Kiwi fruit growth online monitoring device and method

Abstract

This invention discloses an online monitoring device and method for kiwifruit fruit growth, belonging to the field of kiwifruit fruit growth monitoring technology. The device includes at least one intelligent fruit cluster node and a regional gateway. The intelligent fruit cluster node includes a fruit support assembly, a cluster weighing module, at least two laser ranging modules, a flexible thin-film pressure array, and a data acquisition and control module. The fruit support assembly includes a bowl-shaped shell with a preset inner diameter and depth, and the inner surface of the bowl-shaped shell is covered with a flexible silicone layer. The cluster weighing module is used to measure the real-time weight of the entire cluster of kiwifruit fruits. The laser ranging module is embedded in the side wall of the bowl-shaped shell and can penetrate the fruit bag to measure the fruit size. The flexible thin-film pressure array is used to sense the pressure distribution of each fruit within the cluster. This invention solves the monitoring problems of kiwifruit clustering, bagging, and fuzz interference through a combined design of cluster monitoring, laser penetration measurement, and pressure distribution inversion to determine the weight of individual fruits, achieving long-term stable online monitoring.

Description

Technical Field

[0001] This invention relates to the field of kiwifruit fruit growth monitoring technology, and in particular to an online monitoring device and method for kiwifruit fruit growth. Background Technology

[0002] Kiwifruit is an important economic fruit crop in my country, and its fruit quality is closely related to water and fertilizer management and environmental control during the growth process. In kiwifruit breeding research and high-quality cultivation management, real-time acquisition of growth indicators such as fruit weight, size, and volume is of great significance for optimizing cultivation measures, predicting the harvest period, and evaluating varietal characteristics.

[0003] Existing fruit growth monitoring technologies have the following main shortcomings: Firstly, there is the issue of poor adaptability to clustered fruits. Kiwifruit grows in clusters with multiple fruits per stem. Existing monitoring devices are mostly designed for single fruits, making it difficult to effectively monitor the entire cluster simultaneously. Installing a separate sensor on each fruit would require a large number of devices, increase costs, and negatively impact the normal growth of the fruit.

[0004] Secondly, there are monitoring challenges under bagged cultivation conditions. Bagging is commonly used in kiwifruit cultivation to prevent sunburn and pests. Conventional optical cameras or visual sensors cannot penetrate the fruit bags to obtain images of the fruit, making it impossible to monitor indicators such as size and shape online.

[0005] Third, the fuzz on the fruit interferes with measurement accuracy. The skin of kiwifruit is densely covered with fuzz, which can easily damage the skin with contact sensors, and cause large errors in non-contact optical measurements due to diffuse reflection.

[0006] Fourth, the reliability of long-term outdoor operation is insufficient. Most existing devices are powered by independent batteries, which face the problem of insufficient power supply during the growth cycle of 150-180 days; moreover, the sensors are directly exposed to wind and rain, resulting in poor stability and high maintenance costs.

[0007] Therefore, developing a dedicated online monitoring device that can adapt to the characteristics of kiwifruit such as cluster growth, bagging, and fuzz, and has the ability to operate stably for a long time, has become an urgent technical problem to be solved in this field. Summary of the Invention

[0008] The purpose of this invention is to provide an online monitoring device and method for kiwifruit fruit growth, which solves the problem in the prior art that it is impossible to conduct long-term, stable, and high-precision online monitoring of the entire cluster of bagged kiwifruit. It accurately obtains multiple index data such as weight, size, and shape of the fruit during the growth process, providing data support for subsequent cultivation decisions and breeding research.

[0009] To achieve the above objectives, the present invention provides an online monitoring device for kiwifruit fruit growth, comprising: The intelligent fruit cluster node is fixed on the steel wire of the kiwifruit trellis to monitor the entire cluster of kiwifruit. At least one intelligent fruit cluster node is set up. The regional gateway is electrically connected to the smart fruit cluster nodes and is used for power supply, data aggregation, and communication with the cloud server. The intelligent fruit cluster node includes: Fruit support assembly for supporting a cluster of kiwifruit fruits. The fruit support assembly includes a bowl-shaped shell with a preset inner diameter and depth, and the inner surface of the bowl-shaped shell is covered with a flexible silicone layer. The whole cluster weighing module is connected between the fruit support assembly and the trellis wire to measure the real-time weight of the whole cluster of kiwifruit. A laser ranging module, embedded in the side wall of the bowl-shaped shell, is used to emit a laser beam that penetrates the fruit bag and illuminates the surface of the kiwi fruit to measure the relative distance between the kiwi fruit and the fruit support assembly. At least two laser ranging modules are provided. A flexible thin-film pressure array is laid inside the bottom of a bowl-shaped shell to sense the pressure distribution of each fruit in the entire cluster of kiwifruit. The data acquisition and control module is electrically connected to the cluster weighing module, the laser ranging module, and the flexible thin-film pressure array, respectively, and is used to acquire data and upload it to the regional gateway.

[0010] Preferably, the surface of the flexible silicone layer has a diamond-shaped anti-slip texture.

[0011] Preferably, the cluster weighing module includes an S-shaped tension and compression sensor. One end of the S-shaped tension and compression sensor is connected to an upper hook, which is fixed to the steel wire of the trellis by a quick-locking buckle. The other end of the S-shaped tension and compression sensor is connected to a lower hook, which is connected to the fruit support assembly by a flexible connector. The flexible connector is used to absorb wind-induced vibrations and allow the fruit holder assembly to swing freely. The flexible connector includes a rubber shock-absorbing pad and a stainless steel wire rope. The lower hook is fixedly connected to one end of the rubber shock-absorbing pad, and the other end of the rubber shock-absorbing pad is connected to the lifting lug set on the outer side of the top of the bowl-shaped shell through the stainless steel wire rope. The lifting lug is integrally formed on the outer side of the top edge of the bowl-shaped shell and is used to connect and suspend with the stainless steel wire rope.

[0012] Preferably, the laser ranging modules are evenly distributed along the side wall of the bowl-shaped shell, and the downward tilt angle between the laser axis of each laser ranging module and the horizontal plane is 8°-15°. When the fruit shape index of the kiwi fruit variety is ≤1.2, the downward tilt angle is 12°±2°; when the fruit shape index is >1.2, the downward tilt angle is 14°±2°. The flexible thin-film pressure array is a piezoresistive type, covering a circular area on the bottom side of the bowl-shaped shell. The diameter of the flexible thin-film pressure array is configured to be 75% to 80% of the inner diameter of the bowl-shaped shell.

[0013] Preferably, the data acquisition and control module and the laser ranging module are waterproofed by encapsulation with transparent silicone, and the overall protection level of the machine reaches IP67.

[0014] The data acquisition and control module includes: microcontroller; The laser ranging modules share the I2C bus with the microcontroller and are configured with different I2C addresses in a time-division manner through their respective independent XSHUT pins. The 16-channel row drive output pins connected to the microcontroller are used to sequentially select the flexible thin-film pressure array. The 16 analog-to-digital converter input pins connected to the microcontroller are used to read the 16 columns of analog signals from the flexible thin-film pressure array in parallel. The 16-channel voltage follower has its input terminals connected to the 16 column readout lines of the flexible thin-film pressure array, and its output terminals connected to the 16 analog-to-digital converter input pins.

[0015] Preferably, the area gateway includes: A 12V DC bus is laid along the steel wire of the shed and connected to each intelligent fruit cluster node through a waterproof cable to supply power to each intelligent fruit cluster node. Lithium iron phosphate battery packs and solar charging units are used for solar charging, energy storage, and power supply. The edge computing unit is used to filter, remove anomalies, and calculate growth rate of the data uploaded by the smart fruit cluster nodes. The communication unit is used to upload the data processed by the edge computing unit to the cloud server.

[0016] Preferably, it also includes a wind speed sensor installed on the steel wire of the scaffold. The wind speed sensor is connected to a microcontroller. The data acquisition and control module discards weight data collected when the wind speed exceeds a preset threshold based on the signal from the wind speed sensor.

[0017] This invention also provides a method for online monitoring of kiwifruit growth, comprising the following steps: S1. Installation and initialization: In the early stage of kiwi fruit setting, the intelligent fruit cluster node is fixed to the trellis wire with quick-locking buckles. The whole cluster of fruit is gently placed in the fruit support assembly, and the initial weight and laser ranging baseline value are recorded. S2. Data Acquisition: Set the sampling frequency and collect data intensively during the period of lowest wind speed from 3:00 to 5:00 a.m. every day to obtain the total weight of the cluster, the laser ranging values, and the distribution data of the flexible film pressure array; S3. Data Processing: The weight data is smoothed by Kalman filtering to remove wind-induced noise; Based on the laser ranging values ​​and the installation geometric parameters, the transverse diameter, longitudinal diameter, and fruit shape index of the fruit are calculated. The installation geometric parameters include the installation radius of the laser ranging module, the downward tilt angle of the laser axis, and the initial calibration height. Based on the distribution changes of the flexible thin film pressure array, the weight of the entire cluster is decomposed into the weight of each individual fruit using a machine learning model. S4. Index Output: Calculate the daily growth rate, expansion stagnation warning index, and maturity index, and upload them to the cloud server.

[0018] Preferably, the formulas for calculating the transverse and longitudinal diameters of the fruit in step S3 are as follows: ; ; in, The transverse diameter of the fruit. The longitudinal diameter of the fruit, This refers to the number of laser ranging modules. ≥2; The installation radius of the laser ranging module; For the first The measurement distance of each laser ranging module; The downward tilt angle between the laser axis and the horizontal plane; This is the initial calibration height; The machine learning model for estimating the weight of a single fruit uses a random forest regression algorithm. The input features are several pixel values ​​of the flexible thin film pressure array and the weight of the entire cluster. The output is the weight of each individual fruit.

[0019] Preferably, in step S4, the daily growth rate is calculated as the ratio of the difference in weight between two adjacent effective samples to the time difference, and the calculation formula is as follows: ; In the formula: Daily growth rate; This represents the weight of the entire cluster of fruits measured at the current moment. The weight of the entire cluster of fruits was measured at the previous sampling time; The effective sampling interval; The trigger condition for the fruit enlargement stagnation warning index is: the daily fruit growth rate is less than 0.1 g / h for three consecutive days; The formula for calculating the maturity index in step S4 is as follows: ; in, The initial weight of the entire cluster of fruits recorded during installation initialization; The current transverse diameter of the fruit; The initial transverse diameter of the fruit; The maximum daily growth rate during the fruit's growth cycle; , , These are the weighting coefficients. ,and , , The value range is 0.3 to 0.5.

[0020] The advantages and positive effects of the online monitoring device and method for kiwifruit fruit growth described in this invention are as follows: By combining cluster monitoring, laser penetration measurement, and pressure distribution inversion to determine single fruit weight, this invention effectively solves monitoring challenges related to kiwifruit clustering, bagging, and fuzz interference. The device is powered by a 12V DC busbar connected to the trellis steel wire, eliminating the need for built-in batteries in the intelligent fruit cluster nodes and resolving long-term power supply issues. The cluster weighing module is connected to the fruit support assembly via flexible connectors, effectively isolating wind-induced vibrations. Electronic components are waterproofed, achieving an IP67 protection rating for the entire machine, allowing it to adapt to outdoor growth cycles of 150-180 days. This invention significantly improves the accuracy, reliability, and automation of kiwifruit fruit growth monitoring, making it suitable for breeding research and high-quality cultivation management of multiple kiwifruit varieties, including Hongyang, Hayward, and Jinyan.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the overall structure of an embodiment of the online monitoring device for kiwifruit fruit growth according to the present invention; Figure 2 This is a schematic diagram of the fruit support assembly structure of an embodiment of the online monitoring device for kiwifruit fruit growth according to the present invention; Figure 3 This is a schematic diagram of the internal structure distribution of the bowl-shaped shell in an embodiment of the online monitoring device for kiwifruit fruit growth according to the present invention; Figure 4 This is an internal view of the bowl-shaped shell of an embodiment of the online monitoring device for kiwifruit fruit growth according to the present invention; Figure 5 This is a schematic diagram of a 16×16 pixel flexible thin film pressure array structure, illustrating an embodiment of an online monitoring device for kiwifruit fruit growth according to the present invention. Figure 6 This is a circuit connection block diagram of the data acquisition and control module of an embodiment of the online monitoring device for kiwifruit fruit growth according to the present invention; Figure 7 This is a flowchart illustrating an embodiment of an online monitoring method for kiwifruit fruit growth according to the present invention.

[0023] Figure label: 1. Intelligent fruit cluster node; 2. Regional gateway; 3. Trellise steel wire; 4. Cloud server; 5. Bowl-shaped shell; 6. Quick lock; 7. S-shaped tension and compression sensor; 8. Upper hook; 9. Lower hook; 10. Stainless steel wire rope; 11. Lifting lug; 12. Rubber shock-absorbing pad; 13. Laser ranging module; 14. Flexible film pressure array; 15. Flexible silicone layer. Detailed Implementation

[0024] In the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product is in use. They are used only for the convenience of describing the 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. Therefore, they should not be construed as limitations on the invention. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," and "connect" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] In this application, 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 application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.

[0026] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0027] Example: like Figure 1 As shown, the online monitoring device for kiwifruit fruit growth of the present invention includes at least one intelligent fruit cluster node 1 and one regional gateway 2. The intelligent fruit cluster node 1 is fixed on the steel wire 3 of the kiwifruit trellis and monitors the entire cluster of kiwifruit fruits; the regional gateway 2 is electrically connected to the intelligent fruit cluster node 1 and is used for power supply, data aggregation and communication with the cloud server 4.

[0028] like Figure 2 , Figure 3 , Figure 4As shown, the intelligent fruit cluster node 1 includes a fruit support assembly, a cluster weighing module, at least two laser ranging modules 13, a flexible thin film pressure array 14, and a data acquisition and control module.

[0029] The fruit support assembly is used to support a cluster of kiwifruit fruits. The fruit support assembly includes a bowl-shaped shell 5 with a preset inner diameter and depth, and the inner surface of the bowl-shaped shell 5 is covered with a flexible silicone layer 15. The surface of the flexible silicone layer 15 is provided with a diamond-shaped anti-slip texture, and the Shore hardness of the silicone layer is 30A.

[0030] The inner diameter and depth of the bowl-shaped shell 5 of the fruit holder assembly are configured according to the kiwi fruit variety, including: For the Hongyang variety: inner diameter 85±5mm, depth 22±2mm; For Hayward varieties: inner diameter 105±5mm, depth 28±2mm; For the Jin Yan variety: inner diameter 95±5mm, depth 25±2mm; For the Cui Xiang variety: inner diameter 95±5mm, depth 26±2mm; For the Xu Xiang variety: inner diameter 100±5mm, depth 27±2mm; For the hardy kiwifruit variety: inner diameter 45±5mm, depth 12±2mm.

[0031] The cluster weighing module is connected between the fruit support assembly and the trellis wire 3 to measure the real-time weight of the entire cluster of kiwifruit. The cluster weighing module includes an S-shaped tension / compression sensor 7. One end of the S-shaped tension / compression sensor 7 is connected to an upper hook 8, which is fixed to the trellis wire 3 via a quick-locking buckle 6. The other end of the S-shaped tension / compression sensor 7 is connected to a lower hook 9, which is connected to the fruit support assembly via a flexible connector. The S-shaped tension / compression sensor 7 has a measuring range of 0-2kg and an accuracy of not less than 0.1g. The flexible connector is used to absorb wind-induced vibrations and allow the fruit support assembly to swing freely. The flexible connector includes a rubber shock-absorbing pad 12 and a stainless steel wire rope 10. The lower hook 9 is fixedly connected to one end of the rubber shock-absorbing pad 12, and the other end of the rubber shock-absorbing pad 12 is connected to a lifting lug 11 located on the outer side of the top of the bowl-shaped shell 5 via the stainless steel wire rope 10. The lifting lug 11 is integrally formed on the outer edge of the top of the bowl-shaped shell 5 and is used to connect and suspend the stainless steel wire rope 10.

[0032] A laser ranging module 13 is embedded in the side wall of the bowl-shaped housing 5. It emits a laser beam that penetrates the fruit bag and illuminates the surface of the kiwifruit, measuring the relative distance between the kiwifruit and the fruit support assembly. The laser ranging modules 13 are evenly distributed along the side wall of the bowl-shaped housing 5. The downward tilt angle of the laser axis of each laser ranging module 13 relative to the horizontal plane is 8°-15°, specifically configured according to the fruit shape index of the variety. The downward tilt angle of the laser axis of the laser ranging module 13 relative to the horizontal plane is configured according to the following rules: when the fruit shape index of the kiwifruit variety is ≤1.2, the downward tilt angle is 12°±2°; when the fruit shape index is >1.2, the downward tilt angle is 14°±2°. The laser ranging module 13 is a time-of-flight sensor. Each laser ranging module 13 is controlled by an independent XSHUT pin for reset and I2C address configuration. The I2C addresses of the three modules are configured as 0x52, 0x54, and 0x56, respectively.

[0033] A flexible thin-film pressure array 14 is laid inside the bottom side of the bowl-shaped shell 5 to sense the pressure distribution of each fruit in the entire cluster of kiwifruit. The flexible thin-film pressure array 14 is a piezoresistive type, covering a circular area on the bottom side of the bowl-shaped shell 5, and the diameter of the flexible thin-film pressure array 14 is configured to be 75% to 80% of the inner diameter of the bowl-shaped shell 5.

[0034] The data acquisition and control module and the laser ranging module 13 are waterproofed using transparent silicone potting, achieving an overall protection rating of IP67. Figure 5 , Figure 6 As shown, the data acquisition and control module is electrically connected to the cluster weighing module 13, the laser ranging module 13, and the flexible thin-film pressure array 14, respectively, for acquiring data and uploading it to the area gateway 2. The data acquisition and control module includes: a microcontroller, an I2C bus, 16 row drive output pins, 16 analog-to-digital converter input pins, and 16 voltage followers. The I2C bus is connected to the microcontroller; the laser ranging modules 13 share the I2C bus and are configured with different I2C addresses through their independent XSHUT pins in a time-division multiplexing manner. The 16 row drive output pins are connected to the microcontroller and are used for row-by-row selection of the flexible thin-film pressure array 14. The 16 analog-to-digital converter input pins are connected to the microcontroller and are used for parallel reading of the 16 columns of analog signals from the flexible thin-film pressure array 14. The 16 voltage followers have their inputs connected to the 16 column readout lines of the flexible thin-film pressure array 14, and their outputs connected to the 16 analog-to-digital converter input pins.

[0035] The data acquisition and control module has a built-in real-time clock. The sampling strategy is to collect data once per hour from 3:00 AM to 5:00 AM every day, and to suspend data collection during other times to reduce power consumption.

[0036] The regional gateway 2 includes a 12V DC bus, a lithium iron phosphate battery pack and solar charging unit, an edge computing unit, and a communication unit. The 12V DC bus is laid along the scaffolding steel wire 3 and connected to each intelligent fruit cluster node 1 via waterproof cables, supplying power to each node. The lithium iron phosphate battery pack and solar charging unit are used for solar charging, energy storage, and power supply. The edge computing unit is used to filter, remove anomalies, and calculate growth rates from the data uploaded by the intelligent fruit cluster nodes 1. The communication unit is used to upload the data processed by the edge computing unit to the cloud server 4.

[0037] The device of the present invention also includes a wind speed sensor installed on the steel wire 3 of the scaffold. The wind speed sensor is connected to a microcontroller. The data acquisition and control module rejects the weight data collected when the wind speed exceeds a preset threshold based on the signal from the wind speed sensor.

[0038] like Figure 7 As shown, the online monitoring method for kiwifruit growth according to the present invention includes the following steps: S1. Installation and initialization: In the early stage of kiwi fruit setting, the intelligent fruit cluster node 1 is fixed to the trellis wire 3 by quick-locking buckle 6, the whole cluster of fruit is gently placed in the fruit support assembly, and the initial weight and laser ranging baseline value are recorded. S2. Data Acquisition: Set the sampling frequency and collect data intensively during the period of lowest wind speed from 3:00 to 5:00 a.m. every day to obtain the total weight of the cluster, the laser ranging values, and the distribution data of the flexible thin film pressure array 14. S3. Data Processing: The weight data is smoothed by Kalman filtering to remove wind-induced noise; Based on the laser ranging values ​​and the installation geometric parameters, the transverse diameter, longitudinal diameter, and fruit shape index of the fruit are calculated. The installation geometric parameters include the installation radius of the laser ranging module 13, the downward tilt angle of the laser axis, and the initial calibration height.

[0039] The formulas for calculating the transverse and longitudinal diameters of the fruit are: ; ; in, The transverse diameter of the fruit. The longitudinal diameter of the fruit, The number of laser ranging modules is 13. ≥2; The installation radius of the laser ranging module 13; For the first The measurement distance of the laser ranging module 13; The downward tilt angle between the laser axis and the horizontal plane; This is the initial calibration height.

[0040] Based on the distribution variation of the flexible film pressure array 14, the weight of the entire cluster is decomposed into the estimated weight of each individual fruit using a machine learning model. The machine learning model for estimating the weight of each individual fruit uses a random forest regression algorithm. The input features are the 256 pixel values ​​of the pressure array and the weight of the entire cluster, and the output is the weight of each individual fruit. The training dataset of the machine learning model contains no less than 200 pairs of harvested measured data and corresponding growth cycle pressure feature vectors.

[0041] S4. Index Output: Calculate the daily growth rate, expansion stagnation warning index, and maturity index, and upload them to the cloud server 4.

[0042] The daily growth rate is calculated as the ratio of the difference in weight between two adjacent effective samples to the time difference, using the following formula: ; In the formula: Daily growth rate; This represents the weight of the entire cluster of fruits measured at the current moment. The weight of the entire cluster of fruits was measured at the previous sampling time; The effective sampling interval is the time between two sampling operations.

[0043] The trigger condition for the fruit growth stagnation warning index is: the daily fruit growth rate is below 0.1 g / h for 3 consecutive days.

[0044] The formula for calculating the maturity index is: ; in, The initial weight of the entire cluster of fruits recorded during installation initialization; The current transverse diameter of the fruit; The initial transverse diameter of the fruit; This represents the current daily growth rate of the fruit. The maximum daily growth rate during the fruit's growth cycle; , , These are the weighting coefficients. ,and , , The value range is 0.3 to 0.5.

[0045] This example demonstrates online monitoring of the Hongyang kiwifruit variety: The intelligent fruit cluster node 1 includes a fruit support assembly, a cluster weighing module, three laser ranging modules 13, a flexible thin film pressure array 14, and a data acquisition and control module.

[0046] The bowl-shaped shell 5 of the fruit holder assembly is injection molded from nylon material, with an inner diameter of 85mm and a depth of 22mm. The inner surface is covered with a 3mm thick food-grade flexible silicone layer 15. The surface of the flexible silicone layer 15 has diamond-shaped anti-slip patterns with a spacing of 2mm and a depth of 0.2mm to reduce the contact area with the fruit's fuzz and prevent slippage.

[0047] The entire cluster weighing module uses an S-type tension / compression sensor 7 with a range of 2kg and an accuracy of 0.1g. The upper hook 8 of the S-type tension / compression sensor 7 is fixed to the frame steel wire 3 by a quick-lock buckle 6, and the lower hook 9 is connected to the lifting lug 11 of the fruit support assembly by a flexible connector. The flexible connector includes a 5mm thick rubber shock-absorbing pad 12 and a 0.8mm diameter, 50mm long stainless steel wire rope 10, which can effectively filter out high-frequency vibrations and allow the fruit support assembly to sway with the wind.

[0048] All three laser ranging modules 13 employ VL53L1X time-of-flight sensors with an emission wavelength of 940nm. They are evenly distributed along the sidewall of the fruit tray assembly, with adjacent modules having an included angle of 120°. The downward tilt angle between the axis of each laser ranging module 13 and the horizontal plane is 12°. O-rings are installed in the mounting holes of the laser ranging modules 13, and the entire module is encapsulated with transparent silicone to meet waterproof and dustproof requirements.

[0049] The flexible thin-film pressure array 14 is a 16×16 pixel piezoresistive sensor with an effective diameter of 65mm. It is laid in the bottom groove of the fruit holder assembly. The back of the flexible thin-film pressure array 14 is attached to the bottom of the groove, and its surface is flush with the flexible silicone layer 15.

[0050] The data acquisition and control module is based on an STM32F103C8T6 microcontroller. Three laser ranging modules 13 share an I2C1 bus (SCL:PB6, SDA:PB7), with 4.7kΩ pull-up resistors. The XSHUT pin of the first laser ranging module 13 is connected to PA0, the XSHUT pin of the second laser ranging module 13 is connected to PA1, and the XSHUT pin of the third laser ranging module 13 is connected to PA2. After power-on, by controlling the XSHUT pins, the address of the second laser ranging module 13 is configured to 0x54, and the address of the third laser ranging module 13 is configured to 0x56.

[0051] The 16 row drive lines of the flexible thin-film pressure array 14 are connected to the PB0-PB15 pins of the microcontroller, outputting high-level strobe signals row by row. The 16 column readout lines are connected to 16 LMV321 voltage followers, and the follower outputs are connected to the ADC input pins (PA0-PA7, PB0-PB1, PC0-PC5) of the microcontroller, using a 16-channel simultaneous sampling mode with a sampling frequency of 100Hz and a resolution of 12 bits.

[0052] The regional gateway 2 includes a 12V DC bus, a lithium iron phosphate battery pack and solar charging unit, an edge computing unit, and a communication unit. The edge computing unit filters, removes anomalies, and calculates growth rates for the data uploaded by the nodes, while the communication unit uploads the processed data to the cloud server 4.

[0053] The monitoring method in this embodiment includes the following steps: S1. Installation and Initialization: 15 days after flowering (early fruit set stage), when the fruit diameter is approximately 15mm, gently place a cluster of 3-5 fruits into the fruit support assembly. Secure the entire intelligent fruit cluster node 1 to the trellis wire 3 using the quick-locking buckle 6, and record the initial cluster weight. and the initial distance of each laser ranging module 13 , , As a baseline.

[0054] S2. Data Acquisition: The sampling frequency is set to once daily at 3:30 AM (to avoid daytime wind disturbance). During each acquisition, the total cluster weight is recorded. Read three laser ranging values , , The pressure array was scanned sequentially across 16 rows to obtain 256 pressure values. .

[0055] S3. Data Processing: Weight processing: Kalman filtering was used to smooth the weight data for three consecutive days. The filtered weight was denoted as... When the wind speed sensor reading exceeds 3 m / s, the corresponding weight data is marked as invalid and is not included in the filtering calculation.

[0056] Size calculation: Based on three laser ranging values, combined with the known geometric parameters of the laser module installation location (installation radius)... =42.5mm, downslope angle =12°), calculate the transverse diameter of the fruit. and longitudinal diameter : Single fruit weight decomposition: Constructing a feature vector from 256 pressure values. Input a pre-trained random forest regression model and output the estimated weight of each fruit. , … ,in Indicates the first The model estimates the weight of a single kiwifruit. Using the actual weight of a single fruit after harvest as the label and the stress characteristics throughout the entire growth cycle as the training samples, the model achieves an estimation error of less than ±5% after training with 200 sets of data.

[0057] S4. Indicator Output: Daily growth rate ; In the formula: Daily growth rate; This represents the weight of the entire cluster of fruits measured at the current moment. The weight of the entire cluster of fruits was measured at the previous sampling time; The effective sampling interval; Expansion stagnation warning: When for 3 consecutive days A warning message will be sent when the concentration is <0.1g / h; Maturity Index: ; in, ; The initial weight of the entire cluster of fruits recorded during installation initialization; The current transverse diameter of the fruit; The initial transverse diameter of the fruit; The maximum daily growth rate during the fruit's growth cycle; when A value >0.85 indicates the optimal harvesting window has arrived.

[0058] Regional gateway 2 aggregates the processed data from each intelligent fruit cluster node 1 every 6 hours and uploads it to cloud server 4 via 4G communication unit. Cloud server 4 provides a web interface to display the weight growth curve, size change curve, estimated weight of a single fruit, and abnormal warning information for each fruit cluster.

[0059] Example 2, Adaptation of Hayward kiwifruit variety: Unlike Example 1, this example is size-adapted for the Hayward variety. The bowl-shaped shell 5 of the fruit holder assembly has an inner diameter of 105 mm and a depth of 28 mm, the flexible film pressure array 14 has a diameter of 85 mm, and the laser ranging module 13 has a downward tilt angle of 15°. The remaining structure and connection methods are the same as in Example 1.

[0060] Tests showed that, within the 150-180 day growth cycle of Hayward kiwifruit, the device in this embodiment achieved a weight measurement accuracy of ±0.6g and a size measurement accuracy of ±1.8mm. The correlation coefficient between the estimated and measured single-fruit weight was also [not specified]. =0.91, the device has a fault-free operation rate of 97.5%, and the data validity rate is 92.8%.

[0061] Example 3, Adaptation of the Cui Xiang kiwifruit variety: Unlike Example 1, this example is size-adapted for the Cui Xiang variety. The bowl-shaped shell 5 of the fruit support assembly has an inner diameter of 95 mm and a depth of 26 mm, the flexible film pressure array 14 has a diameter of 75 mm, and the laser ranging module 13 has a downward tilt angle of 14°. The inner wall of the flexible silicone layer 15 is reinforced with longitudinal stripes, 0.3 mm deep and spaced 5 mm apart, to guide the slender fruit to be placed in the center.

[0062] Tests showed that, within a 150-160 day growth cycle of the Cui Xiang kiwifruit, the device in this embodiment achieved a weight measurement accuracy of ±0.4g and a size measurement accuracy of ±1.3mm. The correlation coefficient between the estimated and measured single-fruit weight was also [not specified]. =0.94, the device has a fault-free operation rate of 99.0%, and the data validity rate is 95.1%.

[0063] Experimental Verification: In three different major kiwifruit producing areas, five main cultivated varieties—Hongyang, Hayward, Jinyan, Cuixiang, and Xuxiang—were selected, and online monitoring experiments were conducted over three consecutive growth cycles. Twenty intelligent fruit cluster nodes were set up for each variety, with traditional manual measurement (using calipers and electronic scales every 7 days) serving as the control.

[0064] Experimental data show that the device of this invention exhibits high measurement accuracy and stability across different varieties, with a weight measurement error ≤0.6g, a size measurement error ≤1.8mm, and a high correlation between the estimated single fruit weight and the measured value. With a failure rate exceeding 97% and a data validity rate exceeding 92%, it fully meets the accuracy and reliability requirements for kiwifruit growth monitoring. Compared with traditional manual measurement, the data acquisition efficiency is improved by more than 50 times, and continuous monitoring throughout the entire growth cycle is achieved.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An online monitoring device for kiwifruit fruit growth, characterized in that, include: The intelligent fruit cluster node is fixed on the steel wire of the kiwifruit trellis to monitor the entire cluster of kiwifruit. At least one intelligent fruit cluster node is set up. The regional gateway is electrically connected to the smart fruit cluster nodes and is used for power supply, data aggregation, and communication with the cloud server. The intelligent fruit cluster node includes: Fruit support assembly for supporting a cluster of kiwifruit fruits. The fruit support assembly includes a bowl-shaped shell with a preset inner diameter and depth, and the inner surface of the bowl-shaped shell is covered with a flexible silicone layer. The whole cluster weighing module is connected between the fruit support assembly and the trellis wire to measure the real-time weight of the whole cluster of kiwifruit. A laser ranging module, embedded in the side wall of the bowl-shaped shell, is used to emit a laser beam that penetrates the fruit bag and illuminates the surface of the kiwi fruit to measure the relative distance between the kiwi fruit and the fruit support assembly. At least two laser ranging modules are provided. A flexible thin-film pressure array is laid inside the bottom of a bowl-shaped shell to sense the pressure distribution of each fruit in the entire cluster of kiwifruit. The data acquisition and control module is electrically connected to the cluster weighing module, the laser ranging module, and the flexible thin-film pressure array, respectively, and is used to acquire data and upload it to the regional gateway.

2. The online monitoring device for kiwifruit fruit growth according to claim 1, characterized in that, The flexible silicone layer has a diamond-shaped anti-slip texture on its surface.

3. The online monitoring device for kiwifruit fruit growth according to claim 1, characterized in that, The entire cluster weighing module includes an S-shaped tension and compression sensor. One end of the S-shaped tension and compression sensor is connected to an upper hook, which is fixed to the steel wire of the trellis by a quick-locking buckle. The other end of the S-shaped tension and compression sensor is connected to a lower hook, which is connected to the fruit support assembly by a flexible connector. The flexible connector is used to absorb wind-induced vibrations and allow the fruit rack assembly to swing freely. The flexible connector includes a rubber shock-absorbing pad and a stainless steel wire rope. The lower hook is fixedly connected to one end of the rubber shock-absorbing pad, and the other end of the rubber shock-absorbing pad is connected to the lifting lug set on the outer side of the top of the bowl-shaped shell through the stainless steel wire rope. The lifting lug is integrally formed on the outer side of the top edge of the bowl-shaped shell and is used to connect and suspend with the stainless steel wire rope.

4. The online monitoring device for kiwifruit fruit growth according to claim 1, characterized in that, The laser ranging modules are evenly distributed along the side wall of the bowl-shaped shell. The downward tilt angle between the laser axis of each laser ranging module and the horizontal plane is 8°-15°. When the fruit shape index of the kiwi fruit variety is ≤1.2, the downward tilt angle is 12°±2°; when the fruit shape index is >1.2, the downward tilt angle is 14°±2°. The flexible thin-film pressure array is a piezoresistive type, covering a circular area on the bottom side of the bowl-shaped shell. The diameter of the flexible thin-film pressure array is configured to be 75% to 80% of the inner diameter of the bowl-shaped shell.

5. The online monitoring device for kiwifruit fruit growth according to claim 1, characterized in that, The data acquisition and control module and the laser ranging module are waterproofed using transparent silicone potting. The data acquisition and control module includes: microcontroller; The laser ranging modules share the I2C bus with the microcontroller and are configured with different I2C addresses in a time-division manner through their respective independent XSHUT pins. The 16-channel row drive output pins connected to the microcontroller are used to sequentially select the flexible thin-film pressure array. The 16 analog-to-digital converter input pins connected to the microcontroller are used to read the 16 columns of analog signals from the flexible thin-film pressure array in parallel. The 16-channel voltage follower has its input terminals connected to the 16 column readout lines of the flexible thin-film pressure array, and its output terminals connected to the 16 analog-to-digital converter input pins.

6. The online monitoring device for kiwifruit fruit growth according to claim 5, characterized in that, The area gateway includes: A 12V DC bus is laid along the steel wire of the shed and connected to each intelligent fruit cluster node through a waterproof cable to supply power to each intelligent fruit cluster node. Lithium iron phosphate battery packs and solar charging units are used for solar charging, energy storage, and power supply. The edge computing unit is used to filter, remove anomalies, and calculate growth rate of the data uploaded by the smart fruit cluster nodes. The communication unit is used to upload the data processed by the edge computing unit to the cloud server.

7. The online monitoring device for kiwifruit fruit growth according to claim 5, characterized in that, It also includes a wind speed sensor installed on the steel wire of the scaffold. The wind speed sensor is connected to a microcontroller. The data acquisition and control module uses the signal from the wind speed sensor to discard weight data collected when the wind speed exceeds a preset threshold.

8. A method for online monitoring of kiwifruit growth according to any one of claims 1-7, characterized in that, Includes the following steps: S1. Installation and initialization: In the early stage of kiwi fruit setting, the intelligent fruit cluster node is fixed to the trellis wire with quick-locking buckles. The whole cluster of fruit is gently placed in the fruit support assembly, and the initial weight and laser ranging baseline value are recorded. S2. Data Acquisition: Set the sampling frequency and collect data intensively during the period of lowest wind speed from 3:00 to 5:00 a.m. every day to obtain the total weight of the cluster, the laser ranging values, and the distribution data of the flexible film pressure array; S3. Data Processing: The weight data is smoothed by Kalman filtering to remove wind-induced noise; Based on the laser ranging values ​​and the installation geometric parameters, the transverse diameter, longitudinal diameter, and fruit shape index of the fruit are calculated. The installation geometric parameters include the installation radius of the laser ranging module, the downward tilt angle of the laser axis, and the initial calibration height. Based on the distribution changes of the flexible thin film pressure array, the weight of the entire cluster is decomposed into the weight of each individual fruit using a machine learning model. S4. Index Output: Calculate the daily growth rate, expansion stagnation warning index, and maturity index, and upload them to the cloud server.

9. The monitoring method of the online monitoring device for kiwifruit fruit growth according to claim 8, characterized in that, The formulas for calculating the transverse and longitudinal diameters of the fruit in step S3 are as follows: ； ； in, The transverse diameter of the fruit. The longitudinal diameter of the fruit, This refers to the number of laser ranging modules. ≥2; The installation radius of the laser ranging module; For the first The measurement distance of each laser ranging module; The downward tilt angle between the laser axis and the horizontal plane; This is the initial calibration height; The machine learning model for estimating the weight of a single fruit uses a random forest regression algorithm. The input features are several pixel values ​​of the flexible thin film pressure array and the weight of the entire cluster. The output is the weight of each individual fruit.

10. The monitoring method of the online monitoring device for kiwifruit fruit growth according to claim 8, characterized in that, In step S4, the daily growth rate is calculated as the ratio of the difference in weight between two adjacent effective samplings to the time difference, using the following formula: ； In the formula: Daily growth rate; This represents the weight of the entire cluster of fruits measured at the current moment. The weight of the entire cluster of fruits was measured at the previous sampling time; The effective sampling interval; The trigger condition for the fruit enlargement stagnation warning index is: the daily fruit growth rate is less than 0.1 g / h for three consecutive days; The formula for calculating the maturity index in step S4 is as follows: ； in, The initial weight of the entire cluster of fruits recorded during installation initialization; The current transverse diameter of the fruit; The initial transverse diameter of the fruit; The maximum daily growth rate during the fruit's growth cycle; , , These are the weighting coefficients. ,and , , The value range is 0.3 to 0.5.

Images