A crop sorting device and control method

Abstract

This invention belongs to the field of sorting equipment technology and discloses a crop sorting device and control method. The device includes a sorting platform and a core control unit. A platform conveyor belt is installed on top of the sorting platform. A mobile sorting device is installed above the platform conveyor belt. The mobile sorting device includes two three-axis linear slides. A gripping device and a detection device are respectively installed on the final motion platform of the two three-axis linear slides. The three-axis linear slides, the detection device, the gripping device, and the detection device are all electrically connected to the core control unit. The detection device integrates a depth camera, a near-infrared spectral sensor, and a multi-dimensional vision acquisition module. This invention can adapt to the detection and sorting needs of various crops through parameterized configuration and can flexibly add detection indicators or optimize sorting logic. Through real-time data interconnection with the cloud, it can provide support for precise control of the entire chain and improve the efficiency and accuracy of detection and sorting.

Description

Technical Field

[0001] This invention belongs to the field of sorting equipment technology, specifically relating to a sorting device and control method for crops. Background Technology

[0002] Traditional mechanical sorting equipment often uses a single detection dimension for sorting, such as grading based solely on physical parameters like weight and size, or detecting appearance defects using only visual sensors. It lacks the comprehensive detection capabilities for multiple dimensions of agricultural products, including internal quality (such as sweetness and moisture content), pest and disease types, and spatial orientation, resulting in insufficient detection accuracy and grading precision. Furthermore, traditional equipment is often a fixed model adapted to a single crop, lacking flexible parameter configuration and hardware expansion capabilities. This makes it impossible to meet the detection and sorting needs of different types of regularly shaped crops, and it lacks data storage, traceability, and end-to-end data interaction functions, hindering the achievement of end-to-end quality control for agricultural products from planting to sales. Chinese patent publication number CN110125044A, entitled "A Binocular Vision-Based Apple Sorting Device," includes a control system comprising a host computer and a controller. The apple sorting device also includes a conveyor, with a feeder on one side of the conveyor and a detection box on the other side, located on top of the conveyor and fixedly connected to it. A photoelectric sensor is installed on the side of the detection box near the feeder. The detection box contains a detection light and a camera (a binocular camera) located on top of the detection box. A sorter and multiple sorting boxes are located on the side of the detection box opposite the feeder. The photoelectric sensor, camera, and sorter are all connected to the control system. This patent application is difficult to apply to sorting various types of crops and struggles to achieve accurate detection of internal quality and subtle defects in agricultural products. Summary of the Invention

[0003] To overcome the problems existing in the prior art, the present invention aims to provide a crop sorting device and control method. By integrating a depth camera, a near-infrared spectral sensor, and a multi-dimensional visual acquisition module, and configuring target detection algorithms and multimodal data fusion algorithms within the data analysis submodule, it supports parameterized configuration to adapt to the detection and sorting needs of various crops such as fruits, vegetables, and root vegetables. It allows for flexible addition of detection indicators or optimization of sorting logic. Through real-time cloud data interconnection, it provides support for precise control throughout the entire supply chain, improving detection and sorting efficiency and accuracy, reducing labor costs, and ensuring the consistency and traceability of agricultural product quality.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a crop sorting device, comprising: a sorting platform and a core control unit, wherein a platform conveyor belt is disposed on the top of the sorting platform; a mobile sorting device is disposed above the platform conveyor belt, the mobile sorting device comprising two three-axis linear slides, and a gripping device and a detection device are respectively disposed on the final motion platform of the two three-axis linear slides; the three-axis linear slides, the detection device, the gripping device, and the detection device are all electrically connected to the core control unit; the core control unit has a data processing module, the data processing module being communicatively connected to a wireless communication module, and the data processing module having... The system includes a data preprocessing submodule and a data analysis submodule. The data analysis submodule is equipped with a target detection algorithm and a multimodal data fusion algorithm. The detection device integrates a depth camera, a near-infrared spectral sensor, and a multi-dimensional visual acquisition module. The depth camera is used to obtain crop shape regularity parameters, spatial posture information, and external geometric contour features. The near-infrared spectral sensor is used to acquire the crop's sweetness content, moisture content, soluble solids content, and degree of internal defects. The multi-dimensional visual acquisition module includes at least two visual sensors for identifying crop maturity levels, pest and disease types, and surface defect information.

[0005] Optionally, the two three-axis linear slides include: two slide rails parallel to both sides of the platform conveyor belt, and a detection device moving track and a gripping device moving track perpendicular to the slide rails; linear motors are provided at both ends of the detection device moving track and both ends of the gripping device moving track, and the detection device moving track and the gripping device moving track are slidably mounted on the slide rails via the linear motors at both ends; linear motors are slidably mounted on both the detection device moving track and the gripping device moving track; a detection vertical rod is provided on the linear motor of the detection device moving track, and a gripping vertical rod is provided on the linear motor of the gripping device moving track; linear motors are slidably mounted on both the detection vertical rod and the gripping vertical rod, and the detection device and the gripping device are respectively mounted on the bottom of the linear motors on the detection vertical rod and the gripping vertical rod.

[0006] Optionally, the gripping device is a flexible gripper.

[0007] Optionally, the sorting platform is provided with a feeding track mechanism and a conveying track mechanism at both ends along the conveying direction of the platform conveyor belt, and the conveying track mechanism is provided with a plurality of diverting plates.

[0008] Optionally, the core control unit adopts an embedded edge computing core based on Raspberry Pi, and the data processing module and the target detection algorithm are both deployed within this edge computing core.

[0009] Optionally, the data preprocessing submodule is configured to perform noise reduction processing on the original data using a wavelet transform noise reduction algorithm.

[0010] Optionally, the wireless transmission module may be a 5G communication module or a Wi-Fi module.

[0011] Secondly, the present invention provides a control method for a crop sorting device, comprising the following steps: The platform conveyor belt transports the crops to be sorted to the detection area. When the infrared sensor of the detection device detects the entry of the crops, it sends the detection signal to the core control unit. The system collects crop shape regularity parameters, spatial posture information, and external geometric contour features using a depth camera; it collects crop sweetness content, moisture content, soluble solids content, and internal defect degree using a near-infrared spectral sensor; it captures crop images from different angles using a multi-dimensional visual acquisition module, and analyzes the images using a data processing module to identify crop maturity level, pest and disease types, and surface defect information. The data preprocessing submodule performs noise reduction, normalization and feature extraction on the collected raw data, and uploads the preprocessed multimodal data to the cloud through the wireless transmission module. It calls the standard parameters in the cloud agricultural product standard feature library, and uses a multimodal data fusion algorithm to comprehensively score the data collected by the depth camera, near-infrared spectral sensor and multi-dimensional vision acquisition module to generate comprehensive detection results for the corresponding crops. Control the gripping device to move to the target location to grab the crop, and transfer the crop to the corresponding diversion area; The raw data and comprehensive test results of the crop are uploaded to the cloud via a wireless communication module as an electronic tag for the crop.

[0012] Optionally, by controlling the three-axis linear slide to adjust the planar and height positions of the detection device, data on the target crop at different angles and positions can be collected.

[0013] Optionally, the data preprocessing submodule uses a wavelet transform denoising algorithm to denoise the original data and uses the Min-Max normalization method to normalize the data of different dimensions.

[0014] Compared with the prior art, the present invention has the following beneficial effects: This invention integrates a near-infrared spectral sensor to achieve full-dimensional non-contact detection of the geometric features of agricultural products, internal quality indicators, pest and disease types, and surface defects. The comprehensive detection dimensions avoid the inaccurate grading problems caused by a single detection dimension. The sweetness detection accuracy is no less than ±0.5°Bx, the moisture content detection accuracy is no less than ±1%, and the surface defect area detection accuracy is no less than 0.1cm², significantly improving the detection accuracy and grading accuracy. The device uses a Raspberry Pi as its core control unit, along with a data preprocessing unit, a data transmission unit, and a data analysis unit. It employs wavelet transform noise reduction, principal component analysis feature extraction, and weighted average fusion algorithms to achieve efficient processing and real-time analysis of multimodal data. The high data transmission rate ensures the real-time nature of the detection results. Combined with a high-precision mechanical sorting device, the sorting efficiency is 5-8 times higher than manual sorting, significantly reducing labor costs. The equipment supports the detection and sorting needs of various regular-shaped crops such as fruits, vegetables, and root vegetables through parameterized configuration. It has built-in templates of various crop characteristic parameters, which users can directly call or customize. It also reserves hardware expansion interfaces and algorithm upgrade channels, which can flexibly add detection indicators or optimize sorting logic. It has strong adaptability and outstanding scalability, eliminating the need to purchase separate equipment for different crops and reducing equipment investment costs. This invention assigns a unique electronic tag to each agricultural product, incorporating testing data, traceability information, and equipment operation information. Combined with a cloud management platform, it enables end-to-end traceability of agricultural products. Users can query information about the entire process, including planting, harvesting, testing, and sorting, through the electronic tag. At the same time, the cloud platform synchronizes quality data with the planting, warehousing, and logistics ends, providing data support for precise end-to-end control and ensuring the consistency and safety of agricultural product quality. The cloud management platform has data storage, data visualization, and early warning functions. It can realize long-term storage and real-time monitoring of inspection and sorting data, and send early warning signals in a timely manner when agricultural products have abnormal quality or equipment malfunctions, so that relevant personnel can deal with them quickly. At the same time, through data accumulation and analysis, it can provide data reference for the large-scale and standardized development of the agricultural product industry. Attached Figure Description

[0015] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. Furthermore, the shapes and proportions of the components in the drawings are merely schematic to aid in understanding the invention and do not specifically limit the shapes and proportions of the components. In the drawings: Figure 1 This is a top view of the device according to an embodiment of the present invention; Figure 2 This is a left view of the device according to an embodiment of the present invention; Figure 3 This is a three-dimensional structural diagram of the device according to an embodiment of the present invention; The components include: 1. Feeding track mechanism; 2. Slide rail; 3. Detection device moving track; 4. Gripping device moving track; 5. Feeding track mechanism; 6. Diverter plate; 7. Display screen; 8. Gripping vertical rod; 9. Gripping device; 10. Drive motor; 11. Computing unit starter box; 12. Detection device; 13. Detection vertical rod. Detailed Implementation To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0016] Therefore, the following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention.

[0017] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper", "lower", "horizontal", "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product of the invention is usually placed during use, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0018] When an element is referred to as being "set on" another element, it can be directly on the other element or there may be an intervening element. When an element is referred to as being "connected to" another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only and do not represent the only embodiments. The use of the term "horizontal" does not imply that the component is required to be absolutely horizontal, but rather that it may be slightly tilted. "Horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but rather that it may be slightly tilted.

[0019] It should be noted that similar reference numerals and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification and appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0021] The present invention will now be described in detail with reference to the accompanying drawings.

[0022] With the industrialization and large-scale development of agriculture, the quality grading and sorting of agricultural products have become key links affecting the added value, market competitiveness, and supply chain efficiency of agricultural products. Regularly shaped crops (such as apples, carrots, pears, tomatoes, potatoes, etc.) are among the agricultural product categories with the largest market circulation. Their sorting process not only needs to focus on the regularity of appearance, but also needs to accurately detect internal quality indicators, surface defects, and pests and diseases, in order to meet consumers' demand for the consistency and safety of agricultural product quality.

[0023] Currently, the sorting of regularly shaped crops mainly relies on manual sorting or traditional mechanical sorting equipment. Manual sorting is affected by factors such as the operator's experience, physical strength, and subjective judgment, resulting in problems such as low sorting efficiency, high labor intensity, high labor costs, and inconsistent grading standards. Furthermore, it is difficult to accurately detect the internal quality and subtle defects of agricultural products, leading to inconsistent quality of sorted agricultural products that cannot meet the demands of large-scale production and high-quality markets.

[0024] Although some existing intelligent sorting equipment integrates multiple sensors, it suffers from low data processing efficiency and immature multimodal data fusion algorithms, resulting in delayed detection results that cannot meet the needs of real-time sorting. At the same time, the lack of a unified cloud-based standard feature library makes it difficult to guarantee the objectivity and accuracy of detection results, and the absence of a sound agricultural product traceability system makes it impossible to achieve product quality traceability and responsibility determination.

[0025] A crop sorting device of the present invention includes: a sorting platform and a core control unit. A platform conveyor belt is disposed on top of the sorting platform; a mobile sorting device is disposed above the platform conveyor belt, the mobile sorting device including two three-axis linear slides, and a gripping device 9 and a detection device 12 are respectively disposed on the final motion platform of the two three-axis linear slides; the three-axis linear slides, the detection device 12, the gripping device 9, and the detection device 12 are all electrically connected to the core control unit; the core control unit has a data processing module, the data processing module being communicatively connected to a wireless communication module, and the data processing module having… The system includes a data preprocessing submodule and a data analysis submodule. The data analysis submodule is equipped with a target detection algorithm and a multimodal data fusion algorithm. The detection device 12 integrates a depth camera, a near-infrared spectral sensor, and a multi-dimensional visual acquisition module. The depth camera is used to obtain crop shape regularity parameters, spatial posture information, and external geometric contour features. The near-infrared spectral sensor is used to acquire the crop's sweetness content, moisture content, soluble solids content, and degree of internal defects. The multi-dimensional visual acquisition module includes at least two visual sensors for identifying crop maturity levels, pest and disease types, and surface defect information.

[0026] The control method for a crop sorting device includes the following steps: The crops to be sorted are transported to the detection area via the platform conveyor belt. When the infrared sensor of the detection device 12 detects the entry of the crops, it sends the detection signal to the core control unit. The system collects crop shape regularity parameters, spatial posture information, and external geometric contour features using a depth camera; it collects crop sweetness content, moisture content, soluble solids content, and internal defect degree using a near-infrared spectral sensor; it captures crop images from different angles using a multi-dimensional visual acquisition module, and analyzes the images using a data processing module to identify crop maturity level, pest and disease types, and surface defect information. The data preprocessing submodule performs noise reduction, normalization and feature extraction on the collected raw data, and uploads the preprocessed multimodal data to the cloud through the wireless transmission module. It calls the standard parameters in the cloud agricultural product standard feature library, and uses a multimodal data fusion algorithm to comprehensively score the data collected by the depth camera, near-infrared spectral sensor and multi-dimensional vision acquisition module to generate comprehensive detection results for the corresponding crops. Control the gripping device 9 to move to the target position to grab the crop, and transfer the crop to the corresponding diversion area; The raw data and comprehensive test results of the crop are uploaded to the cloud via a wireless communication module as an electronic tag for the crop.

[0027] This invention provides a crop sorting device with multi-dimensional feature comprehensive detection, parameterized configuration adaptation, cloud data interconnection and full-chain traceability functions.

[0028] Example 1 The present invention provides a crop sorting device, comprising a feeding conveyor belt mechanism 1, a conveyor belt mechanism 5, and a sorting platform.

[0029] The sorting platform has a drive roller and a driven roller at its top two ends, respectively. A platform conveyor belt, specifically a rubber conveyor belt, is fitted over the drive roller and driven roller. A drive motor 10 is installed on the sorting platform, and the output shaft of the drive motor 10 is connected to the drive roller.

[0030] The sorting platform has two slide rails 2, a detection device moving track 3, and a gripping device moving track 4 on both sides of the platform conveyor belt at the top. The two slide rails 2 are parallel to each other and parallel to the transmission direction of the feeding conveyor mechanism 1 and the conveying conveyor mechanism 5. The feeding conveyor mechanism 1 and the conveying conveyor mechanism 5 are respectively located at both ends of the sorting platform along the direction of the slide rails 2.

[0031] Linear motors are provided at both ends of the moving track 4 of the gripping device, and linear motors are provided at both ends of the moving track 3 of the detection device; the linear motors at both ends of the moving track 4 of the gripping device and the linear motors at both ends of the moving track 3 of the detection device are slidably mounted on the slide rail 2.

[0032] The moving track 3 of the detection device and the moving track 4 of the gripping device are both perpendicular to the slide rail 2.

[0033] Linear motors are slidably mounted on both the detection device moving track 3 and the gripping device moving track 4. The linear motors on the detection device moving track 3 and the gripping device moving track 4 are respectively capable of moving along the axial direction of the detection device moving track 3 and the gripping device moving track 4.

[0034] A detection vertical rod 13 is mounted on the linear motor of the moving track 3 of the detection device; a gripping vertical rod 8 is mounted on the linear motor of the moving track 4 of the gripping device. Linear motors are slidably mounted on both the gripping vertical rod 8 and the detection vertical rod 13.

[0035] The bottom of the linear motor that grips the vertical rod 8 is connected to a gripping device 9; the bottom of the linear motor that detects the vertical rod 13 is provided with a detection device 12.

[0036] Specifically, the gripping device 9 is a flexible gripper.

[0037] The feeding track mechanism 5 is provided with a plurality of diverting plates 6. Specifically, side plates are provided on both sides of the feeding track mechanism 5 in the conveying direction, and a fixed connector is provided between the two side plates. The fixed connector is a rod or a plate. The two ends of the fixed connector are respectively connected to the side plates on both sides of the feeding track mechanism 5, and the top of the diverting plate 6 is connected to the fixed connector.

[0038] A core control unit is located below the sorting platform. The core control unit has a data processing module, which is communicatively connected to a wireless communication module.

[0039] Optionally, in this embodiment, the core control unit is a Raspberry Pi.

[0040] Optionally, in this embodiment, the detection device 12 integrates a depth camera, a near-infrared spectral sensor, and a multi-dimensional visual acquisition module, enabling the acquisition of full-dimensional feature data of regularly shaped crops through a non-contact detection method.

[0041] The linear motors on the slide rail 2, the detection device moving track 3, the gripping device moving track 4, the gripping vertical rod 8, and the detection vertical rod 13 are all electrically connected to the core control unit, and the detection device 12, the drive motor 10, and the gripping device 9 are all electrically connected to the core control unit.

[0042] Optionally, the wireless transmission module uses a 5G communication module or a Wi-Fi module to achieve real-time data interaction with the cloud database.

[0043] The data processing module includes a data preprocessing submodule and a data analysis submodule. Data preprocessing is used to perform noise reduction, normalization, and feature extraction on the raw data acquired by the near-infrared spectroscopy sensor.

[0044] The data analysis submodule is equipped with a target detection algorithm and a multimodal data fusion algorithm.

[0045] Optionally, the data preprocessing submodule uses wavelet transform denoising algorithm to denoise the original data, uses Min-Max normalization method to unify data of different dimensions to the [0,1] interval, and uses principal component analysis algorithm to extract features and remove redundant data.

[0046] Multimodal data fusion algorithms are used to comprehensively analyze appearance geometric feature data, internal quality index data, and surface defect data, and compare them with threshold parameters in the cloud-based agricultural product standard feature library to generate quantitative comprehensive test results. The data processing module uploads the collected multimodal feature data to a cloud database in real time via a wireless communication module. By comparing and analyzing the data with a pre-set standard feature library of agricultural products in the cloud, a comprehensive test result is generated, including maturity level, sweetness index, pest and disease severity, and shape rating. Each agricultural product to be tested is assigned a unique electronic tag, which contains the corresponding original test data, grading result, traceability identifier, and equipment operation information. Based on the grading information stored in the electronic tag, automated grading and sorting of agricultural products is achieved, and products that fail the test are isolated separately.

[0047] The core control unit transmits parameters to the grasping device 9, enabling parameterized configuration to adapt to the detection and sorting needs of various regularly shaped agricultural products such as fruits, vegetables, and root vegetables. The core control unit also has reserved hardware expansion interfaces and algorithm upgrade channels, allowing for flexible addition of detection indicators or optimization of sorting logic. Real-time cloud data interconnection is achieved through a wireless transmission module, synchronizing agricultural product quality data with the planting, warehousing, and logistics ends. This provides data support for precise control of the entire agricultural product supply chain, improving detection and sorting efficiency and accuracy, reducing labor costs, and ensuring the consistency and traceability of agricultural product quality.

[0048] The depth camera is used to accurately capture the shape regularity parameters, spatial posture information, and appearance geometric contour features of regularly shaped crops. The shape regularity parameters include the crop's size deviation, symmetry, corner roundness, and surface flatness. The spatial posture information includes the crop's placement angle, height coordinates, and spatial position parameters within the detection area. The appearance geometric contour features include the crop's contour integrity, edge smoothness, and geometric parameter consistency. The near-infrared spectral sensor accurately acquires the internal quality indicators of regularly shaped crops through near-infrared spectral detection technology. The internal quality indicators include sweetness content, moisture content, soluble solids content, and degree of internal defects.

[0049] The multi-dimensional visual acquisition module includes at least two visual sensors. These visual sensors, combined with a target detection algorithm preset by the data processing module, enable the identification of maturity levels, determination of pest and disease types, and acquisition of surface defect parameters for crops with regular shapes.

[0050] Optionally, the target detection algorithm adopts the YOLO algorithm based on deep learning, the pest and disease types include fungal diseases, bacterial diseases and insect damage, and the surface defect parameters include defect area, defect location, defect type and defect severity level. The electronic tag is a virtual tag, and the detection data written in the tag includes shape regularity detection value, sweetness detection value, moisture content detection value, maturity measurement value, pest and disease detection results, and surface defect parameters.

[0051] The parameterized configuration method can be set through the device's built-in touch screen and remote terminal. The configuration parameters include crop type parameters, detection index thresholds, sorting and grading standards, conveyor belt running speed and detection accuracy level. The crop type parameters cover characteristic parameter templates of common regular-shaped crops such as apples, carrots, pears, tomatoes, carrots, and potatoes, which users can directly call or customize. The hardware expansion interface includes a sensor expansion interface, a communication interface, and an actuator expansion interface. The sensor expansion interface supports the addition of Raman spectroscopy sensors, humidity sensors, or pesticide residue sensors. The communication interface supports Ethernet or LoRa communication protocols. The algorithm upgrade channel enables the optimization and updating of target detection algorithms, data fusion algorithms, and sorting control algorithms through remote cloud upgrades or local USB flash drive upgrades. Optionally, the sorting platform is equipped with a display screen 7 on its exterior. The display screen is electrically connected to the core control unit and is used to display parameter information and equipment operation information.

[0052] Example 2 This embodiment takes the intelligent sorting of apples and carrots (regularly shaped fruits) as an example. The core control unit of this embodiment uses a Raspberry Pi 4B (4GB memory version), with an integrated near-infrared spectral sensor. The data transmission unit is equipped with a 5G communication module, a high-precision mechanical sorting device, a food-grade PU material sorting conveyor belt, and hardware expansion interfaces including GPIO interface, I2C interface, RS485 interface, RJ45 Ethernet interface and RS232 interface. The remote terminal supports mobile APP (Android / iOS system) and PC software (Windows / Linux system). Operators access the built-in "Apple and Carrot" crop characteristic parameter templates via the equipment's touchscreen. The preset parameters are as follows: size range 60mm-90mm, size deviation threshold ±3mm, symmetry threshold ≥90%, sweetness threshold ≥10°Bx (high-quality ≥12°Bx, qualified 10-11.9°Bx, products awaiting re-inspection 8-9.9°Bx, <8°Bx is unqualified), moisture content threshold 75%-90% (beyond this range is unqualified), surface defect area threshold ≤0.5cm² (minor defect), 0.5-1.0cm² (moderate defect), >1.0cm² (serious defect, judged as unqualified), maturity level corresponding to sugar content: immature <8°Bx, semi-mature 8-9.9°Bx, mature 10-14°Bx, over-mature >14°Bx, conveyor belt speed set to 0.5m / s, and detection accuracy level set to high. Feeding: Harvested apples and carrots are transported to the sorting platform via the feeding conveyor 1. The platform conveyor belt then transports the apples and carrots to the detection area in sequence. After the infrared sensor of the detection device 12 in the detection area detects the apples and carrots entering, it sends a signal to the core control unit.

[0053] First, the apple and carrot are scanned using a depth camera to collect shape regularity parameters such as size deviation (e.g., actual size 75mm, deviation +1mm), symmetry (e.g., 96%), corner roundness (e.g., 98%), and surface flatness (e.g., 97%), as well as spatial pose information such as placement angle (e.g., 30°), height coordinate (e.g., 15cm), and spatial position parameters (e.g., center position of the detection area). At the same time, appearance geometric contour features such as contour integrity (e.g., 100%) and edge smoothness (e.g., 95%) are acquired.

[0054] By controlling the linear motors on the moving track 3 and at both ends of the detection device, the planar position of the detection device 12 is adjusted. The height position of the detection device 12 is adjusted by the linear motor on the vertical rod 13, so that the multi-dimensional visual acquisition module and the near-infrared spectral sensor can fully scan every angle of the target.

[0055] The near-infrared spectroscopy sensor emits near-infrared light with wavelengths of 900nm-1700nm. After penetrating the skin of apples and carrots, it receives the reflected spectrum. The sweetness content (e.g., 12.8°Bx), moisture content (e.g., 86%), and soluble solids content (e.g., 13.5%) of apples and carrots are calculated by spectral analysis algorithms, and the degree of internal defects is detected. The multi-dimensional visual acquisition module uses two visual sensors to capture images of apples and carrots from the front and side, respectively. The images are then transmitted to the data processing module, where the YOLO target detection algorithm is used to analyze the images, identify the maturity level of the apples and carrots, determine the type of pests and diseases, and collect surface defect parameters.

[0056] The data preprocessing unit performs wavelet transform noise reduction on the collected raw data to remove environmental interference noise. It then uses the Min-Max normalization method to unify the data of different dimensions to the [0, 1] interval and uses principal component analysis to extract key features. The data transmission unit uploads the preprocessed multimodal data to the cloud database in real time through the 5G communication module. The data analysis unit calls the standard parameters of apples and carrots in the cloud agricultural product standard feature library and uses a weighted average fusion algorithm to comprehensively score the appearance geometric feature data, internal quality index data, and surface defect data (e.g., 92 points) to generate a comprehensive test result.

[0057] After the detection device 12 completes the detection, it controls the drive motor 10 to drive the platform conveyor belt to transport the apples and carrots to the sorting area.

[0058] Based on the grading results in the electronic tag, the core control unit sends control signals to the linear motors on the moving track 3 of the detection device and at both ends, as well as the linear motors on the gripping vertical rod 8, to move the gripping device 9 to the target position (positioning accuracy error within 0.1mm).

[0059] After picking up the apples and carrots separately, they are moved to the corresponding positions marked by the diversion plate 6. The diversion plate 6 divides the feeding conveyor mechanism 5 into different areas, each corresponding to a different hopper position.

[0060] Apples and carrots are transported to their respective grade grading bins via a platform conveyor belt and feeding conveyor mechanism 5. If the test result is unqualified (e.g., sweetness 7.5°Bx, surface defect area 1.2cm²), the product is pushed to the unqualified product isolation bin, where the warning light automatically illuminates and the ultraviolet disinfection module is activated.

[0061] The testing data and grading results of apples and carrots are transmitted to the cloud via a wireless transmission module, and then synchronized to the planting, storage, and logistics ends via the cloud. Growers can adjust their fertilization and irrigation plans for the next season based on sweetness and moisture content data. Storage managers can set the storage temperature to 0-2℃ and the humidity to 85%-90% based on the moisture content data. The logistics end prioritizes the transportation of this batch of apples and carrots according to their maturity level. Optionally, a pesticide residue sensor (such as the QT-P01 pesticide residue rapid detector) can be connected to the detection device 12 via a sensor expansion interface (GPIO interface) to incorporate the pesticide residue detection data into the scoring system of the multimodal data fusion algorithm.

[0062] Optionally, the YOLOv8 algorithm upgrade package can be downloaded from the cloud via a wireless transmission module to optimize the target detection algorithm. The crop sorting equipment of this invention achieves full feature detection, intelligent grading and sorting, and full-chain traceability of regular-shaped crops through multi-dimensional sensor integration, efficient data processing, cloud data interconnection, and high-precision mechanical execution. It is highly adaptable and flexible in expansion, and can be widely used in the post-harvest processing of various regular-shaped crops, effectively improving sorting efficiency and accuracy, reducing labor costs, and providing technical support for the large-scale and standardized development of the agricultural product industry.

[0063] Unless otherwise specified, the equipment components involved in the above embodiments are all conventional equipment components, and the structural settings, working methods or control methods involved are all conventional settings, working methods or control methods in the art unless otherwise specified.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Any other modifications or equivalent substitutions made by those skilled in the art to the technical solutions of the present invention, as long as they do not depart from the spirit and scope of the technical solutions of the present invention, should be covered within the scope of the claims of the present invention.

Claims

1. A crop sorting device, characterized in that, include: The sorting platform and the core control unit are described above. A platform conveyor belt is installed on top of the sorting platform. A mobile sorting device is installed above the platform conveyor belt. The mobile sorting device includes two three-axis linear slides. A gripping device (9) and a detection device (12) are respectively installed on the final motion platform of the two three-axis linear slides. The three-axis linear slides, the detection device (12), the gripping device (9), and the detection device (12) are all electrically connected to the core control unit. The core control unit has a data processing module, which is communicatively connected to a wireless communication module. The data processing module has data preprocessing capabilities. The detection device (12) includes a depth camera, a near-infrared spectral sensor, and a multi-dimensional visual acquisition module. The depth camera is used to obtain crop shape regularity parameters, spatial posture information, and appearance geometric contour features. The near-infrared spectral sensor is used to obtain the sweetness content, moisture content, soluble solids content, and internal defect degree of the crop. The multi-dimensional visual acquisition module includes at least two visual sensors for identifying crop maturity level, pest and disease type, and surface defect information.

2. The crop sorting equipment according to claim 1, characterized in that, The two three-axis linear slides include: two slide rails (2) arranged parallel to both sides of the platform conveyor belt, and a detection device moving track (3) and a gripping device moving track (4) perpendicular to the slide rails (2); both ends of the detection device moving track (3) and both ends of the gripping device moving track (4) are provided with linear motors, and the detection device moving track (3) and the gripping device moving track (4) are slidably mounted on the slide rails (2) through the linear motors at both ends; both the detection device moving track (3) and the gripping device moving track (4) are slidably mounted with linear motors; a detection vertical rod (13) is provided on the linear motor of the detection device moving track (3), and a gripping vertical rod (8) is provided on the linear motor of the gripping device moving track (4); both the detection vertical rod (13) and the gripping vertical rod (8) are slidably mounted with linear motors, and the detection device (12) and the gripping device (9) are respectively mounted on the bottom of the linear motors on the detection vertical rod (13) and the gripping vertical rod (8).

3. The crop sorting equipment according to claim 1, characterized in that, The gripping device (9) is a flexible gripper.

4. The crop sorting equipment according to claim 1, characterized in that, The sorting platform is provided with a feeding conveyor belt mechanism (1) and a feeding conveyor belt mechanism (5) at both ends along the conveying direction of the platform conveyor belt. The feeding conveyor belt mechanism (5) is provided with a number of diversion plates (6).

5. A crop sorting device according to claim 1, characterized in that, The core control unit adopts an embedded edge computing core based on Raspberry Pi, and the data processing module and the target detection algorithm are both deployed within this edge computing core.

6. The crop sorting equipment according to claim 1, characterized in that, The data preprocessing submodule is configured to use a wavelet transform denoising algorithm to denoise the original data.

7. A crop sorting device according to claim 1, characterized in that, The wireless transmission module adopts a 5G communication module or a Wi-Fi module.

8. A control method for a crop sorting device as described in any one of claims 1 to 7, characterized in that, Includes the following steps: The crops to be sorted are transported to the detection area via the platform conveyor belt. When the infrared sensor of the detection device (12) detects the entry of the crops, it sends the detection signal to the core control unit. The system collects crop shape regularity parameters, spatial posture information, and external geometric contour features using a depth camera; it collects crop sweetness content, moisture content, soluble solids content, and internal defect degree using a near-infrared spectral sensor; it captures crop images from different angles using a multi-dimensional visual acquisition module, and analyzes the images using a data processing module to identify crop maturity level, pest and disease types, and surface defect information. The data preprocessing submodule performs noise reduction, normalization and feature extraction on the collected raw data, and uploads the preprocessed multimodal data to the cloud through the wireless transmission module. It calls the standard parameters in the cloud agricultural product standard feature library, and uses a multimodal data fusion algorithm to comprehensively score the data collected by the depth camera, near-infrared spectral sensor and multi-dimensional vision acquisition module to generate comprehensive detection results for the corresponding crops. Control the gripping device (9) to move to the target position to grab the crop and transfer the crop to the corresponding diversion area; The raw data and comprehensive test results of the crop are uploaded to the cloud via a wireless communication module as an electronic tag for the crop.

9. The control method for a crop sorting device according to claim 8, characterized in that, By controlling the three-axis linear slide to adjust the planar and height positions of the detection device (12), data of the target crop at different angles and positions are collected.

10. The control method for a crop sorting device according to claim 8, characterized in that, The data preprocessing submodule uses wavelet transform denoising algorithm to denoise the original data and Min-Max normalization method to normalize data of different dimensions.

Images