator for determining fruit maturity. By comparing fruit size and spine size, if the fruit size exceeds a preset multiple of the spine size, the fruit is classified as overripe, suitable, or rotten. Overripe refers to a state where the fruit is too ripe and oil quality may deteriorate, while suitable indicates that the fruit is in the most optimal state of maturity for harvesting. Rotten is classified as a state where the fruit has rotted and can no longer be used.

[0116] Conversely, if the fruit size is less than a preset multiple of the spine size, the fruit is classified as semi-ripe, unripe, or poorly pollinated. Semi-ripe refers to a state where the fruit is under-ripe but maintains a certain level of quality, while unripe indicates a case where the fruit fails to meet harvest standards and its quality is inferior. Poorly pollinated means that a defect occurred because pollination did not take place in the flower.

[0117] As a specific example, let's assume that a particular palm fruit is predicted to be in a fit state. In this case, the AI ​​model analyzes the size of the fruit and its spines to classify it as fit, and "Fit State (0.85)" is displayed inside the bounding box. This indicates that there is an 85% probability that the fruit is in a fit state. If another fruit is predicted to be overripe, it is displayed as "Overripe State (0.75)" inside the bounding box, providing information that there is a 75% probability that the fruit is overripe. Through these probability displays, users can intuitively understand how reliable the condition of each fruit is.

[0118] In addition, the system automatically calculates the predicted quantity for each condition for multiple fruits. For example, if 10 fruits are detected in an image and classified as 5 suitable, 3 overripe, 1 semi-ripe, and 1 unripe, the quantity for each condition is automatically recorded as 5 suitable, 3 overripe, 1 semi-ripe, and 1 unripe. This information is transmitted to a server and used by suppliers and factories to evaluate and manage the quality of the fruits in real time.

[0119] The system of the present invention significantly improves the efficiency of palm fruit quality control and supply chain management by rapidly and accurately classifying the condition of palm fruits through such automated analysis and calculating the corresponding quantity. Furthermore, the prediction probability displayed in the bounding box allows the user to verify the accuracy of the fruit condition prediction, thereby enabling a more reliable quality assessment.

[0120]

[0121] Additional embodiments of the present invention are described.

[0122] In one embodiment, the server of the present invention compares the first state-quantity information (230) and the second state-quantity information (310), and if the difference is less than or equal to a preset value, combines (ensembles) the two information to generate third state-quantity information. For example, assuming that the first information collected from the supplier terminal is analyzed to have 600 suitable state heats and the second information collected from the factory device is analyzed to have 580 suitable state heats, if this difference is within a preset allowable range, the server averages them to calculate the quantity of suitable state heats as 590. Based on this ensembled information, the server generates a final report and provides it to the supplier.

[0123] In the present invention, "ensemble" refers to a process of combining two or more data to derive more reliable final data. For example, when first state-quantity information (230) and second state-quantity information (310) are collected from a supplier terminal and a factory device, respectively, the server compares these two data and, if the difference is within an acceptable range, combines them to generate third state-quantity information. At this time, an ensemble technique is applied to derive a final result by simply averaging the two data or by prioritizing the more reliable data.

[0124] In one embodiment of the present invention, the process includes a server collecting first image data and second image data, comparing first state-quantity information generated based thereon with second state-quantity information, and transmitting data to an inspector terminal if the difference exceeds a preset allowable range. The purpose of this embodiment is to make a more accurate judgment through the intervention of an inspector when automated analysis is inaccurate or inconsistent. For example, if a fruit captured by a supplier terminal is classified as suitable but analyzed as semi-ripe by a factory device, and this difference is greater than the allowable range, the inspector directly verifies the first and second image data to generate fourth state-quantity information. This increases the reliability of the system's automated analysis results, reduces errors, and ultimately enables accurate quality control.

[0125] In the following embodiment, a process of additionally training an artificial intelligence model using fourth state-quantity information reflecting the inspector's judgment is included. The purpose of this embodiment is to continuously improve the performance of the artificial intelligence model based on data provided by the inspector. For example, when the inspector newly determines a suitable fruit quantity from the fourth state-quantity information, the server trains the first artificial intelligence model and the second artificial intelligence model with this data along with the first image data and the second image data. As a result, the artificial intelligence model learns new patterns and rules, enabling it to make more accurate predictions in the same situation thereafter. Through this, the analysis accuracy of the artificial intelligence model is continuously improved, and the automatic analysis results of the system become more reliable.

[0126] In another embodiment, the process includes a server calculating the difference between the fourth state-quantity information and the existing first state-quantity information and second state-quantity information, and further training an artificial intelligence model based on the one with the larger difference. The purpose of this embodiment is for the system to determine which artificial intelligence model requires more training based on inconsistent data and to proceed with efficient training. For example, if the difference between the first state-quantity information and the fourth state-quantity information is large, it means that the first artificial intelligence model requires additional training, and the server retrains and improves the first artificial intelligence model. By training the artificial intelligence model based on the data with the larger difference in this way, the learning efficiency of the system can be increased, and the accuracy of the analysis results can be further enhanced.

[0127] These embodiments contribute to improving the accuracy and efficiency of the system for each respective purpose and providing high reliability in palm fruit condition analysis and quantity evaluation.

[0128] In an additional embodiment, the inspector uses an inspector terminal to verify the first and second image data, and then generates new fourth state-quantity information through direct judgment. Depending on the inspector's judgment, the state of the heat transfer fluid may be changed, and the fourth state-quantity information entered by the inspector is transmitted back to the server. The server compares this fourth state-quantity information with the existing first and second state-quantity information to calculate the quantity difference for each state.

[0129] The server retrains the artificial intelligence model based on the calculated difference. For example, if the difference between the fourth state-quantity information and the first state-quantity information is large, the server further trains the first artificial intelligence model; conversely, if the difference with the second state-quantity information is large, the server retrains the second artificial intelligence model. However, if both differences exceed the allowable range, the server retrains both models, but prioritizes training on the one with the larger difference.

[0130] Although preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above. Various modifications are possible by those skilled in the art without departing from the essence of the invention as claimed in the patent claims, and such modifications should not be understood individually from the technical spirit or perspective of the present invention.

Claims

1. As a system for determining the condition of palm fruits, Server(100); Provider terminal (200); and Factory device (300); Includes, The above supplier terminal (200) is, Receive supplier information (210) from the supplier, The fruit is photographed at the supply site to generate the first image data, Generate GPS information (220) of the fruit shooting time at the above supply site, and Based on the above first image data, a first artificial intelligence model is used to generate first state-quantity information (230), and The above factory device (300) is, The supplier information (210); GPS information (220); and the first state-quantity information (230) are obtained from the supplier terminal (200), and The above fruit transported by a conveyor belt is photographed to generate second image data, Based on the above second image data, a second artificial intelligence model is used to generate second state-quantity information (310), and The above server (100) is, Acquiring the above supplier information (210); the above shooting time GPS information (220); the above first state-quantity information (230); and the second state-quantity information (310) and storing them in a DB A system characterized by 2. In Paragraph 1, The above server (100) is, If the above first state-quantity information (230) or the above second state-quantity information (310) exists, A feed report for a harvester is generated based on at least one of the first state-quantity information (230) or the second state-quantity information (310), and When both the first state-quantity information (230) and the second state-quantity information (310) exist, Calculate the error between the first state-quantity information (230) and the second state-quantity information (310) to generate a pay report for the transporter, The payroll report for the above-mentioned transporter includes a predetermined incentive inversely proportional to the above error. A system characterized by 3. In Paragraph 1, The above server (100) is, When comparing the first state-quantity information (230) and the second state-quantity information (310), if the difference is less than or equal to a predetermined value, The first state-quantity information (230); and the second state-quantity information (310) are ensembled to generate third state-quantity information, and Generating a report based on at least one third-state quantity information A system characterized by 4. In Paragraph 2 or 3, The above report is, It includes a salary report corresponding to the above supplier information (210), and The above salary report is, Salary information determined in proportion to the proportion of the suitable quantity relative to the total quantity obtained based on at least one of the first state-quantity information (230), the second state-quantity information (310), or the third state-quantity information, System.

5. In Paragraph 1, The above supplier terminal (200) is, Receive supplier name information, supplier ID information, and supplier transportation information from the above supplier, and Based on an application or the web, output a data creation button, a data modification button, and a report view button, and When the above data generation button is interacted with, A fruit is photographed at the supply site to generate first image data, and Generates GPS information (220) of the fruit shooting time at the above supply site, Based on the first image data, at least one bounding box is generated in the area corresponding to the fruit using the first artificial intelligence model, and Predicting the first state-quantity information (230) using the first artificial intelligence model based on the above at least one bounding box, and Based on the above first image data, a first artificial intelligence model is used to generate first state-quantity information (230), and When the above data modification button is interacted with, The first state-quantity information (230) is modified and updated from the above supplier, When the above report view button is interacted with, Outputting a report generated based on multiple first state-quantity information (230) obtained from a terminal from multiple suppliers. A system characterized by 6. In Paragraph 1, The above-mentioned first artificial intelligence model is, The above first image data is received, and Based on the first image data above, at least one bounding box is created in an area corresponding to a palm fruit, and Predicting the size of the palm fruit and the size of the thorns included in each of the above at least one bounding box, and When comparing the size of the palm fruit and the size of the thorn, if the size of the palm fruit is greater than or equal to a predetermined multiple of the size of the thorn, The above palm fruit is classified into one of the following states: overripe, suitable, or overripe. If, when comparing the size of the palm fruit and the size of the spine, the size of the palm fruit is less than the predetermined multiple of the size of the spine, Classifying the above palm fruit into one of the following: semi-ripe, unripe, or poor moisture state. A system characterized by 7. In Paragraph 1, The above server (100) is, Acquiring the first image data; and the second image data, When comparing the first state-quantity information (230) and the second state-quantity information (310), if the difference is greater than or equal to a predetermined value, The first image data; and the second image data are transmitted to the inspector terminal, and The inspector obtains the fourth state-quantity information generated by the inspector based on the first image data and the second image data from the inspector terminal. A system characterized by 8. In Paragraph 7, The above server (100) is, The first artificial intelligence model is further trained based on the above-mentioned fourth state-quantity information and the above-mentioned first image data, and Further training the second artificial intelligence model based on the fourth state-quantity information and the second image data. A system characterized by 9. In Paragraph 7, The above server (100) is, Calculate the first difference, which is the difference between the above-mentioned fourth state-quantity information and the above-mentioned first state-quantity information (230), and Calculate the second difference, which is the difference between the fourth state-quantity information and the second state-quantity information (310), and When comparing the first difference and the second difference, if the first difference is larger, The first artificial intelligence model is further trained based on the above-mentioned fourth state-quantity information and the above-mentioned first image data, and When comparing the first difference and the second difference, if the second difference is larger, Further training the second artificial intelligence model based on the fourth state-quantity information and the second image data. A method characterized by 10. In Paragraph 1, The above supplier terminal (200) is, If communication with the server (100) associated with the above factory device (300) is impossible, The first image data, the GPS information (220), and the first state-quantity information (230) are temporarily stored in a storage device, and after the fruit is transported to a factory device (300) by a transport means, the information is transmitted via a wired or wireless network. When communication with the server (100) associated with the above factory device (300) is possible, the first image data, the GPS information (220), and the first status-quantity information (230) are transmitted in real time to the server (100) or a designated cloud server via a wireless network. A system characterized by

Images