Automated systems and associated methods for grading and sorting palm oil fresh fruit bunches

Abstract

of the Disclosure Automated systems and associated methods for grading and sorting palm oil fresh fruit bunches (FFBs) are provided. A plurality of FFBs are received, and one or more quality indicator characteristics of the plurality of FFBs are determined by performing analysis on the plurality of FFBs. The analysis comprises at least one of a color and thermal imaging analysis, a reflection analysis, and a weight measurement analysis. Further the plurality of FFBs are categorized into a number of groups based on the determined one or more quality indicator characteristics. Moreover, the plurality of FFBs are sorted based on the categorization, and a sterilization recipe is determined for the FFBs based on the categorization.

Description

AUTOMATED SYSTEMS AND ASSOCIATED METHODS FOR GRADING AND SORTING PALM OIL FRESH FRUIT BUNCHESField of the invention

[0001] The present invention relates generally to automation systems and, more specifically, to automated systems and associated methods for grading and sorting palm oil fresh fruit bunches (FFBs).Background of the invention

[0002] Palm oil industry is a cornerstone of global agriculture, providing a versatile and widely used edible oil. Within the palm oil industry, palm oil mills serve as vital hubs where the processing of Fresh Fruit Bunches (FFBs) takes place, shaping the quality and efficiency of palm oil production. However, traditional practices in grading the FFBs pose significant challenges to the optimal functioning of the palm oil mills.

[0003] Conventional methods of manual grading of FFBs, which is reliant on visual inspection of outer color of the FFBs, are riddled with inaccuracies. This manual approach not only consumes substantial labor but also leads to delays in the grading process, contributing to inefficient use of time. Further, the conventional method of manual grading lacks accurate data on FFB count, weight, ripeness levels, and bad FFB return rates, which hampers effective monitoring and process improvement across palm oil mills. This deficiency in data not only impedes strategic decisionmaking but also limits the ability to implement targeted improvements in the oil extraction operations.

[0004] Furthermore, storage practices in many palm oil mills exacerbate the above challenges, with the FFBs often piled up outdoors while awaiting sterilizer loads. Exposure to environmental elements, such as rainwater, raises concerns about the potential degradation of the FFBs, thereby impacting the quality of the FFBs and Oil Extraction Rates (OERs).

[0005] Moreover, the absence of information on FFB quality parameters as the FFBs enter the sterilizer station creates uncertainties in setting optimal processing parameters. Parameters such as ripeness level, Free Fatty Acid (FFA), Peroxide Value (PV), and Deterioration of Bleachability Index (DOBI) are crucial for efficient oil extraction and setting the optimal processing parameters, and the lack of this information compromises the overall oil extraction process. The absence of information on the FFB quality parameters also leads to challenges in ensuring fair compensation for the smallholder farmers.

[0006] Accordingly, there is a need for a solution to the aforementioned problems. For example, there exists a need for an automated grading and sorting system and method for FFBs. Moreover, there is a need for system in which the FFBs are treated for oil extraction based on the automated grading and sorting.Summary of the Invention

[0007] According to an implementation of the present disclosure, there is provided an automated grading and sorting system. The automated grading and sorting system comprises an input module configured to receive a plurality of fresh fruit bunches (FFBs), and a grading module, operatively coupled to the input module, configured to determine one or more quality indicator characteristics of the plurality of FFBs by performing analysis on the plurality of FFBs, wherein the analysis comprises at least one of color and thermal imaging analysis, reflection analysis, and weight measurement analysis, and categorize the plurality of FFBs into a number of groups based at least on the determined one or more quality indicator characteristics. The automated grading and sorting system also comprises a sorting module, operatively coupled to the grading module, configured to sort the plurality of FFBs based on the categorization.

[0008] In an aspect, the grading module may be configured to obtain user-defined profile data associated with the categorization of the FFBs; compare the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data; and based at least upon the comparison, categorize the plurality of FFBs into the number of groups.

[0009] In an aspect, to obtain user-defined profile data, the grading module may be configured to: obtain a training set, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs; obtain one or more quality indicator characteristics of the test group of FFBs, wherein the one or more quality indicator characteristics are obtained from a user based on manual inspection and lab reports for the test group of FFBs; and determine the user-defined profile data based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results.

[0010] In an aspect, to determine the user-defined profile data, the grading module may be configured to apply the obtained training set and the obtained one or more quality indicator characteristics to a machine learning algorithm to determine the acceptable range.

[0011] In an aspect, the automated grading and sorting system may further comprise a storage module that comprises a plurality of bins, wherein the storage module may be coupled to the sorting module, and configured to store each of the plurality of FFBs in a respective one of the plurality of bins based on the categorization of the plurality of FFBs, and wherein each of the one or more bins corresponds to a respective one of the number of groups.

[0012] In an aspect, the grading module may be configured to employ a machine learning algorithm to determine the one or more quality indicator characteristics.

[0013] In an aspect, the grading module may be further configured to employ the machine learning algorithm to determine a respective sterilization recipe for each of the number of groups, wherein the sterilization recipe comprises at least one of a time value, a pressure value, and a temperature value, which may be employed as process parameters in a sterilization center, to maximize oil extraction rate (OER) from the plurality of FFBs.

[0014] In an aspect, the grading module may be configured to calculate at least one value that reflects at least one quality indicator characteristic of each of the plurality of FFBs based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis.

[0015] In an aspect, the at least one value may comprise one or more of: a first value that indicates a ripeness of the FFB, a second value that indicates an oil quality of the FFB, and a third value that indicates a weight of the FFB.

[0016] In an aspect, the grading module may be configured to calculate the first value and the second value from the color and thermal imaging analysis, and / or the reflection analysis, wherein the grading module may be configured to calculate the third value from the weight measurement analysis.

[0017] In an aspect, the grading module may comprise a first imaging module that captures visible color and thermal images of the plurality of FFBs to perform the color and thermal imaging analysis of the plurality of FFBs; a second imaging module that captures a reflection response of the plurality of FFBs to illuminated light from the second imaging module to perform the reflection analysis, wherein the second imaging module comprises at least one of: a hyperspectral imaging device and a near-infrared (NIR) imaging device; and a weight load cell that measures a weight of each of the plurality of FFBs.

[0018] In an aspect, the input module may comprise a conveyor unit for automated reception of the plurality of FFBs from one or more sources, wherein the one or more sources may comprise vehicles carrying the plurality of FFBs.

[0019] In an aspect, the one or more quality indicator characteristics may comprise at least one of: a ripeness level, an aesthetic condition, one or more diseases associated with the plurality of FFBs, an oil quality of the plurality of FFBs, a brix percentage, species, and firmness of the plurality of FFBs.

[0020] In an aspect, the storage module may further comprise an unloading feature that allows for unloading of the plurality of FFBs from the plurality of bins to a loading ramp station.

[0021] In an aspect, the one or more quality indicator characteristics may comprise an oil quality of the FFB, wherein the grading module may be configured to predict the oil quality of the FFB based on one or more oil quality parameters that may comprise Free Fatty Acid (FFA), peroxide value (PV), Deterioration of Bleachability Index (DOBI), and Carotene.

[0022] According to another implementation of the present disclosure, there is provided an automated method for grading and sorting fresh fruit bunches (FFBs). The method comprises receiving, by an automated system, a plurality of FFBs, and determining, by the automated system, one or more quality indicator characteristics of the plurality of FFBs by performing analysis on the plurality of FFBs. The analysis comprises at least one of: color and thermal imaging analysis, reflection analysis, and weight measurement analysis. The method also comprises categorizing, by the automated system, the plurality of FFBs into a number of groups based on the determined one or more quality indicator characteristics.

[0023] According to yet another implementation of the present disclosure, there is provided a computer-implemented method for handling Fresh Fruit Bunches (FFBs). The method comprises determining one or more quality indicator characteristics of a plurality of FFBs, and receiving user-defined profile data associated with categorization of the FFBs. The method also comprises comparing the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data, and based at least upon the comparison, categorizing the plurality of FFBs into a number of groups.

[0024] In an aspect, determining the one or more quality indicator characteristics may comprise determining one or more of: species, ripeness level, aesthetic condition, diseases, oil quality, brix percentage, and firmness of the plurality of FFBs.

[0025] In an aspect, determining the ripeness level may comprise determining whether the FFBs are ripe, over ripe, under ripe, unripe, rotten, or empty.

[0026] In an aspect, determining the aesthetic condition may comprise determining one or more of: dirtiness, size, color, and weight of the plurality of FFBs.

[0027] In an aspect, determining the diseases may comprise determining one or more of: defect and damage in the FFBs, and / or determining whether the FFBs are long stalk FFBs and / or poorly pollinated FFBs.

[0028] In an aspect, determining the oil quality may comprise determining one or more of oil content, Free Fatty Acid (FFA), peroxide value (PV), Deterioration of Bleachability Index (DOB1), and Carotene of the plurality of FFBs.

[0029] In an aspect, determining the one or more quality indicator characteristics may comprise performing one or more of image analysis, reflection analysis, and weight analysis on the plurality of FFBs.

[0030] In an aspect, performing the image analysis may comprise performing at least one of: color image analysis and thermal imaging analysis.

[0031] In an aspect, performing the color image analysis may comprise analyzing one or more of: color range, hue, color variance, appearance, color coverage, color coverage group of fruitlets on the FFBs.

[0032] In an aspect, performing the color image analysis may comprise determining at least some information about the one or more quality indicator characteristics including one or more of a ripeness level and a size of the plurality of FFBs.

[0033] In an aspect, performing the thermal image analysis may comprise performing a comparison between a temperature profile of the FFBs, and atmospheric temperature.

[0034] In an aspect, performing the thermal imaging analysis may comprise determining at least some information about the one or more quality indicator characteristics including one or more of oil content, Free Fatty Acid (FFA), peroxide value (PV), Deterioration of Bleachability Index (DOB I) and Carotene of the plurality of FFBs.

[0035] In an aspect, performing the reflection analysis may comprise analyzing one or more of: hyperspectral and Near Infrared (NIR) illumination reflected from the FFBs.

[0036] In an aspect, performing the reflection analysis may comprise determining at least some information about the one or more quality indicator characteristics that comprises one or more of ripeness level, BRIX percentage, firmness, defect, damage, total percentage of oil content and FreeFatty Acid (FFA), peroxide value (PV), Deterioration of Bleachability Index (DOBI), and Carotene of the plurality of FFBs.

[0037] In an aspect, performing the weight analysis may comprise determining a weight of individual FFBs of the plurality of FFBs.

[0038] In an aspect, receiving the user-defined profile data may comprise: receiving a training set, wherein the training set may comprise at least one of color images, thermal images, and spectral images for a test group of FFBs, receiving one or more quality indicator characteristics of the test group of FFBs, wherein the one or more quality indicator characteristics may be received from a user based on manual inspection and lab reports for the test group of FFBs; and determining the user-defined profile data based on the received training set and the received one or more characteristics, wherein the user-defined profile data may comprise an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results.

[0039] Tn an aspect, determining the user-defined profile data may comprise applying the received training set and the received one or more quality indicator characteristics to a machine learning algorithm to determine the acceptable range.

[0040] In an aspect, the method may further comprise sorting the plurality of FFBs based at least upon the categorization.

[0041] According to yet another implementation of the present disclosure, there is provided a non-transitory computer-readable storage medium comprising instructions, which in response to execution by one or more processors cause the one or more processors to perform or control performance of any of the methods disclosed herein.

[0042] The automated grading and sorting system and associated methods of the present disclosure overcome one or more of the shortcomings of the prior art. Additional features and advantages may be realized through the techniques of the present disclosure. Other embodiments and aspects of the disclosure are described in detail herein and are considered a part of the claimed disclosure.

[0043] These and other objects, features, and advantages of the present invention will become more readily apparent from the attached drawings and the detailed description of the preferred embodiments, which follow.Brief Description of the Drawings

[0044] The preferred embodiments of the invention will hereinafter be described in conjunction with the appended drawings provided to illustrate and not to limit the invention, where like designations denote like elements, and in which:

[0045] FIG. 1 shows a block diagram of an example automated system for grading and sorting of FFBs, in accordance with some embodiments of the present disclosure;

[0046] FIG. 2 shows a detailed block diagram illustrating a grading module of the automated system of FIG. 1, in accordance with some embodiments of the present disclosure;

[0047] FIG. 3 shows another block diagram of an example automated system for grading and sorting of FFBs, in accordance with some embodiments of the present disclosure;

[0048] FIG. 4 shows a flow diagram of an example automated grading and sorting method, in accordance with some embodiments of the present disclosure;

[0049] FIG. 5 shows a flow diagram of another example automated grading and sorting method, in accordance with some embodiments of the present disclosure;

[0050] FIG. 6 shows a flow diagram of yet another example automated grading and sorting method, in accordance with some embodiments of the present disclosure;

[0051] FIG. 7 shows a flow diagram of an example machine learning functionality implemented by the grading module, in accordance with some embodiments of the present disclosure;

[0052] FIG. 8 shows an example automated grading and sorting modules, in accordance with some embodiments of the present disclosure;

[0053] FIG. 9 shows an example unloading and singulation module, in accordance with some embodiments of the present disclosure;

[0054] FIG. 10 shows an example grading, weighing, and sorting module in accordance with some embodiments of the present disclosure;

[0055] FIG. 11 shows a flow diagram of an example grading method implemented by a color imaging module, in accordance with some embodiments of the present disclosure;

[0056] FIG. 12 shows a flow diagram of an example grading method implemented by a thermal imaging module, in accordance with some embodiments of the present disclosure;

[0057] FIG. 13 shows a flow diagram of an example grading method implemented by a hyperspectral imaging module, in accordance with some embodiments of the present disclosure;

[0058] FIG. 14 shows an exemplary temperature plot associated with a training dataset for training the thermal imaging module, in accordance with some embodiments of the present disclosure;

[0059] FIG. 15 shows an exemplary Free Fatty Acid (FFA) plot for training the thermal imaging module, in accordance with some embodiments of the present disclosure;

[0060] FIG. 16 shows an exemplary Deterioration of Bleachability Index (DOB I) plot for training the thermal imaging module, in accordance with some embodiments of the present disclosure; and

[0061] FIG. 17 shows an exemplary Peroxide Value (PV) plot for training the thermal imaging module, in accordance with some embodiments of the present disclosure.

[0062] Like reference numerals refer to like parts throughout the several views of the drawings.Detailed Description

[0063] The following detailed description is merely exemplary in nature and is not intended to limit the described embodiments or the application and uses of the described embodiments. As used herein, the word “exemplary” or “illustrative” means “serving as an example, instance, or illustration ” Any implementation described herein as “exemplary” or “illustrative” is not necessarily to be construed as preferred or advantageous over other implementations. All of the implementations described below are exemplary implementations provided to enable persons skilled in the art to make or use the embodiments of the disclosure and are not intended to limit the scope of the disclosure, which is defined by the claims. Furthermore, there is no intention to be bound by any expressed or implied theory presented in the preceding technical field, background, brief summary or the following detailed description. It is also to be understood that the specific devices and processes illustrated in the attached drawings, and described in the following specification, are simply exemplary embodiments of the inventive concepts defined in the appended claims. Hence, specific dimensions and other physical characteristics relating to the embodiments disclosed herein are not to be considered as limiting, unless the claims expressly state otherwise.

[0064] Unless the context requires otherwise, throughout the specification and claims which follow, the word “comprise” and variations thereof, such as, “comprises” and “comprising” are to be construed in an open, inclusive sense, that is as “including, but not limited to.”

[0065] As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the content clearly dictates otherwise. It should also be noted that the term “or” is generally employed in its broadest sense, that is as meaning “and / or” unless the content clearly dictates otherwise.

[0066] The headings and Abstract of the Disclosure provided herein are for convenience only and do not interpret the scope or meaning of the implementations.

[0067] The present disclosure describes an automated system for grading and sorting Palm Oil Fresh Fruit Bunches (FFB) in palm oil mills. The automated system addresses manual grading inefficiencies by employing image, reflection, and weight analysis. The automated system employs advanced technologies such as hyperspectral and thermal imaging, backed by a machinelearning algorithm. The automated system is configured to grade and / or categorize the FFBs based on the image, reflection, and weight analysis. Further, the automated system is configured to determine an optimal setting required for a sterilization process by employing the machine learning algorithm based on the grade associated with the FFBs, thereby enhancing the efficiency of the sterilization process by providing known properties of FFBs for better machine settings, improving Oil Extraction Rate (OER), and overall mill efficiency.

[0068] Referring initially to FIG. 1, which shows a block diagram of a networked environment 100 for grading and sorting of FFBs, in accordance with some embodiments of the present disclosure. As illustrated in FIG. 1, the networked environment 100 comprises an automated grading and sorting system 102, a user device 104, and a data repository 106. The automated grading and sorting system 102, the user device 104, and the data repository 106 are coupled via a network 108.

[0069] The automated grading and sorting system 102 comprises several key components, such as, but not limited to, an input module 120, a grading module 122, a sorting module 124, and a storage module 126, which work in harmony and enhance the efficiency of the automated grading and sorting system 102. In some embodiments, the automated grading and sorting system 102 may be employed between a Fresh Fruit Bunches (FFBs) reception area of a palm oil mill and a sterilizer station. The FFBs reception area comprises an area where the FFBs arrive at the palm oil mill on vehicles such as, but not limited to, trucks. The sterilizer station comprises an area where the FFBs are subjected to heat to facilitate the extraction of oil from the FFBs.

[0070] The input module 120 comprises an unloading module and a singulation module. The unloading module is configured to receive the FFBs transported by vehicles, for example, trucks, from plantations and prepare the FFBs for the grading process in the grading module 122. In some embodiments, the unloading module is an area where the vehicles carrying the FFBs arrive at the palm oil mill. In order to unload the FFBs, the unloading module comprises one or more components, such as a conveyor system or cranes, to unload the FFBs from the vehicles. In some embodiments, the unloading module comprises automated sensors that may track the arrival of the vehicles, and initiate the unloading process. Further, in some other embodiments, the unloading module may be a ramp platform where the vehicles can be positioned for unloading. The rampplatform is usually inclined to facilitate the automatic movement of the FFBs from the vehicles to the mill without any external device.

[0071] The singulation module is coupled to the unloading module. The singulation module is configured to separate a cluster of FFBs from the unloading module into individual FFBs. Generally, the FFBs are transported in clusters or bunches, especially when unloaded from the vehicles. The singulation module is configured to separate the cluster of FFBs into individual FFBs, thereby ensuring that the individual FFBs are presented to the processing one at a time in the grading module 122. In some embodiments, the singulation module comprises a stopper such that proper spacing is achieved between the individual FFBs when processed to the grading module 122. In some embodiments, the singulation module comprises one or more channels or lanes to separate the cluster of FFBs into the individual FFBs.

[0072] The automated grading and sorting system 102 further comprises the grading module 122 which is a crucial component of the automated grading and sorting system 102. In some embodiments, the grading module 1 2 receives the FFBs from the singulation module of the input module 120 through the conveyor system. The grading module 122, as discussed in detail in relation to FIG. 2, is configured to determine one or more quality indicator characteristics of each FFB when the FFBs pass through the conveyor system. In some embodiments, the grading module 122 is equipped with multiple image sensors to accurately determine the one or more quality indicator characteristics of each FFB and categorize the FFBs based on the determined one or more quality indicator characteristics.

[0073] The automated grading and sorting system 102 further comprises the sorting module 124, which is an integral part of the automated grading and sorting system 102. The sorting module 124 is configured to efficiently handle and organize the FFBs based on the categorization of the FFBs, which is based the one or more quality indicator characteristics. For example, the sorting module 124 may be equipped with one or more robotic arms that are coupled to the grading module 122. In some embodiments, the one or more robotic arms may sort the FFBs based on the categorization. In some other embodiments, the sorting module 124 may be a part of the same conveyor system as the grading module 122. In such embodiments, the one or more robotic arms are configured to re-arrange the FFBs in an appropriate lane of the conveyor system, such that theFFBs with similar quality indicator characteristics are processed in the same lane, thereby sorting, and enabling the storage of the FFBs in the storage module 126 based on the one or more quality indicator characteristics. Further, in some embodiments, in order to maintain the quality of the FFBs during the sorting process, the sorting module 1 4 may implement measures to protect the FFBs from environmental factors. In some embodiments, the measures may comprise sheltering the FFBs from sunlight, rainwater, and other contaminants that could potentially affect the quality of the FFBs.

[0074] The automated grading and sorting system 102 further comprises the storage module 126 that is coupled to the sorting module 124. The storage module 126 serves as a temporary holding for the FFBs. The storage module 126 may comprise a plurality of bins that correspond to the number of groups. The plurality of bins enable the storage of the FFBs according to the group associated with the FFBs. Further, in some embodiments, the storage module 126 is configured to provide controlled environment conditions, which are essential for maintaining the quality of the FFBs. Factors such as temperature, humidity, and exposure to sunlight are carefully regulated in the storage module 126 to prevent deterioration and ensure the freshness of the FFBs during the storage period. Further, storing the FFBs in the plurality of bins based on the group allows for targeted processing in subsequent stages, which ensures that the FFBs with similar characteristics are treated together in the sterilization center.

[0075] Further, the storage module 126 is equipped with one or more sensors that monitor the storage utilization of the storage module 126 and generate an alert, thereby avoiding inefficient utilization of the storage module 126. Furthermore, the storage module 126 is also equipped with one or more components, such as the conveyor system or cranes, that enable the unloading of the FFBs from each bin to a loading system.

[0076] The networked environment 100 further comprises the user device 104 that comprises an integrated digital platform which is a user-friendly platform for one or more users to oversee, control, and interact with the automated grading and sorting system 102. In some embodiments, the integrated digital platform of the user device 104 enables the one or more users associated with the user device 104 to wirelessly connect with the automated grading and sorting system 102. The integrated digital platform may be a software application installed in the user device 104, a webapplication that can be accessed through a web browser associated with the user device 104, or an Original Equipment Manufacturer (OEM) software. The user device 104 may be any electronic device such as a desktop computer, laptop computer, portable or mobile device, cell phone, smartphone, tablet computer, personal digital assistant (PDA), wearable device, or the like.F0077] Moreover, the networked environment 100 comprises a data repository 106 that comprises information related to each batch of the FFBs that has undergone processing in the automated grading and sorting system 102. In some embodiments, the data repository 106 may be integrated within the automated grading and sorting system 102 and the user device 104. In some other embodiments, the data repository 106 may be a standalone repository communicatively coupled with the automated grading and sorting system 102 and the user device 104 over the network 108. In some embodiments, the data repository may be implemented over a cloud (e.g., using cloud computing).

[0078] Entities of the networked environment 100 may be coupled via the network 108, which may be a LAN (local area network), WAN (wide area network), wireless network, point-to-point network, or another configuration. One of the most common types of the network 108 in current use is a TCP / IP (Transfer Control Protocol and Internet Protocol) network for communication between different devices. Other common Internet protocols used for such communication include HTTPS, FTP, AFS, and WAP and using secure communication protocols etc. In some embodiments, the network 108 may be any type of communication network, including one or more of the Internet, local area networks (LAN), wireless networks, switch or hub connections, a telephone network (e.g., a public switched telephone (PSTN) network, a cellular network, or the like), or the like.

[0079] It is to be noted that though FIG. 1 shows a single input module 120, a single grading module 122, a single sorting module 124, and a single storage module 126, a person skilled in the art would appreciate the automated grading and sorting system 102 may comprise a plurality of such modules, which are not shown herein for the sake of brevity.

[0080] FIG. 2 shows a detailed block diagram illustrating the grading module 122 of the automated grading and sorting system 102 of FIG. 1, in accordance with some embodiments of the present disclosure.

[0081] The grading module 122 is a critical component within the automated grading and sorting system 102, responsible for processing the FFBs from the input module 120 to determine the one or more quality indicator characteristics associated with each FFB. In an implementation, the grading module 122 may comprise a processor 202, a memory 204, and one or more modules (interchangeably referred to as modules) 206. The processor 202 may be configured to perform one or more functions to fulfill one or more requirements of the grading module 122. The memory 204 may be communicatively coupled to the processor 202 and may store data (not shown) related to information associated with each FFB which is undergoing analysis within the grading module 122.

[0082] In some implementations, the grading module 122 may comprise the one or more modules 206 for performing various operations in accordance with some embodiments of the present disclosure. In some embodiments, the data may be stored within the memory 204 in the form of various data structures. Additionally, the data may be organized using data models, such as relational or hierarchical data models. Further, the memory 204 also includes other data, which may comprise other temporary data generated by other modules 232 for performing various functions of the grading module 122. In some embodiments, each of the data stored in the memory 204 may be processed by the one or more modules 206 of the grading module 122.

[0083] In some implementations, the modules 206 may be stored as a part of the processor 202. In some other implementations, the modules 206 may be communicatively coupled to the processor 202 to perform one or more functions of the grading module 122. The modules 206 may include, without limiting to, a receiving module 220, a color imaging module 222, a thermal imaging module 224, a spectral imaging module 226, a weighing module 228, a quality assessment module 230, and other modules 232. Further, each of the one or more modules 206 may comprise an Artificial Intelligence (Al) engine that provides intelligent decision-making capabilities for the corresponding module.

[0084] As used herein, the term module refers to an application-specific integrated circuit (ASIC), an electronic circuit, a processor (shared, dedicated, or group) and memory that executes one or more software or firmware programs, a combinational logic circuit, and / or other suitable components that provide the described functionality. Tn an embodiment, the other modules 232 may be used to perform various miscellaneous functionalities of the grading module 122. It would be appreciated that such other modules 232 may be represented as a single module or a combination of different modules.

[0085] In some examples, the processor 202 may comprise at least one controller in communication with at least one non-transitory processor-readable medium. The processor- readable medium may have instructions stored thereon which when executed cause the processors to perform or control performance of the operations as described herein. Furthermore, in some examples, the processor 202 or its functionality may be implemented in other ways, including: via Application Specific Integrated Circuits (ASICs), in standard integrated circuits, as one or more computer programs executed by one or more computers (e.g., as one or more programs running on one or more computer systems), as one or more programs executed by on one or more controllers (e.g., microcontrollers), as one or more programs executed by one or more processors (e.g., microprocessors, central processing units, graphical processing units), as firmware, and the like, or as a combination thereof.

[0086] The receiving module 220 is configured to begin the process implemented by the grading module 122 by receiving the FFBs from the input module 120 through the conveyor system. Further, the receiving module 220 is enabled with one or more components such as rollers, dividers, or guides that enable the processing of the FFBs in a sequential fashion, thereby preventing the FFBs from bunching up or overlapping and facilitating the grading of the FFBs.

[0087] The grading module 122 further comprises the color imaging module 222, which is coupled to the receiving module 220. The color imaging module 222 comprises a color image capturing module and a color image analysis module. The color image capturing module is configured to capture a color image of the FFBs. In some embodiments, the color image of the FFBs is captured when the FFBs move along the conveyor system. In some embodiments, the color image capturing module comprises a plurality of cameras, which in an exampleimplementation, may be suitably positioned, or disposed around the receiving module (e.g., the conveyor system), to capture color images of the FFBs from multiple angles. Further, in some other embodiments, the color imaging module 222 may comprise a mechanism to flip the FFBs once initial color images are captured by the color image capturing module, such that additional color images of the FFBs may be captured by the color image capturing module. By capturing the additional color images, the color image capturing module is configured to provide a complete view of the FFBs. The color image analysis module is then configured to analyze the color characteristics of the FFBs from the captured images (which comprises the initial color images and the additional color images), such as the outer color ranging from yellow to red, to detect and categorize the FFBs into different ripeness levels such as unripe, under-ripe, ripe, over-ripe, and empty. In some embodiment, the color characteristics comprise color range, hue, color variance, appearance, color coverage, and color coverage group of fruitlets on the FFBs. In order to detect and categorize the FFBs from the captured images, an accurate color threshold is required. However, due to variations in cultivars, the color threshold for FFB recognition may differ significantly and a fixed color threshold may not be suitable for all FFB types, which leads to potential misclassifications.

[0088] Accordingly, the color image analysis module is configured to employ a machine learning algorithm, as discussed in detail in relation to FIG. 7, to adaptively determine the color threshold (which is also referred to as a user-defined threshold) for the FFB recognition. Upon determining the color threshold, the color image analysis module is configured to utilize the determined color threshold to detect FFBs from a background in the captured images and to determine at least one value that reflects at least one quality indicator characteristic of the FFB . In order to determine the at least one value that reflects the at least one quality indicator characteristic of the FFB, the color image analysis module is configured to initially remove the background from the captured images which are captured from multiple views, thereby detecting the FFBs in the captured images. Further, the color image analysis module is configured to process the detected FFBs using advanced machine vision algorithms to extract one or more features associated with each FFB. In some embodiments, the one or more features may comprise a length, a width, and a size of the FFB.

[0089] Further, upon extracting the one or more features, the color image analysis module is configured to process the extracted one or more features by employing one or more machine learning algorithms and the color threshold determined by the machine learning algorithm to accurately determine the at least one value that reflects the at least one quality indicator characteristic of each FFBs. In some embodiments, the at least one value comprises a first value that indicates a ripeness of the FFBs. Further, in some embodiments, the first value may indicate the ripeness level of the FFB. In some embodiments, the ripeness level may be indicative of the FFB being unripe, under-ripe, ripe, over-ripe, or empty.

[0090] The grading module 122 further comprises the thermal imaging module 224, which is coupled to the receiving module 220. The thermal imaging module 224 comprises a thermal image capturing module and a thermal image analysis module. The thermal image capturing module is configured to capture thermal images of the FFBs. In some embodiments, the thermal image capturing module comprises thermal cameras or sensors that are equipped with detectors, which in an example implementation, may be suitably positioned, or disposed around the receiving module (e.g., the conveyor system), to sense and measure infrared radiation emitted by the FFBs. The amount of radiation emitted by the FFBs is directly related to the temperature of the FFBs. Further, the thermal image capturing module is configured to translate the detected infrared radiation into a thermal image, which represents the temperature variations across the FFBs.

[0091] Upon generating the thermal image, the thermal image analysis module is configured to determine at least one value that reflects at least one quality indicator characteristic of each FFB by analyzing the thermal image. In one embodiment, the at least one value comprises one or more of: a first value that indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB. In some embodiments, the thermal image analysis module is configured to determine at least one value that reflects at least one quality indicator characteristic of each FFB based on a temperature difference (AT) between a temperature of each FFB and the atmospheric surface. Further, in some embodiments, the second value, which indicates the oil quality, is determined by measuring the FFA, the DOBI, and the PV associated with each FFB based on the temperature difference (AT).

[0092] In some embodiments, the thermal image analysis module is configured to employ one or more machine learning algorithms that determine the first value and the second value associated with each FFB based on the temperature difference (AT). In some embodiments, the one or more machine learning algorithms may be initially trained on a significant amount of labeled data. The labeled data comprises images of the FFBs along with corresponding information such as ripeness levels, oil content, and other relevant parameters. Generally, human experts may qualitatively determine the ripeness level of FFBs in the training dataset, and the FFBs may undergo laboratory chemical analysis for oil content such as the FFA, the PV, the DOBI, and carotene. In some embodiments, the thermal imaging module 224 may employ a machine learning algorithm similar to that used in the color imaging module 222 in order to determine the first value and the second value.

[0093] In some example embodiments, the thermal image used by the thermal imaging module 224 may be used to predict the oil quality in the FFBs, whereas the color image used by the color imaging module 222 may be used to assess the ripeness level of the FFBs.

[0094] The grading module 122 further comprises the spectral imaging module 226, which is coupled to the receiving module 220. In some embodiments, the spectral imaging module 226 comprises a spectral image capturing module and a reflection analysis module. The spectral image capturing module may be a hyperspectral camera and a Near Infra-Red (NIR) camera. The hyperspectral camera and the NIR camera are configured to capture a reflection of the FFBs for an electromagnetic spectrum ranging from 400-1700nm. Similar to color image cameras, and thermal image cameras, or sensors, the hyperspectral camera and the Near Infra-Red (NIR) camera may also be suitably positioned, or disposed around the receiving module (e.g., the conveyor system). Upon capturing the reflection response, the reflection analysis module is configured to generate a reflectance and a plot between the reflectance and wavelength of the electromagnetic spectrum. The plot may used to determine one or more quality indicator characteristics of the FFBs such as, but not limited to, ripeness level value and oil quality value.

[0095] Upon generating the plot, the spectral imaging module 226 is configured to determine at least one value that reflects at least one quality indicator characteristic of each FFB by analyzing the generated plot. In one embodiment, the at least one value comprises one or more of a first valuethat indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB. In some embodiments, the spectral imaging module 226 is configured to employ one or more machine learning algorithms that determines the first value and the second value using the generated plot. In some embodiments, the one or more machine learning algorithms may be initially trained on a significant amount of labeled data. The labeled data comprises spectral response images of the FFBs along with corresponding information such as ripeness levels, oil content, and other relevant parameters. Generally, human experts may qualitatively determine the ripeness level of FFBs in the training dataset, and the FFBs may undergo laboratory chemical analysis for oil content such as the FFA, the PV, the DOBI, and carotene. In some embodiments, the spectral imaging module 226 may employ a machine learning algorithm similar to that used in the color imaging module 222.F0096] Further, in some embodiments, the color imaging module 222, the thermal imaging module 224, and the spectral imaging module 226 may be employed as a single module, thereby capturing the color image, the thermal image, and the spectral response image simultaneously. Further, multiple such single modules may be employed by the grading module 122 to capture multiple views of the FFBs. In some embodiments, the grading module 122 may employ three sets of the single modules.

[0097] The grading module 122 further comprises the weighing module 228, which is coupled to the receiving module 220. In some embodiments, the weighing module 228 is integrated into the conveyor system such that the weighing module 228 is configured to measure and analyze the weight of individual FFBs. In some embodiments, the weighing module 228 may utilize load cells or other weight-sensing mechanisms to accurately measure a weight of each FFB. In some embodiments, the weight of the FFBs measured by the weighing module 228 is used by other components, such as the quality assessment module 230, as a value that reflects at least one quality indicator characteristic of the corresponding FFB.

[0098] The grading module 122 further comprises the quality assessment module 230, which is coupled to the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228. In some embodiments, the quality assessment module 230 is configured to receive output from each of the color imaging module 222, the thermalimaging module 224, the spectral imaging module 226, and the weighing module 228. In some embodiments, the output from each of the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228 comprises the at least one value that reflects the at least one quality indicator characteristic of each FFB based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis. In one embodiment, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB, a second value that indicates an oil quality of the FFB, and a third value that indicates a weight of the FFB.

[0099] In some other embodiments, the quality assessment module 230 is configured to determine at least one value that reflects at least one quality indicator characteristic of each of the plurality of FFBs based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis from the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228.

[0100] Further, in some embodiments, the quality assessment module 230 is configured to determine the one or more quality indicator characteristics based on the determined at least one value. In some embodiments, the one or more quality indicator characteristics comprise at least one of: the ripeness level, the aesthetic condition, the one or more diseases associated with the FFB, the oil quality of the FFB, the brix percentage, the species, and the firmness of the FFB.

[0101] In some embodiments, the quality assessment module 230 is configured to determine the one or more quality indicator characteristics by employing one or more advanced machine learning algorithms. The one or more quality indicator characteristics may be used to categorize the FFBs into the number of groups.

[0102] In some embodiments, in order to categorize the FFBs into the number of groups, the grading module 122 is configured to obtain user-defined profile data associated with categorization of the FFBs and compare the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data. The grading module 122 is then configured to categorize the plurality of FFBs into the number of groups based on the comparison.

[0103] In some embodiments, in order to obtain user-defined profile data, as described previously with respect to functionality of the color imaging module 222, thermal imaging module 224, and spectral imaging module 226, the grading module 122 is configured to obtain a training set, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs and obtain one or more quality indicator characteristics of the test group of FFBs. The one or more quality indicator characteristics are obtained from a user based on manual inspection and lab reports for the test group of FFBs. The grading module 122 is then configured to determine the user-defined profile data based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results.

[0104] In some embodiments, in order to determine the user-defined profile data, the grading module 122 is configured to apply the obtained training set and the obtained one or more quality indicator characteristics to a machine learning algorithm.

[0105] Further, the grading module 122 is also configured to determine a respective sterilization recipe for each of the number of groups using one or more advanced machine learning algorithms. In some embodiments, the sterilization recipe comprises at least one of a time value, a pressure value, and a temperature value, which may be employed as process parameters in a sterilization center, to maximize Oil Extraction Rate (OER) from the plurality of FFBs.

[0106] Referring now to FIG. 3, which shows another block diagram of an example implementation of an automated system 300 for grading and sorting of the FFBs, in accordance with some embodiments of the present disclosure.

[0107] As illustrated in FIG.3, the automated system 300 comprises a color imaging device 302, a thermal imaging device 304, a spectral imaging device 306, a weighing device 308, a computing device 310, a server 312, and a data repository 314. The color imaging device 302, the thermal imaging device 304, the spectral imaging device 306, the weighing device 308, the computing device 310, the server 312, and the data repository 314 are coupled via a network 316. In some embodiments, each of the color imaging device 302, the thermal imaging device 304, the spectralimaging device 306, the weighing device 308 may be mounted around the conveyor system through which FFBs pass to the sterilization center.

[0108] The color imaging device 302 is configured to capture a color image of the FFBs. In some embodiments, the color image of the FFBs is captured when the FFBs move along the conveyor system. In some embodiments, the color imaging device 302 comprises a plurality of cameras to capture color images of the FFBs from multiple angles. Further, in some other embodiments, the color imaging device 302 is also configured to capture additional images when the FFBs are flipped and when initial color images are captured. By capturing the additional color images, the color imaging device 302 is configured to provide a complete view of the FFBs. The color imaging device 302 is then configured to analyze the color characteristics of the FFBs from the captured images (which comprises the initial color images and the additional color images), such as the outer color ranging from yellow to red, to categorize the FFBs into different ripeness levels such as unripe, under-ripe, ripe, over-ripe, rotten, and empty. In some embodiment, the color characteristics comprise color range, hue, color variance, appearance, color coverage, and color coverage group of fruitlets on the FFBs.

[0109] In some embodiments, the color imaging device 302 is configured to employ a machine learning algorithm, as discussed in detail in relation to FIG. 7, to adaptively determine the color threshold (which is also referred to as user-defined threshold) for the FFB recognition from the captured images. Upon determining the color threshold, the color imaging device 302 is configured to utilize the determined color threshold to detect FFBs from a background in the captured images and to determine the at least one value that reflects the at least one quality indicator characteristic of the FFB. In order to determine the at least one value that reflects the at least one quality indicator characteristic of the FFB, the color imaging device 302 is configured to initially remove the background from the captured images which are captured from multiple views, thereby detecting the FFBs in the captured images. Further, the color imaging device 302 is configured to process the detected FFBs using advanced machine vision algorithms to extract one or more features associated with each FFB. In some embodiments, the one or more features may comprise a length, a width, and a size of the FFB. Further, upon extracting the one or more features, the color imaging device 302 is configured to process the extracted one or more features by employing one or moremachine learning algorithms and the color threshold determined by the machine learning algorithm to accurately determine the at least one value that reflects the at least one quality indicator characteristic of each FFBs. In some embodiments, the at least one value comprises a first value that indicates a ripeness of the FFBs. ronoi The automated system 300 further comprises the thermal imaging device 304, which is configured to capture thermal images of the FFBs. In some embodiments, the thermal imaging device 304 comprises thermal cameras or sensors that are equipped with detectors that may sense and measure an infrared radiation emitted by the FFBs. The amount of radiation emitted by the FFBs is directly related to the temperature of the FFBs. Further, the thermal imaging device 304 is configured to translate the detected infrared radiation into a thermal image, which represents the temperature variations across the FFBs.

[0111] Upon generating the thermal image, the thermal imaging device 304 is configured to determine the at least one value that reflects the at least one quality indicator characteristic of each FFB by analyzing the thermal image. In one embodiment, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB. In some embodiments, the thermal imaging device 304 is configured to determine at least one value that reflects at least one quality indicator characteristic of each FFB based on a temperature difference (AT) between a temperature of each FFB and the atmospheric surface. In some embodiments, the thermal imaging device 304 is configured to employ one or more machine learning algorithms that determines the first value and the second value associated with each FFB based on the temperature difference (AT). In some embodiments, the one or more machine learning algorithms may be initially trained on a significant amount of labeled data. The labeled data comprises images of the FFBs along with corresponding information such as ripeness levels, oil content, and other relevant parameters. Generally, human experts may qualitatively determine the ripeness level of FFBs in the training dataset, and the FFBs may undergo laboratory chemical analysis for oil content such as the FFA, the PV, the DOBI, and carotene.

[0112] The automated system 300 further comprises the spectral imaging device 306, which comprises at least one a hyperspectral camera and a Near Infra-Red (NIR) camera. The hyperspectral camera and the NIR camera are configured to capture a reflection of the FFBs foran electromagnetic spectrum ranging from 400-1700nm. Upon capturing the reflection response, the spectral imaging device 306 is configured to generate a reflectance and a plot between the reflectance and wavelength of the electromagnetic spectrum. The plot may used to determine the one or more quality indicator characteristics of the FFBs such as ripeness level value and oil content value.

[0113] Upon generating the plot, the spectral imaging device 306 is configured to determine the at least one value that reflects the at least one quality indicator characteristic of each FFB by analyzing the generated plot. In one embodiment, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB. In some embodiments, the spectral imaging device 306 is configured to employ one or more machine learning algorithms that determines the first value and the second value using the generated plot. In some embodiments, the one or more machine learning algorithms may be initially trained on a significant amount of labeled data. The labeled data comprises spectral response images of the FFBs along with corresponding information such as ripeness levels, oil content, and other relevant parameters. Generally, human experts may qualitatively determine the ripeness level of FFBs in the training dataset, and the FFBs may undergo laboratory chemical analysis for oil content such as the FFA, the PV, the DOB1, and carotene.

[0114] The automated system 300 further comprises the weighing device 308, which is integrated into the conveyor system such that the weighing device 308 is configured to measure and analyze the weight of individual FFBs. In some embodiments, the weighing device 308 may utilize load cells or other weight-sensing mechanisms to accurately measure a weight of each FFB. In some embodiments, the weight of the FFBs measured by the weighing device 308 is used by other components, such as the computing device 310 or the server 312, to determine characteristics of the FFBs such as the quality indicator characteristics.

[0115] The automated system 300 further comprises the computing device 310, which is coupled to the color imaging device 302, the thermal imaging device 304, the spectral imaging device 306, and the weighing device 308. In some embodiments, the computing device 310 is configured to receive output from each of the color imaging device 302, the thermal imaging device 304, the spectral imaging device 306, and the weighing device 308. In some embodiments,the output from each of the color imaging device 302, the thermal imaging device 304, the spectral imaging device 306, and the weighing device 308 comprises the at least one value that reflects the at least one quality indicator characteristic of each of the plurality of FFBs based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis. In one embodiment, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB, a second value that indicates an oil quality of the FFB, and a third value that indicates a weight of the FFB. In some other embodiments, the output from each of the color imaging device 302, the thermal imaging device 304, the spectral imaging device 306, and the weighing device 308 comprises color image analysis results, thermal image analysis results, spectral response analysis results, and weighing analysis results.

[0116] The computing device 310 is further configured to determine the one or more quality indicator characteristics based on the determined at least one value. In some embodiments, the one or more quality indicator characteristics comprise at least one of: the ripeness level, the aesthetic condition, the one or more diseases associated with the FFB, the oil quality of the FFB, the brix percentage, the species, and the firmness of the plurality of FFB.

[0117] Further, the computing device 310 is configured to categorize the FFBs into a number of groups based on grading results, which are based the one or more quality indicator characteristics. In some embodiments, in order to categorize the FFBs into the number of groups, the computing device 310 is configured to obtain user-defined profile data associated with the categorization of the FFBs from the data repository 314 and compare the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data. The computing device 310 is then configured to categorize the plurality of FFBs into the number of groups based on the comparison.

[0118] In some embodiments, in order to obtain user-defined profile data, the computing device 310 is configured to obtain a training set, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs and obtain one or more quality indicator characteristics of the test group of FFBs, wherein the one or more quality indicator characteristics are obtained from a user based on manual inspection and lab reports for the test group of FFBs. The computing device 310 is then configured to determine the user-defined profiledata based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results. In some embodiments, in order to determine the user-defined profile data, the computing device 310 is configured to apply the obtained training set and the obtained one or more quality indicator characteristics to a machine learning algorithm.

[0119] The automated system 300 further comprises the server 312, which is configured to host an integrated platform through which a remote user may access the information related to each FFB using the user device 104 that is provided with the integrated platform.

[0120] The automated system 300 further comprises the data repository 314 which comprises information related to each batch of FFB s that has undergone processing in the automated system 300. The information includes color image analysis results, thermal image analysis results, spectral response analysis results, and characteristics of FFBs. Further, the data repository 314 also comprises user defined profiles and an acceptable value for each user defined profile.

[0121] The network 316 may be a LAN (local area network), WAN (wide area network), wireless network, point-to-point network, or another configuration. One of the most common types of the network 316 in current use is a TCP / IP (Transfer Control Protocol and Internet Protocol) network for communication between different devices. Other common Internet protocols used for such communication include HTTPS, FTP, AFS, and WAP and using secure communication protocols etc. In some embodiments, the network 108 may be any type of communication network, including one or more of the Internet, local area networks (LAN), wireless networks, switch or hub connections, a telephone network (e.g., a public switched telephone (PSTN) network, a cellular network, or the like), or the like.

[0122] FIG. 4 shows a flow diagram of an example automated grading and sorting method in accordance with some embodiments of the present disclosure. As illustrated in FIG. 4, the method 400 comprises one or more blocks implemented by the automated grading and sorting system 102 or by the automated system 300. The method 400 may be described in the general context of computer executable instructions. Generally, computer executable instructions can includeroutines, programs, objects, components, data structures, procedures, modules, and functions, which perform particular functions or implement particular abstract data types.

[0123] The order in which the method 400 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method 400. Additionally, individual blocks may be deleted from the method 400 without departing from the spirit and scope of the subject matter described herein. Furthermore, the method 400 can be implemented in any suitable hardware, software, firmware, or combination thereof.

[0124] At block 410, the Fresh Fruit Bunches (FFBs) are received from one or more sources. In some embodiments, the FFBs are received from one or more sources by the input module 120, which comprises the unloading module and the singulation module. In some embodiments, the FFBs transported by the one or more sources, such as trucks, from plantations are received by the unloading module. In some embodiments, the unloading module is an area where the trucks carrying the FFBs arrive at the palm oil mill. In order to unload the FFBs, the unloading module comprises one or more components, such as a conveyor system or cranes, to remove the FFBs from the trucks. In some embodiments, the unloading module comprises automated sensors that may track the arrival of the trucks and initiate the unloading process. Further, in some other embodiments, the unloading module may be a ramp platform where trucks can be positioned for unloading. The ramp platform is usually inclined to facilitate the automatic movement of the FFBs from the trucks to the mill without any external device. Upon receiving the FFBs from the one or more sources, a cluster of FFBs from the unloading module is separated into individual FFBs by the singulation module. Generally, the FFBs are transported in clusters or bunches, especially when unloaded from the trucks. In some embodiments, the singulation module comprises a stopper such that proper spacing is achieved between the individual FFBs when processed to the grading module 122. In some embodiments, the singulation module comprises one or more channels or lanes to separate the cluster of FFBs into the individual FFBs.

[0125] At block 420, one or more quality indicator characteristics of each FFB is determined. In some embodiments, the quality indicator characteristics of each FFB is determined by the grading module 122. In some embodiments, the FFBs from the singulation module are receivedby the receiving module 220 of the grading module 122. In some embodiments, the receiving module 220 is enabled with one or more components such as rollers, dividers, or guides that enable the processing of the FFBs in order to determine the one or more quality indicator characteristics.

[0126] Further, in some embodiments, the received FFBs are processed using the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228. In some embodiments, the color image of the FFBs is captured by the color image capturing module of the color imaging module 222. In some embodiments, the color image of the FFBs is captured when the FFBs move along the conveyor system. In some embodiments, the color image capturing module comprises a plurality of cameras to capture color images of the FFBs from multiple angles. Further, in some other embodiments, the color imaging module 222 may comprise a mechanism to flip the FFBs once initial color images are captured by the color image capturing module, such that additional color images of the FFBs may be captured by the color image capturing module. By capturing the additional color images, a complete view of the FFBs is provided by the color image capturing module. The captured image is then analyzed by the color image analyzing module to determine the color characteristics of the FFBs from the captured images (which comprises the initial color images and the additional color images), such as the outer color ranging from yellow to red, to categorize the FFBs into different ripeness levels such as unripe, under-ripe, ripe, over-ripe, and empty. In order to detect or categorize the FFBs, an accurate color threshold is required. However, due to variations in cultivars, the color threshold for FFB recognition may differ significantly and a fixed color threshold may not be suitable for all FFB types, which leads to potential misclassifications.

[0127] Accordingly, the machine learning algorithm, as discussed in detail in relation to FIG. 7, is employed by the color imaging module 222 to adaptively determine the color threshold (which is also referred to as user-defined threshold) for the FFB recognition. Upon determining the color threshold, the determined color threshold is utilized by the color image analysis module to detect FFBs from a background in the captured images and to determine at least one value that reflects at least one quality indicator characteristic of the FFB. In order to determine the at least one value that reflects the at least one quality indicator characteristic of the FFB, the background is removed from the captured images which are captured from multiple views, thereby detecting the FFBs inthe captured images. Further, the detected FFBs are processed by the color image analysis module using advanced machine vision algorithms to extract one or more features associated with each FFB.

[0128] Further, upon extracting the one or more features, the extracted one or more features are processed, by the color image analysis module, by employing one or more machine learning algorithms and the color threshold determined by the machine learning algorithm to accurately determine at least one value that reflects at least one quality indicator characteristic of each FFBs. In some embodiments, the at least one value comprises a first value that indicates a ripeness of the FFBs. In some embodiments, the one or more machine learning algorithms may be classification algorithms. In some embodiments, the one or more machine learning algorithms may be a trained machine learning algorithm which is employed by the machine learning algorithm.

[0129] Further, in some embodiments, thermal images of the FFBs are captured by a thermal image capturing module of the thermal imaging module 224. Upon capturing the thermal image, at least one value that reflects at least one quality indicator characteristic of each FFB is determined by the thermal image analysis module of the thermal imaging module 224 by analyzing the thermal image. In some embodiments, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB. In some embodiments, the at least one value that reflects at least one quality indicator characteristic of each FFB is determined by the thermal image analysis module based on a temperature difference (AT) between a temperature of each FFB and the atmospheric surface. In some embodiments, one or more machine learning algorithms are employed by the thermal image analysis module to determine the at least one value. In some embodiments, the at least one value comprises a first value that indicates a ripeness of the FFBs and a second value that indicates an oil content of the FFBs.

[0130] Further, in some embodiments, a reflection of the FFBs is captured by the spectral imaging module 226 for an electromagnetic spectrum ranging from 400-1700nm. Upon capturing the reflection response, the reflectance is generated and the plot between the reflectance and wavelength of the electromagnetic spectrum is also generated by the spectral imaging module 226. The plot may used to determine characteristics of the FFBs such as ripeness level and oil content.

[0131] Upon generating the plot, at least one value that reflects at least one quality indicator characteristic of each FFB is determined by the spectral imaging module 226, by analyzing the generated plot. In one embodiment, the at least one value comprises one or more of a first value that indicates a ripeness of the FFB and a second value that indicates an oil quality of the FFB.

[0132] . In order to determine the first value and the second value, one or more machine learning algorithms are employed by the spectral imaging module 226.

[0133] Further, in some embodiments, the weight of individual FFBs is measured by the weighing module 228. In some embodiments, the weighing module 228 may utilize load cells or other weight-sensing mechanisms to accurately measure the weight of each FFB. In some embodiments, the weight of the FFBs measured by the weighing module 228 is used by other components, such as the quality assessment module 230, to as a value that reflects at least one quality indicator characteristic of the corresponding FFB.

[0134] Further, in some embodiments, output from each of the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228 is received by the quality assessment module 230. In some embodiments, the output from each of the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228 comprise at least one value that reflects at least one quality indicator characteristic of each of the plurality of FFBs based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis. In one embodiment, the at least one value comprises one or more of the first value that indicates a ripeness of the FFB, the second value that indicates an oil quality of the FFB, and the third value that indicates the weight of the FFB. In some other embodiments, the at least one value that reflects at least one quality indicator characteristic of each of the plurality of FFBs is determined by the quality assessment module 230 based on one or more of the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis from the color imaging module 222, the thermal imaging module 224, the spectral imaging module 226, and the weighing module 228.

[0135] Further, in some embodiments, the one or more quality indicator characteristics is determined by the quality assessment module 230 based on the calculated at least one value. Insome embodiments, the one or more quality indicator characteristics comprise at least one of: the ripeness level, the aesthetic condition, the one or more diseases associated with the FFB, the oil quality of the FFB, the brix percentage, the species, and the firmness of the FFB.

[0136] In some embodiments, the oil quality is indicated by at least one of the FFA, the DOBI, and the PV.

[0137] At block 430, the FFBs are categorized into a number of groups. In some embodiments, the FFBs are categorized into the number of groups by the grading module 122 based on the determined quality indicator characteristics. Further, in some embodiments, one or more robotic arms are equipped by the sorting module 124 that may sort the FFBs into the number of groups based on the categorization. Further, in some embodiments, in order to maintain the quality of FFBs during the sorting process, one or more measures are implemented by the sorting module 124 to protect the FFBs from environmental factors. In some embodiments, the one or more measures may comprise sheltering the FFBs from sunlight, rainwater, and other contaminants that could potentially affect the quality of the FFBs.

[0138] At block 440, the FFBs are stored in the plurality of bins. In some embodiments, the FFBs are stored in the one or more bins of the storage module 126 according to the category or group associated with the FFBs. Further, in some embodiments, controlled environment conditions are provided by the storage module 126, which are essential for maintaining the quality of FFBs. Factors such as temperature, humidity, and exposure to sunlight are carefully regulated in the storage module 126 to prevent deterioration and ensure the freshness of the FFBs during the storage period. In some embodiments, one or more components, such as a conveyor system or cranes, are equipped in the storage module 126 that enable the unloading of the FFBs from each bin to a loading system.

[0139] FIG. 5 is a flow diagram of another example automated grading and sorting method in accordance with an embodiment of the present disclosure. As illustrated in FIG. 5, the method 500 comprises one or more blocks implemented by the automated grading and sorting system 102. The method 500 may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components,data structures, procedures, modules, and functions, which perform particular functions or implement particular abstract data types.

[0140] The order in which the method 500 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method 500. Additionally, individual blocks may be deleted from the method 500 without departing from the spirit and scope of the subject matter described herein. Furthermore, the method 500 can be implemented in any suitable hardware, software, firmware, or combination thereof.

[0141] At block 510, one or more quality indicator characteristics of a plurality of FFBs are determined. The description of block 510 is analogous to block 420, which is not repeated herein for the sake of brevity.

[0142] At block 520, user-defined profile data associated with the categorization of the FFBs is received. In some embodiments, the user-defined profile data associated with the categorization of the FFBs is received by the grading module 122. Tn some embodiments, in order to receive user- defined profile data, initially a training set is obtained by the grading module 122, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs. Further, one or more quality indicator characteristics of the test group of FFBs is also obtained by the grading module 122, wherein the one or more quality indicator characteristics are obtained from a user based on manual inspection and lab reports for the test group of FFBs. Upon obtaining the training set and one or more quality indicator characteristics, the user-defined profile data is determined by the grading module 122 based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results. In some embodiments, in order to determine the user-defined profile data, a machine learning algorithm may be employed by the grading module 122.

[0143] At block 530, the determined one or more characteristics of the FFBs are compared with the user-defined profile data. In some embodiments, the determined one or more characteristics of the plurality of FFBs are compared with the user-defined profile data by the grading module 122.

[0144] At block 540, the plurality of FFBs is categorized into the number of groups based on the comparison. In some embodiments, the plurality of FFBs is categorized by the grading module 122 into the number of groups based on the comparison.

[0145] FIG. 6 is a flow diagram of yet another example automated grading and sorting method in accordance with an embodiment of the present disclosure. As illustrated in FIG. 6, the method 600 comprises one or more blocks implemented by the automated grading and sorting system 102. The method 600 may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform particular functions or implement particular abstract data types.

[0146] The order in which the method 600 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method 600. Additionally, individual blocks may be deleted from the method 600 without departing from the spirit and scope of the subject matter described herein. Furthermore, the method 600 can be implemented in any suitable hardware, software, firmware, or combination thereof.

[0147] At block 610, one or more analysis results are received. In some embodiments, the one or more analysis results comprise an image analysis result, a reflection analysis result, and a weight analysis result which are obtained from various modules of the grading module 122.

[0148] At block 620, at least one value that reflects at least one characteristic of each FFB is determined. In some embodiments, the at least one value that reflects the at least one characteristic of each FFB is determined by the grading module 122 based on the received image analysis results, the received reflection analysis results, and the received weight analysis results.

[0149] At block 630, one or more quality indicator characteristics are determined. In some embodiments, the one or more quality indicator characteristics are determined by the grading module 122 based on the at least one value that reflects the at least one characteristic.

[0150] At block 640, training dataset, human inspection results, and lab results are received. In some embodiments, the training dataset comprises at least one of color images, thermal images, and spectral images for a test group of FFBs. In some embodiments, the human inspection results and the lab results comprise one or more quality indicator characteristics of the test group of FFBs. In some embodiments, the human inspection results comprise ripeness level of the test group of FFBs, and the lab results comprises oil quality of the test group of FFBs.

[0151] At block 650, one or more user defined profiles are obtained. In some embodiments, the one or more user defined profiles are obtained by the grading module 122 based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profiles comprise an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results. Further, the obtained user-defined profiles are stored in the data repository 106 for future reference.

[0152] At block 660, the FFBs are categorized into a number of groups based on comparison of one or more quality indicator characteristics with the user defined profiles. In some embodiments, the FFBs are categorized into the number of groups by the grading module 122 based on comparison of one or more quality indicator characteristics with the user defined profiles.

[0153] FIG. 7 shows a flow diagram of an example machine learning functionality implemented by the grading module 122, in accordance with some embodiments of the present disclosure.

[0154] At step 710, a labeled data set (X0, Y0) is received, which is used as training data to train a machine learning algorithm. In general, X0 represents the set of labeled input data and Y0 represents the set of corresponding labels or categories for the input data X0. In the present disclosure, X0 may comprise images of FFBs. Further, Y0 may comprise labels indicating whether each image in the input data X0 contains an FFB and / or belongs to a specific class (ripe, unripe, overripe, etc.).

[0155] At block 720, the machine learning algorithm, such as a supervised learning algorithm, is trained on the labeled dataset (X0, Y0) to learn patterns and characteristics associated with the FFBs. By training the machine learning algorithm, the machine learning algorithm may predict and classify images as the FFB or background for any new incoming dataset.

[0156] At block 730, the machine learning algorithm trained on the dataset (X0, Y0) predicts an output Y 1 for an input dataset XI which is unlabeled data set.

[0157] At block 740, one or more features associated with the input dataset XI are extracted. In some embodiments, in order to extract the one or more features, initially a background is removed from the images included in the input dataset XI. Upon removing the background, one or more scaly spikes and stalk of the FFBs are detected and removed using a CNN algorithm. Further, the one or more features are extracted upon removing the background, scaly spikes, and stalk. In some embodiments, the one or more features comprises at least one of length, width, and defect of each FFB.

[0158] At block 750, the one or more images in the input data set (XI) and corresponding extracted one or more features are processed to determine an extraction error. In some embodiments, the extraction error may refer to an error in the extracted features and expected features for the input data set (XI).

[0159] At block 760, dataset in the input data set (XI ) with the extraction error greater than a threshold is discarded.

[0160] At block 770, dataset in the unlabelled data set XI with the extraction error less than the threshold are processed as training data to re-train the machine learning model, thereby iteratively refining the machine learning algorithm to improve the accuracy and adapt to variations in FFB characteristics. The refined machine learning algorithm is used as a processing algorithm or as the machine learning algorithm by at least one of the color imaging module 222, the thermal imaging module 224, and the spectral imaging module 226 which incorporates an accurate threshold for categorizing the FFBs.

[0161] FIG. 8 shows an example automated grading and sorting module, in accordance with some embodiments of the present disclosure. As shown in Fig. 8, the automated grading and sorting system 102 comprises the input module 120, the grading module 122, the sorting module 124, and the storage module 126. The input module 120, as shown in Fig. 8, comprises the unloading module 802 and the singulation module 804. The input module 120 is configured to receive FFBs from the truck 806 and prepare the FFBs for the subsequent grading and sortingprocesses. The FFBs are unloaded from the truck 806 at the unloading module 802. In some embodiments, the unloading module 802 and the singulation module 804 are slanted, allowing the FFBs to roll downwards naturally. The singulation module 804, as shown in the Fig. 8, is a gravity- driven singulation module which is used to separate cluster of FFBs into individual FFBs. Further, the singulation module 804 is configured to ensure a proper spacing between the FFBs when the FFBs are directed toward the grading process.

[0162] The grading module 122 comprises a grading and weighting module 808. The grading and weighting module 808 is responsible for evaluating the quality of individual FFBs by employing various sensing and analysis techniques such as color imaging, thermal imaging, and spectral response imaging. As the FFBs move through a conveyor system 810 with one or multiple lanes, a first set of multiple sensors 812 equipped within the grading and weighting module 808, such as visible color cameras, thermal cameras, hyperspectral devices, and NIR devices capture data from different angles. Further, a second set of multiple sensors 814 captures data from other angles as the FFBs are flipped in the conveyor system as they move. The captured data is then processed, as explained in FIG. 2, to determine one or more quality indicator characteristics such as the ripeness level, the oil quality, and the weight. Using the determined quality indicator characteristics, the FFBs may be categorized by the grading module 122.

[0163] The sorting module 124 is configured to sort the FFBs based on the quality indicator characteristics and the categorization of the FFBs. The storage module 126 comprises one or more bins 816 which are used to store the FFBs based on the categorization. The storage module 126, as shown in FIG. 8, interfaces seamlessly with the loading ramp station 818 without requiring modifications.

[0164] FIG. 9 shows an example unloading and singulation module, in accordance with some embodiments of the present disclosure.

[0165] As shown in Fig. 6, the FFBs are initially unloaded from the truck 806 and directed to the unloading module 802, which is typically a slanted area. The slanted area of the unloading module 802 allows the FFBs to move downward due to gravity. After descending through the unloading module 802, the FFBs enter the singulation module 804. In the singulation module 804,the movement of the FFBs is carefully controlled to ensure that they are evenly spaced and do not cluster together. The singulation process implemented by the singulation module 804 separates the cluster of FFBs into individual FFBs, thereby preventing the FFBs from bunching up or overlapping. The singulation process helps to avoid jams or blockages in the automated grading and sorting system 102 by preventing the FFBs from colliding with each other using an actuated stopper 905. Once the FFBs are singulated, the FFBs are now organized for the subsequent grading and weighing processes in a controlled and systematic manner using an interface 910 to the grading and weighing module 808.

[0166] FIG. 10 shows an example grading and weighing module, in accordance with some embodiments of the present disclosure.

[0167] As shown in Fig. 10, the FFBs 1005 are moved through the grading and weighing module 808 using the conveyor system 810. The conveyor system 810 may be one or multiple lanes and is powered to transport the FFBs efficiently. The grading and weighing module 808 comprise a plurality of sensor sets. Each sensor set (SI, S2, S3) comprises a combination of a camera, hyperspectral device, and near-infrared (NIR) sensors. These sensors are employed to capture various types of information from the FFBs, including visual images, thermal images, hyperspectral data, and NIR reflectance. A weight load cell (Wl) is mounted on top of the conveyor system 1010. The weight load cell is used to measure the weight of the FFBs as they pass through the conveyor system 1010. To enhance accuracy, the FFBs are inspected from six different angles. This involves capturing three views using the first sensor sets (SI, S2, S3) 812 and then flipping to expose the other sides of the FFBs, to capture images, reflectance, and weight results from the opposite sides using the second sensor sets (SI, S2, S3) 814. This multi-view approach is aimed at obtaining a comprehensive understanding of the FFB characteristics as all six sides of the FFB are thoroughly inspected.

[0168] FIG. 11 shows a flow diagram of an example grading method implemented by the color imaging module, in accordance with some embodiments of the present disclosure.

[0169] At block 1110, a color image of the FFBs is captured by the color image capturing module.

[0170] At block 1120, a background is removed from the captured color image of the FFB. In some embodiments, the background is removed from the captured color image using a color threshold which is dynamically determined using the machine learning algorithm. The background is removed from the captured images by the color imaging module 222, thereby detecting the FFBs in the captured images.

[0171] At block 1130, the detected FFBs are processed using advanced machine vision algorithms to extract one or more features associated with each FFBs.

[0172] At block 1140, upon extracting the one or more features, the extracted one or more features are processed in a processing center which employs one or more machine learning algorithms and the color threshold which are determined by the machine learning algorithm to accurately categorize the FFB into the number of profile groups. In some embodiments, the one or more machine learning algorithms may be classification algorithms.

[0173] At block 1 150, the FFBs are categorized into different profile groups based on an output of the processing center. In an example embodiment, the profile group 1 corresponds to overripe, the profile group 2 corresponds to ripe, and the profile group 3 corresponds to underripe.

[0174] FIG. 12 shows a flow diagram of an example grading method implemented by the thermal imaging module, in accordance with some embodiments of the present disclosure.

[0175] At block 1210, a thermal image of the FFBs is captured by the thermal image capturing module.

[0176] At block 1220, a background is removed from the captured thermal image of the FFB. In one embodiment, the background is removed from the captured thermal image using a thermal threshold which is dynamically determined using the machine learning algorithm. The background is removed from the captured images by the thermal imaging module, thereby detecting the FFBs in the captured images.

[0177] At block 1230, the detected FFBs are processed using advanced machine vision algorithms to extract the temperature of the FFBs and the temperature of a surface. Further, basedon the temperature of the FFBs and the temperature of a surface, a temperature difference (AT) is determined.

[0178] At block 1240, upon extracting the temperature difference, the obtained temperature difference is processed by employing one or more machine learning algorithms to categorize the FFB into a number of profile groups. In some embodiments, the one or more machine learning algorithms may be classification algorithms.

[0179] At block 1250, the FFBs are categorized into different profile groups based on an output of the classification algorithm.

[0180] FIG. 13 shows a flow diagram of an example grading method implemented by the hyperspectral imaging module, in accordance with some embodiments of the present disclosure.

[0181] At block 1310, a hyperspectral image of the FFBs is captured by the spectral imaging module.

[0182] At block 1320, a background is removed from the captured hyperspectral image of the FFB. In one embodiment, the background is removed from the captured hyperspectral image using a pixel discrimination and a pixel spectral processing. The background is removed from the captured images by the spectral imaging module 226, thereby detecting the FFBs spectral response.

[0183] At block 1330, the detected FFBs spectral response are processed using advanced imaging processing algorithms and using a low pass filter to remove the noise component from the detected FFBs spectral response.

[0184] At block 1340, upon removing the noise from the FFBs spectral response, the spectral imaging module 226 is configured to process the FFB spectral response by employing one or more machine learning algorithms to categorize the FFB into a number of groups. In some embodiments, the one or more machine learning algorithms may be a classification algorithm which is based on an artificial neural network (ANN).

[0185] At block 1350, the FFBs are grouped into different profile groups based on an output of the one or more machine learning algorithms.

[0186] FIG. 14 shows an exemplary temperature plot associated with a training dataset for training the thermal imaging module 224, in accordance with some embodiments of the present disclosure.

[0187] As shown in Fig. 14, an exemplary temperature plot associated with the training dataset for training the thermal imaging module 224 is described. The exemplary temperature plot represents a ripeness level correction with the temperature difference. Further, as shown in Fig. 14, the FFBs which are having in the training dataset, average value around 3 are categorized as under ripe. Similarly, the FFBs which are having in the training dataset, average value around 2 are categorized as ripe, the FFBs which are having in the training dataset, average value around 1 are categorized as over ripe.

[0188] FIG. 15 shows an exemplary Free Fatty Acid (FFA) plot associated with a training dataset for training the thermal imaging module 224, in accordance with some embodiments of the present disclosure.

[0189] As shown in Fig. 15, the exemplary FFA plot associated with the training dataset for training the thermal imaging module 224 is described. The exemplary FFA plot represents a relationship between the FFA and the temperature difference. The training dataset may be used to train the thermal imaging module 224, thereby accurately capturing the FFA level associated with the FFBs in real-time.

[0190] FIG. 16 shows an exemplary Deterioration of Bleachability Index (DOB1) plot for training the thermal imaging module 224, in accordance with some embodiments of the present disclosure.

[0191] As shown in Fig. 16, the exemplary DOBI plot associated with the training dataset for training the thermal imaging module 224 is described. The exemplary DOBI plot represents a relationship between the DOBI and the temperature difference. The training dataset may be used to train the thermal imaging module 224, thereby accurately capturing the DOBI level associated with the FFBs in real-time.

[0192] FIG. 17 shows an exemplary PV plot associated with a training dataset for training the thermal imaging module 224, in accordance with some embodiments of the present disclosure.

[0193] As shown in Fig. 17, the exemplary PV plot associated with the training dataset for training the thermal imaging module 224 is described. The exemplary plot represents a relationship between the PV and the temperature difference. The training dataset may be used to train the thermal imaging module 224, thereby accurately capturing the PV associated with the FFBs in real-time.

[0194] Fruit plantations are paid by mills based on grade of FFBs. The automated grading and sorting of FFBs as disclosed herein allows mill owners or operators to match manual grading of FFBs as indicated by plantations or vehicle drivers, which bring FFBs to the mill. In case of any discrepancies, the mill owners or operators may provide grading results of the automated system to the vehicle drivers, which may then become a basis for payment to the plantations. Moreover, the FFBs that are determined or graded as rotten can be returned back to the vehicles in real-time, thereby reducing the losses of mill owners or operators.

[0195] The methods, systems, and CRSMs described herein may include the features or perform the functions described herein in association with any one or more of the other methods, systems, and CRSMs described herein.

[0196] The above description of shown example implementations, including what is described in the Abstract, is not intended to be exhaustive or to limit the implementations to the precise forms disclosed. Although specific implementations of and examples are described herein for illustrative purposes, various equivalent modifications can be made without departing from the spirit and scope of the disclosure, as will be recognized by those skilled in the relevant art. Moreover, the various example implementations described herein may be combined to provide further implementations.

[0197] In some embodiments, the method or methods described above may be executed or carried out by a computing system (for example, the automated grading and sorting system 102) including a tangible computer-readable storage medium, also described herein as a storage machine, that holds machine -readable instructions executable by a logic machine (e.g., a processoror programmable control device) to provide, implement, perform, and / or enact the above described methods, processes and / or tasks. When such methods and processes are implemented, the state of the storage machine may be changed to hold different data. For example, the storage machine may include memory devices such as various hard disk drives, CD, or DVD devices. The logic machine may execute machine -readable instructions via one or more physical information and / or logic processing devices. For example, the logic machine may be configured to execute instructions to perform tasks for a computer program. The logic machine may include one or more processors to execute the machine-readable instructions. The computing system may include a display subsystem to display a graphical user interface (GUI), or any visual element of the methods or processes described above. For example, the display subsystem, storage machine, and logic machine may be integrated such that the above method may be executed while visual elements of the disclosed system and / or method are displayed on a display screen for user consumption. The computing system may include an input subsystem that receives user input. The input subsystem may be configured to connect to and receive input from devices such as a mouse, keyboard, or gaming controller. For example, a user input may indicate a request that certain task is to be executed by the computing system, such as requesting the computing system to display any of the above-described information or requesting that the user input updates or modifies existing stored information for processing. A communication subsystem may allow the methods described above to be executed or provided over a computer network. For example, the communication subsystem may be configured to enable the computing system to communicate with a plurality of personal computing devices. The communication subsystem may include wired and / or wireless communication devices to facilitate networked communication. The described methods or processes may be executed, provided, or implemented for a user or one or more computing devices via a computer-program product such as via an API.

[0198] Since many modifications, variations, and changes in detail can be made to the described preferred embodiments of the disclosure, it is intended that all matters in the foregoing description and shown in the accompanying drawings be interpreted as illustrative and not in a limiting sense. Thus, the scope of the invention should be determined by the appended claims and their legal equivalents.

Claims

What is claimed is:

1. An automated system for grading and sorting fresh fruit bunches (FFBs), comprising: an input module configured to receive a plurality of FFBs; a grading module, operatively coupled to the input module, and configured to determine one or more quality indicator characteristics of the plurality of FFBs by performing analysis on the plurality of FFBs, wherein the analysis comprises at least one of: color and thermal imaging analysis, reflection analysis, and weight measurement analysis; and a sorting module, operatively coupled to the grading module, and configured to categorize the plurality of FFBs into a number of groups based at least on the determined one or more quality indicator characteristics.

2. The system as claimed in claim 1 , wherein the sorting module is further configured to: obtain user-defined profile data associated with the categorization of the FFBs; compare the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data; and based at least upon the comparison, categorize the plurality of FFBs into the number of groups.

3. The system as claimed in claim 2, wherein to obtain user-defined profile data , the grading module is configured to: obtain a training set, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs; obtain the one or more quality indicator characteristics for the test group of FFBs, wherein the one or more quality indicator characteristics are obtained from a user based on manual inspection and lab reports for the test group of FFBs; and determine the user-defined profile data based on the obtained training set and the obtained one or more quality indicator characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results.

4. The system as claimed in claim 3, wherein to determine the user-defined profile data, the grading module is configured to apply the obtained training set and the obtained one or more quality indicator characteristics to a machine learning algorithm to determine the acceptable range.

5. The system as claimed in any of the previous claims, wherein the grading module is configured to employ a machine learning algorithm to determine the one or more quality indicator characteristics.

6. The system as claimed in claim 5, wherein the grading module is further configured to employ the machine learning algorithm to determine a respective sterilization recipe for each of the number of groups, wherein the sterilization recipe comprise at least one of a time value, pressure value, and a temperature value, which may be employed as process parameters in a sterilization center, to maximize oil extraction rate (OER) from the FFBs.

7. The system as claimed in any of the previous claims, wherein the grading module is configured to calculate at least one value that reflects at least one quality indicator characteristic of each of the plurality of FFBs based on one or more of: the color and thermal imaging analysis, the reflection analysis, and the weight measurement analysis.

8. The system as claimed in claim 7, wherein the at least one value comprises one or more of: a first value that indicates a ripeness of the FFB, a second value that indicates an oil quality of the FFB, and a third value that indicates a weight of the FFB.

9. The system as claimed in any of the previous claims, wherein the grading module comprises: a first imaging device that captures visible color and thermal images of the plurality of FFBs to perform the color and thermal imaging analysis of the plurality of FFBs; a second imaging device that captures a reflection response of the plurality of FFBs to illuminated light from the second imaging device to perform the reflection analysis, wherein the second imaging device comprises at least one of: a hyperspectral imaging device and a near-infrared (NIR) imaging device; anda weight load cell that measures a weight of each of the plurality of FFBs.

10. The system as claimed in any of the previous claims, wherein the one or more quality indicator characteristics comprise at least one of: a ripeness level, an aesthetic condition, one or more diseases associated with the FFB, an oil quality of the FFB, a brix percentage, species, and firmness of the FFB.

11. An automated method for grading and sorting fresh fruit bunches (FFBs), comprising: receiving, by an automated system, a plurality of FFBs; determining, by the automated system, one or more quality indicator characteristics of the plurality of FFBs by performing analysis on the plurality of FFBs, wherein the analysis comprises at least one of: color and thermal imaging analysis, reflection analysis, and weight measurement analysis; and categorizing, by the automated system, the plurality of FFBs into a number of groups based at least on the determined one or more quality indicator characteristics.

12. A computer-implemented method for handling Fresh Fruit Bunches (FFBs), the method comprising: determining one or more quality indicator characteristics of a plurality of FFBs; receiving user-defined profile data associated with categorization of the FFBs; comparing the determined one or more quality indicator characteristics of the plurality of FFBs with the user-defined profile data; and based at least upon the comparison, categorizing the plurality of FFBs into a number of groups.

13. The method as claimed in claim 12, wherein determining the one or more quality indicator characteristics comprises determining one or more of: species, ripeness level, aesthetic condition, diseases, oil quality, brix percentage, and firmness of the plurality of FFBs.

14. The method as claimed in claim 12 or claim 13, wherein determining the one or more quality indicator characteristics comprises performing one or more of image analysis, reflection analysis, and weight analysis on the plurality of FFBs.

15. The method as claimed in claim 14, wherein performing the image analysis comprises performing one or more of color image analysis and thermal imaging analysis.

16. The method as claimed in claim 14 or claim 15, wherein performing the color image analysis comprises analyzing one or more of: color range, hue, color variance, appearance, color coverage, color coverage group of fruitlets on the FFBs.

17. The method as claimed in any of claims 14 to 16, wherein performing the color image analysis comprises determining at least some information about the one or more quality indicator characteristics including one or more of a ripeness level and a size of the plurality of FFBs.

18. The method as claimed in any of claims 14 to 17, wherein performing the thermal image analysis comprises performing a comparison between a temperature profile of the FFBs, and atmospheric temperature.

19. The method as claimed in any of claims 14 to 18, wherein receiving the user-defined profile data comprises: receiving a training set, wherein the training set comprises at least one of color images, thermal images, and spectral images for a test group of FFBs; receiving the one or more quality indicator characteristics for the test group of FFBs, wherein the one or more quality indicator characteristics are received from a user based on manual inspection and lab reports for the test group of FFBs; and determining the user-defined profile data based on the received training set and the received one or more characteristics, wherein the user-defined profile data comprises an acceptable range associated with each of image analysis results, thermal analysis results, and spectral analysis results.

20. The method as claimed in claim 19, wherein determining the user-defined profile data comprises applying the received training set and the received one or more quality indicator characteristics to a machine learning algorithm to determine the acceptable range.

21. A non-transitory computer-readable storage medium comprising instructions, which in response to execution by one or more processors cause the one or more processors to perform or control performance of any of the methods disclosed herein.

Images