Portable fruit appearance phenotypic character three-dimensional reconstruction modeling instrument for fruit and vegetable crops

Abstract

The invention discloses a portable three-dimensional reconstruction modeling instrument for fruit appearance phenotypic characters of fruit and vegetable crops. Comprising a fruit appearance phenotypic character shooting sensor support, a fruit overhanging type fixing device, a multi-head sensor module, a model geometry and color calibration device, an instrument shooting data storage module and a fruit appearance phenotypic character three-dimensional reconstruction modeling instrument control module. According to the invention, by designing a fruit omnibearing video shooting platform and a multi-head sensor group, a portable fruit and vegetable crop fruit appearance phenotypic character three-dimensional reconstruction modeling instrument is constructed; a modeling target is fixed in a suspension mode, and shooting parameters of a sensor group and the rotating speed of a sensor support can be controlled, so that fruit appearance omnibearing images are quickly and efficiently acquired, image transmission, modeling, storage and data output are realized, and the method is used for subsequent fruit appearance three-dimensional true color reconstruction modeling. And model geometry and color feature measurement is supported.

Description

Technical Field

[0001] This invention relates to the field of intelligent analysis equipment for the appearance and phenotypic traits of crop fruits, and in particular to a portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops. Background Technology

[0002] With the development of intelligent breeding technology, crop breeding work requires existing breeding programs to not only identify genetic markers among germplasm resources but also to model, characterize, and evaluate their phenotypic traits. Fruit appearance phenotypic traits are crucial in the evaluation of crop germplasm resources, involving the analysis of traits such as fruit size, shape, peel color, and texture. Currently, the evaluation of appearance phenotypic traits of germplasm resources mainly involves measuring the overall length and diameter of the fruit, combined with qualitative analysis of fruit color traits. However, due to the complex appearance structure of some crop fruits, uneven color, and intricate textures, detailed structural traits are difficult to measure directly. There is a lack of efficient and accurate quantitative observation techniques and equipment for fruit surface detail structural features and color traits such as color uniformity. This leads to a need to expand research on the genetic laws of related traits and the molecular mechanisms in the breeding process, hindering the development of intelligent breeding technology. There is an urgent need to develop relevant intelligent breeding equipment to achieve efficient quantitative modeling and phenotypic trait extraction of germplasm resources.

[0003] 3D modeling of fruit appearance enables quantitative analysis of detailed phenotypic traits. While existing methods based on LiDAR, depth cameras, and structured light scanning can achieve millimeter-level precision in 3D modeling of fruit appearance structure, they struggle to accurately depict the color and texture of the fruit. Furthermore, most high-performance devices are not portable, and post-processing of scanned data is complex, making it difficult to efficiently and cost-effectively support the analysis of 3D phenotypic traits of crop fruit appearance.

[0004] With the development of 3D reconstruction technology, 3D reconstruction modeling of fruit appearance traits can be achieved based on omnidirectional video, accurately depicting the fruit's color and texture characteristics, providing technical support for the analysis and predictive modeling of fruit appearance phenotypic traits in the breeding process. To quickly and efficiently obtain a true-color 3D model of fruit appearance, a portable 3D reconstruction modeling instrument for fruit appearance phenotypic traits needs to be designed based on 3D reconstruction technology and feasible data acquisition methods.

[0005] In the early stages of 3D reconstruction technology development, the need to rely on target object point clouds and depth maps to improve modeling quality limited its universality. Subsequently, modeling based on neural network architectures and using multi-view images for training to generate scene models became a research hotspot. These technologies utilize camera pose transformation and image completion techniques to iteratively synthesize continuous viewpoint images or virtual scenes, possessing high-fidelity reconstruction capabilities for target object structures and surface colors. Representative technologies include NeRF and 3DGS. NeRF uses multi-view images and shooting pose information to self-supervised train a multilayer perceptron, outputting a neural radiation field model that implicitly expresses the appearance, structure, and color information of objects using neural networks. However, the results are difficult to edit and perform structural measurements. 3DGS technology uses a large number of anisotropic Gaussian ellipsoidal voxels to explicitly represent the scene. During scene rendering, the Gaussian ellipsoidal voxels are optimized to generate high-precision scene images through differentiable splash rendering according to the set rendering mode, significantly improving the rendering quality and efficiency of scene models. It supports model editing and spatial reference settings and is easily converted into point clouds and mesh models for output. This technology is now used in 3D simulation reconstruction and phenotypic trait extraction of crops such as wheat and corn.

[0006] Therefore, during the fruit modeling data acquisition process, the target fruit's pose needs to be fixed, and local structural occlusion must be avoided. During modeling, the fruit must remain stationary, and the imaging quality from different sensors must be high-definition and stable to prevent structural defects and blurred details in subsequent modeling processes. The instrument also needs to be adaptable to various indoor and outdoor environments, suppressing the influence of external light on the model's color fidelity. The instrument should be lightweight and easy to operate to lower the barrier to entry for users with different professional backgrounds. The modeling results also require reliable spatial dimension markers and standard color charts to assist in the calibration of structural and color features, supporting the extraction of spatial and fruit color correlation traits.

[0007] Therefore, providing a method for unobstructed modeling of fruit appearance characteristics by fixing the target pose through fruit stem suspension, and continuously capturing omnidirectional images of the target object by multiple high-definition cameras around it at a set time step, is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0008] In view of this, the present invention provides a portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops. By acquiring omnidirectional images of the fruit appearance and combining them with image capture pose analysis, a three-dimensional model of the target fruit appearance is realized, providing reliable data for subsequent intelligent analysis of detailed phenotypic traits.

[0009] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: A portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops includes: a fruit appearance and phenotypic trait imaging sensor bracket, a fruit suspension fixing device, a multi-head sensor module, a model geometry and color calibration device, an instrument imaging data storage module, and a three-dimensional reconstruction modeling instrument control module for the appearance and phenotypic traits of fruit. The fruit appearance phenotypic morphology imaging sensor bracket is used to suspend and fix the fruit to be modeled. The above operations can keep the fruit in the same position during the modeling data collection process; The side of the fruit appearance phenotypic imaging sensor bracket is a multi-head sensor module of a 3D reconstruction modeler, which is used to collect omnidirectional images of the fruit and perform 3D reconstruction. The fruit suspension fixing device secures the fruit stem of the fruit or the stem with a tie by means of a buckle. The above operations allow the fruit to be suspended and fixed on the instrument, preventing directional occlusion during the omnidirectional image acquisition of the fruit and ensuring that the fruit maintains a fixed pose during the omnidirectional image data acquisition process; and avoiding damage to the appearance of the fruit during the modeling process. The multi-head sensor module consists of five high-definition cameras that capture images of the fruit's appearance from different angles. Through the above operations, during the omnidirectional rotation of the sensor bracket around the fruit, omnidirectional images of the fruit can be acquired from different directions at the same time step, which can be used for three-dimensional reconstruction modeling of the fruit's appearance. Before data acquisition, the parameters of a single sensor can be adjusted through the control module to ensure image capture quality. The capture data from each sensor are collected into a separate folder for users to edit. The model geometry and color calibration device consists of a geometry calibration base and a color calibration caliper. Through the above operations, during the process of acquiring images of the fruit from all directions, the images will record the geometric and color calibration devices; after the three-dimensional reconstruction modeling, the geometric calibration chassis supports the manual picking of spatial reference marks in the scene model to give spatial reference information to the three-dimensional model of the fruit appearance, and supports the structural shape analysis and measurement of the fruit appearance model; through the color calibration caliper, the color characterization fidelity of the model can be evaluated, and the color feature optimization of the model appearance can be achieved. The instrument's image data storage module is used to store images acquired by multiple sensors respectively; The control module of the fruit appearance phenotypic 3D reconstruction modeler allows users to turn the device on and off, set the time step interval during the multi-sensor module's sequential shooting process, set the shooting parameters of a single sensor, adjust the rotation speed of the fruit appearance phenotypic trait shooting sensor bracket, control the auxiliary lighting during the shooting process, and allows users to browse, query, edit, and copy the stored data.

[0010] Preferably, the fruit appearance phenotypic trait imaging sensor bracket fixes the fruit stem part of the fruit by suspension, and acquires and records the fruit appearance phenotypic trait image set of fruit and vegetable crop based on multi-head sensor all-round shooting, supports three-dimensional reconstruction modeling of target fruit, and realizes non-destructive high-fidelity recording of fruit appearance geometry and color traits.

[0011] Preferably, the fruit appearance phenotypic morphology imaging sensor bracket is an approximately arc-shaped structure with a fruit suspension fixing device in the middle; The fruit appearance and phenotypic characteristics imaging sensor bracket has three layers, with the inner and outer layers fixed and the middle layer installing a multi-head sensor module. Through the above operations, images of the modeled target can be captured from multiple angles; The middle layer can rotate clockwise around the fruit-hanging fixing device; Through the above operations, an all-around image of the modeling target can be recorded for three-dimensional reconstruction modeling of fruit appearance phenotypic traits; The outermost part of the fruit appearance phenotypic trait imaging sensor bracket is the control module of the fruit appearance phenotypic trait three-dimensional reconstruction modeling instrument, which is equipped with a control screen and a simple operation keyboard.

[0012] Preferably, the multi-head sensor module is fixed on the middle support at the upper oblique end of the top of the fruit when it is hanging, the upper, middle and lower parts of the fruit body, and the lower side of the tail of the fruit. During the rotation, it can realize the all-round image acquisition of the appearance of the fruit, especially the images of the top and tail of the fruit. The above operations can improve the integrity of 3D modeling.

[0013] Before the modeling data is captured, the shooting parameters of each sensor can be adjusted separately. After the instrument is started, as the sensor support slowly rotates, the sensor captures color or single-band images of the fruit in the corresponding pose by setting the focal length and frequency and taking equal time steps. The images are numbered and saved in the order of shooting time and used for model iterative training in the process of shooting pose analysis and 3D reconstruction.

[0014] Preferably, the fixing bracket of the multi-head sensor module is embedded in the middle layer of the fruit appearance phenotypic imaging sensor bracket; By performing the above operations, the sensor can be protected during the power-off period; After the instrument is moved, the fixing bracket of the multi-head sensor module will rotate clockwise at a uniform speed around the fruit suspension fixing device and the modeling target object according to the user-set speed and shooting time step. If the user pauses the instrument during the modeling data acquisition process, the sensor bracket will remain in the corresponding position until the user restarts the instrument and continues to complete the multi-view image capture of the remaining positions clockwise. If the user stops capturing the image midway, the sensor bracket will move clockwise back to the initial position to wait for the next startup, that is, it will be embedded in the instrument bracket, and all sensors will be in a powered-off state, waiting for the next startup.

[0015] Preferably, the instrument is provided with the geometric and color calibration device at the bottom. The geometric calibration device is a disc with four scales embedded around the center of the disc to form an inscribed square. By performing the above operations, it can be ensured that at least one ruler image can be seen on each of the four sides of the 3D reconstruction result of the scene.

[0016] A red LED is installed at each node of the scale. By performing the above operations, the start and end points of the scale can be quickly found in the 3D reconstruction modeling results of the scene, reducing the geometric calibration error of the modeling results.

[0017] The scale on the side of the caliper facing the sensor bracket is set to a different color than the other three sides to aid in spatial matching of multiple 3D modeling results of the same object.

[0018] Preferably, a white LED light strip with adjustable brightness is installed around the perimeter of the disc; By performing the above operations, supplemental lighting can be provided to the lower part of the shooting environment when necessary, so as to avoid local color distortion of the fruit peel in the 3D reconstruction modeling results.

[0019] Before operating the instrument, insert the bar colorimeter into the slot; The above operations can be used to evaluate the surface color difference of the 3D reconstructed model.

[0020] After the instrument is shut down, remove or replace the calipers; By performing the above operations, you can avoid the color of the color calipers from becoming contaminated or fading.

[0021] Preferably, during the slow rotation of the multi-head sensor bracket, by setting the focal length and shooting frequency of each camera, images of the corresponding pose are captured around the fruit at equal time steps. Images captured by a single camera are numbered and saved in chronological order of capture time. A small display and shooting parameter control device are embedded on the side of the instrument, allowing users to view the image quality of each individual sensor in real time before shooting, and manually adjust shooting parameters based on the imaging results, including: exposure time, ISO sensitivity, aperture, white balance, focal length, and focus mode sensor parameters.

[0022] By performing the above operations, you can obtain the best image quality.

[0023] Preferably, the control module of the three-dimensional reconstruction modeling instrument for fruit appearance phenotypic traits includes: support rotation control, sensor performance control, and shooting light environment control; The single sensor uses the control screen on the side of the bracket to observe the shooting parameter settings and uses the keyboard next to the control screen to adjust the parameters; the control module controls the rotation speed of the multi-sensor bracket. Through the above operations, users can manually adjust in real time to ensure efficient data collection and that the fruit's hanging posture is not affected by the rotation of the sensor bracket during the collection process.

[0024] If the user stops or finishes shooting, the corresponding dataset will be displayed on the controller's visual interface, and a dialog box will appear asking whether to save; during the shooting process, the shooting parameters of each sensor remain unchanged; To suppress the impact of environmental shadows on modeling quality, the brightness of the ring light strip on the bottom disc can be adjusted using the control keyboard to ensure that the appearance, color, and texture information of the fruit can be accurately depicted.

[0025] During the shooting process, the sensor will be activated to analyze the shaking of the target object. If shaking occurs, the user will be reminded to pause the instrument. After the user confirms, the controller will adjust the sensor bracket to rotate counterclockwise and move it to a position similar to the position before the image shaking occurred. The instrument will also automatically delete the images taken during the warning period.

[0026] Preferably, the control module of the fruit appearance phenotypic 3D reconstruction modeling instrument connects the portable fruit and vegetable crop fruit appearance phenotypic 3D reconstruction modeling instrument to a mobile storage device or computer, transmits the scanned data to the memory via a data cable, and the user transfers the data and performs modeling analysis on a high-performance server to realize the fruit appearance 3D reconstruction modeling. Through the above operations, users can easily view saved folders on the controller screen and perform data deletion, renaming, and renaming operations.

[0027] If some sensors malfunction or fail to store data, the control module will display a warning message on the visual screen before starting the shooting, allowing the user to decide whether to shoot and store the data.

[0028] The present invention achieves the following technical effects compared to the prior art: The instrument of this invention is portable and easy to operate, and is suitable for three-dimensional reconstruction modeling of the appearance and phenotypic characteristics of small and medium-sized elliptical and elongated fruits and vegetables. Attached Figure Description

[0029] Figure 1The main view of the portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops of this invention is shown below. Figure 2 This is a schematic diagram of the operation process of the portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops of the present invention; Figure 3 This is a design drawing of the instrument model geometry and color calibration device of the present invention; Figure 4 This is a structural design diagram of the external control module of the instrument of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] This invention discloses a portable three-dimensional reconstruction modeling instrument for the appearance phenotypic traits of fruit and vegetable crops, comprising: a fruit appearance phenotypic trait imaging sensor bracket, a fruit suspension fixing device, a multi-head sensor module, a model geometry and color calibration device, an instrument imaging data storage module, and a three-dimensional reconstruction modeling instrument control module; wherein... The fruit appearance phenotypic imaging sensor holder features a fruit stalk fixing buckle at the top of the instrument. This buckle can be used to directly secure slender fruit stalks, or it can be used to tie fruit stalks of different diameters together and then secure the ties with the buckle, keeping the fruit suspended in the center of the multi-sensor's field of view. This suspension method maintains the fruit's pose during data acquisition, ensuring no obstruction of the fruit body, top, or bottom. The sensor module can then continuously capture omnidirectional images depicting the fruit's phenotypic characteristics without damaging the fruit's appearance. During instrument operation, the buckle prevents the fruit from rotating with the sensor holder, facilitating subsequent image capture and pose analysis. After capturing the target fruit sample image data, the user can manually open the buckle, remove the fruit, and replace the modeling object.

[0032] The fruit-suspending fixing device includes a multi-sensor fixing mechanism, a multi-sensor shooting bracket, and auxiliary sensors for calibrating shooting parameters. The bracket, located within the instrument support frame, has five sensor interfaces from top to bottom, each used to capture fruit images from different angles. The fruit fixing device remains stationary during bracket rotation. Below the bracket is a fixed disk with multiple size markings, used for dimensional calibration of the subsequent 3D reconstruction results, enabling spatial geometric analysis of the model. Two removable colorimetric plates are mounted on the side of the disk for image color correction. Adjustable white LED light strips can be installed around the disk to provide supplementary lighting for the shooting environment when necessary, preventing distortion of the fruit's external color characteristics in the 3D reconstruction model. During bracket operation, the sensors embedded in the bracket slowly rotate out of the main bracket's interlayer, rotating clockwise around the fruit at a set speed. After shooting, the bracket retracts into the interlayer to protect the multi-sensor fixed to it. The rotation speed of the support can be set to fast, medium, and low through the instrument's main control system. During rapid shooting, it is also necessary to suppress artifacts in the acquired images and prevent the fruit from shaking.

[0033] The multi-sensor module houses five image sensors mounted in the middle layer of the 3D reconstruction modeling instrument's support frame for fruit appearance phenotypic characteristics. These sensors can be true-color cameras, customized single- or multi-band combinations, or video cameras. Because the instrument support frame is nearly semi-circular, the sensors are positioned at the top of the fruit (dangling position), the upper, middle, and lower parts of the fruit body (suspended and fixed), and the lower side of the tail section. This allows for comprehensive image acquisition of the fruit's appearance during rotation, particularly focusing on the concave top and tail areas, thus enhancing the completeness of the 3D modeling.

[0034] During the slow rotation of the support, each sensor captures color or single-band images of the fruit at equal time steps, based on its focal length and frequency. The images are saved in chronological order of capture time and used for pose analysis and iterative model training in subsequent 3D reconstruction. A small display and control panel are embedded in the side of the instrument, allowing for real-time viewing of the image quality of each individual sensor before shooting. Shooting parameters, including exposure time, ISO sensitivity, aperture, white balance, focal length, and focus mode, can be manually adjusted based on the results to achieve optimal image quality. After shooting, the sensor data is stored in a data storage device and numbered chronologically and by sensor, providing a data source for subsequent modeling.

[0035] A geometric and color calibration device is installed on the circular disk at the bottom of the instrument. Four scales are embedded around the center of the disk, forming an inscribed square to ensure that at least one scale image is visible on all four sides of the 3D scene reconstruction result. A red LED is placed at each node of the scale to quickly locate the start and end points of the scale in subsequent scene modeling results, reducing geometric calibration errors. The scale facing the sensor support is colored differently from the other three sides to aid in spatial matching of multiple 3D models of the same object. Two mounting slots for colorimetric plates are provided on the vertical side of the sensor rotating support. Before instrument operation, strip-shaped colorimetric calipers are embedded in the slots to evaluate the surface color difference of the 3D reconstruction model. After the instrument is stopped, the calipers can be removed or replaced to prevent surface color contamination or fading.

[0036] After the fruit's omnidirectional data acquisition is completed, the images are aggregated into the instrument's image data storage module according to the individual sensor's capture sequence and sensor number. This data is then used for subsequent image pose analysis on a high-performance hardware and software platform. A database mapping is used to record the capture time and corresponding image of the time-series image set for subsequent 3D reconstruction modeling. During this process, users can customize the deletion of images and the editing of file names within the library to save storage space. During single-target scanning, the data storage module receives capture data from each sensor in real time and stores it in folders based on the sensor's unique number. After scanning is complete, the omnidirectional images of the single target are stored in five subfolders for easy viewing of the image data quality captured by each sensor.

[0037] The storage module of the 3D reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops is divided into internal and external storage modules. The internal storage module is built into the instrument's control module, providing temporary storage for multi-directional image data of the scanned target. The internal storage module monitors the remaining storage space in real time to prevent memory overflow. Once the instrument's memory exceeds 98%, the storage module will activate an early warning function, reminding the user to move data to the external storage module or delete data from memory through both an alarm on the control panel and a warning light on the start-up bracket. Users can access and view the data in the internal storage module through the instrument's control module. The external storage module is a user-defined mobile storage device or computer, storing data in the same format as the internal storage module, and is used for 3D reconstruction modeling on a high-performance hardware platform.

[0038] The control module of the fruit appearance phenotypic 3D reconstruction modeler includes support rotation control, sensor performance control, and shooting light environment control. To acquire omnidirectional images of the fruit, the support rotates 360° clockwise around the fruit. Therefore, the rotation speed of the support needs to be controlled, with images taken at equal time intervals during this rotation. The rotation speed is controlled by a knob with a default speed limit, which can be manually adjusted in real time by the user to ensure efficient data acquisition and to ensure that the fruit's hanging state is not affected during the acquisition process.

[0039] During shooting, if the user pauses shooting, the sensor remains in the corresponding position until the user restarts the instrument. If the user stops shooting midway, the sensor bracket moves clockwise back to its initial position to await the next start. The corresponding dataset will be displayed on the controller interface, along with a dialog box asking whether to save. For a single sensor, the user can use the control screen on the side of the bracket to check if the shooting parameter settings are appropriate and adjust the parameters using the keyboard next to the control screen. The user can compare the image with the color calibration device within the field of view until the image is clear and the colors are accurate.

[0040] During the shooting process, all sensors maintain constant shooting parameters to suppress systematic errors in the 3D reconstruction modeling process. When shooting indoors using a top light source, to suppress the impact of environmental shadows on modeling quality, the brightness of the bottom circular light strip can be adjusted using the control keyboard to ensure that the fruit's appearance, color, and texture information are accurately depicted. During shooting, the sensors will also analyze the target object's movement. If movement occurs, the user will be prompted to pause the instrument. After user confirmation, the controller will adjust the sensor bracket to rotate counterclockwise, moving it to a position similar to where the image movement occurred before, and deleting the target fruit images captured during the warning period.

[0041] The control module connects the portable 3D reconstruction modeler of fruit and vegetable appearance phenotypic traits to a mobile storage device or computer, transferring scanned data to the memory via a data cable. This allows users to easily transfer data and store large amounts of data on a high-performance hardware platform for 3D reconstruction modeling and phenotypic analysis. Users can easily view saved folders on the controller screen and perform operations such as searching, deleting, and renaming. If some sensors experience imaging problems or data cannot be stored, the control module will display a warning message on the visual screen before starting shooting, allowing the user to decide whether to proceed with shooting and data storage.

[0042] The working principle of this invention is: The fruit appearance phenotypic trait imaging sensor bracket features a detachable clip at the top to secure the fruit, allowing for natural fruit hanging and capturing unobstructed images from top to bottom. Five small high-definition cameras are vertically arranged from top to bottom on the middle layer of the instrument bracket's side. A controllable speed rotation mechanism drives the camera bracket layer to slowly rotate clockwise, continuously capturing multi-view images around the fruit at set time intervals. These images are then stored and their poses analyzed. Based on the image sequence and corresponding pose analysis results, a fruit appearance simulation model is constructed, resulting in a true-color 3D model. A scale disk at the bottom of the instrument is used to calibrate the spatial dimensions of the modeling results, supporting subsequent calculations of the fruit's spatial structure phenotypic traits after outputting a solid point cloud or mesh model.

[0043] Example 1: Fruits and vegetables less than 50cm in length, including their stems, were harvested and suspended within the data acquisition space of a portable 3D reconstruction modeling instrument for fruit and vegetable appearance phenotypic traits using a fruit-fixing bracket, ensuring the fruit remained stable without significant shaking. After the instrument was powered on, the imaging quality of the five sensors from top to bottom was observed through the visualization screen of the control module. Sensor shooting parameters, including exposure time, ISO sensitivity, aperture, white balance, focal length, and focus mode, were controlled via buttons to suppress image blurring and artifacts during data acquisition.

[0044] Adjusting the sensor bracket's rotation speed via the controller knob allows the multi-head sensor to acquire multi-view images of the target fruit at a uniform speed and time step, providing data for 3D reconstruction modeling. In low-light conditions or with strong overhead lighting, the light intensity knob can be turned on and adjusted to control the brightness of the circular light strip, suppressing the effects of insufficient ambient light and shadows, thus ensuring modeling quality. After setting the sensor and shooting parameters, start the data acquisition program.

[0045] Due to the limited height of the support frame, to ensure the integrity of subsequent modeling, before shooting, the user needs to view the initial images captured by the sensors above and below the sides of the fruit through the controller interface to ensure that the top and bottom images of the fruit can be completely acquired. During data acquisition, the multi-head sensor acquires multi-view images of the target fruit's appearance and phenotypic characteristics at a uniform rotation speed and fixed time step, and stores the images to the instrument's internal memory in real time. If the user pauses shooting during the process, the sensor support frame will remain in the corresponding position and will no longer rotate clockwise until the user restarts the instrument. If the user stops shooting midway, the sensor support frame will move clockwise back to the initial position to await the next start. The corresponding dataset will be displayed on the controller interface, and a dialog box will appear asking whether to save.

[0046] To ensure the quality of 3D reconstruction modeling of fruit appearance phenotypic traits, the control system is activated to analyze the shaking of the target object during multi-view image capture. If shaking occurs, the controller will issue a reminder to the user to pause the instrument on the visual interface. After the user confirms, the controller will adjust the sensor bracket to rotate counterclockwise and move it to a position similar to the position before the image shaking occurred, and delete the target fruit images captured during the warning period.

[0047] During instrument data acquisition, the control module monitors memory usage in real time. If data overflow occurs, the controller activates a memory alarm light to remind the user to export historical data and clear instrument memory. Upon memory overflow, the instrument pauses operation and alerts the user to process stored data by illuminating the memory alarm light. After a single data acquisition cycle, the user can retrieve and view stored images through the controller interface and delete redundant or unwanted images. Users can also batch export acquired images to external storage and editing devices via the data transmission module for further analysis and editing. The images acquired in this invention are subsequently analyzed for shooting pose, optimized for color using the instrument's color chart, and input into general 3D reconstruction modeling software for true-color modeling of fruit appearance phenotypic traits.

[0048] Because four length scales are set on the bottom disk during image acquisition, it can be ensured that at least one scale is present in the multi-view images of the modeling target. Using the spatial reference provided by the scales, accurate spatial geometric reference information can be given to the 3D reconstruction modeling results of the fruit, which can then be used for the extraction and statistical analysis of spatial structural features of the phenotypic traits of fruit appearance in fruit and vegetable crops.

[0049] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the technical scope of the present invention. Therefore, any minor modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A portable three-dimensional reconstruction modeling instrument for the phenotypic traits of fruit and vegetable crops, characterized in that, include: Fruit appearance phenotypic characteristics imaging sensor bracket, fruit suspension fixing device, multi-head sensor module, model geometry and color calibration device, instrument imaging data storage module, and fruit appearance phenotypic characteristics three-dimensional reconstruction modeling instrument control module. The fruit appearance phenotypic morphology imaging sensor bracket is used to suspend and fix the fruit to be modeled. The side of the fruit appearance phenotypic imaging sensor bracket is a multi-head sensor module of a 3D reconstruction modeler, which is used to collect omnidirectional images of the fruit and perform 3D reconstruction. The fruit suspension fixing device secures the fruit stem of the fruit or the stem with a tie by means of a buckle. The multi-head sensor module consists of five high-definition cameras, which capture images of the fruit's appearance from different angles. The model geometry and color calibration device consists of a geometry calibration base and a color calibration caliper. The instrument's image data storage module is used to store images acquired by multiple sensors respectively; The control module of the fruit appearance phenotypic 3D reconstruction modeler allows users to turn the device on and off, set the time step interval during the multi-sensor module's sequential shooting process, set the shooting parameters of a single sensor, adjust the rotation speed of the fruit appearance phenotypic trait shooting sensor bracket, control the auxiliary lighting during the shooting process, and allows users to browse, query, edit, and copy the stored data.

2. The portable three-dimensional reconstruction modeling instrument for the appearance phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The fruit appearance phenotypic trait imaging sensor bracket fixes the fruit stem part of the fruit by suspension. Based on multi-head sensor, it captures and records the image set of the appearance phenotypic traits of fruit and vegetable crops from all directions. It supports three-dimensional reconstruction modeling of the target fruit and realizes non-destructive high-fidelity recording of the appearance structure and color traits of the fruit.

3. The portable three-dimensional reconstruction modeling instrument for the appearance phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The fruit appearance and phenotypic characteristics imaging sensor bracket is an approximately arc-shaped structure with a fruit suspension fixing device in the middle. The fruit appearance and phenotypic characteristics imaging sensor bracket has three layers, with the inner and outer layers fixed and the middle layer installing a multi-head sensor module. The middle layer can rotate clockwise around the fruit suspension fixing device; the outermost part of the fruit appearance phenotypic trait imaging sensor bracket is the control module of the fruit appearance phenotypic trait three-dimensional reconstruction modeling instrument, which is equipped with a control screen and a simple operation keyboard.

4. The portable three-dimensional reconstruction modeling instrument for the appearance phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The multi-head sensor module is fixed on the middle support at the upper oblique end of the top of the fruit when it is hanging, the upper, middle and lower parts of the fruit body, and the lower side of the tail of the fruit. During the rotation, it can realize the all-round image acquisition of the appearance of the fruit, especially the images of the top and tail of the fruit. Before the modeling data is captured, the shooting parameters of each sensor can be adjusted separately. After the instrument is started, as the sensor support slowly rotates, the sensor captures images of the corresponding pose around the fruit at equal time steps by setting the focal length and frequency. The images are numbered and saved in the order of shooting time and used for model iterative training in the process of shooting pose analysis and 3D reconstruction.

5. A portable three-dimensional reconstruction modeling instrument for the phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The mounting bracket of the multi-head sensor module is embedded in the middle layer of the fruit appearance and phenotypic characteristics imaging sensor bracket. After the instrument is moved, the fixing bracket of the multi-head sensor module will rotate clockwise at a uniform speed around the fruit suspension fixing device and the modeling target object according to the user-set speed and shooting time step. If the user pauses the instrument during the modeling data acquisition process, the sensor bracket will remain in the corresponding position until the user restarts the instrument and continues to complete the multi-view image capture of the remaining positions clockwise. If the user stops capturing the image midway, the sensor bracket will move clockwise back to the initial position to wait for the next startup, that is, it will be embedded in the instrument bracket, and all sensors will be in a powered-off state, waiting for the next startup.

6. A portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The instrument is equipped with the aforementioned geometric and color calibration device at its bottom. The geometric calibration device is a disc with four scales embedded around its center, forming an inscribed square. A red LED is installed at each node of the scale. The scale on the side of the caliper facing the sensor bracket is set to a different color than the other three sides to aid in spatial matching of multiple 3D models of the same object.

7. A portable three-dimensional reconstruction modeling instrument for the phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, A white LED light strip with adjustable brightness is installed around the disc; Before operating the instrument, insert the bar colorimeter into the slot; after the instrument is stopped, remove or replace the caliper.

8. A portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, During the slow rotation of the multi-head sensor bracket, by setting the focal length and shooting frequency of each camera, images of the corresponding poses are captured around the fruit at equal time steps. Images captured by a single camera are numbered and saved in chronological order of capture time. A small display and shooting parameter control device are embedded on the side of the instrument, allowing users to view the image quality of each individual sensor in real time before shooting, and manually adjust shooting parameters based on the imaging results, including: exposure time, ISO sensitivity, aperture, white balance, focal length, and focus mode sensor parameters.

9. A portable three-dimensional reconstruction modeling instrument for the appearance and phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The control module of the three-dimensional reconstruction modeling instrument for fruit appearance phenotypic traits includes: support rotation control, sensor performance control, and shooting light environment control; The single sensor uses the control screen on the side of the bracket to observe the shooting parameter settings and uses the keyboard next to the control screen to adjust the parameters; the control module controls the rotation speed of the multi-sensor bracket. If the user stops or finishes shooting, the corresponding dataset will be displayed on the controller's visual interface, and a dialog box will appear asking whether to save; during the shooting process, the shooting parameters of each sensor remain unchanged; During the shooting process, the sensor will be activated to analyze the shaking of the target object. If shaking occurs, the user will be reminded to pause the instrument. After the user confirms, the controller will adjust the sensor bracket to rotate counterclockwise and move it to a position similar to the position before the image shaking occurred. The instrument will also automatically delete the target fruit images taken during the warning period.

10. A portable three-dimensional reconstruction modeling instrument for the appearance phenotypic traits of fruit and vegetable crops according to claim 1, characterized in that, The control module of the fruit appearance phenotypic 3D reconstruction modeling instrument connects the portable fruit and vegetable crop fruit appearance phenotypic 3D reconstruction modeling instrument to a mobile storage device or computer, and transmits the scanned data to the memory via a data cable. The user can then transfer the data and perform modeling analysis on a high-performance server to achieve 3D reconstruction modeling of the fruit appearance. If some sensors have imaging problems or the data cannot be stored, the control module will display a warning message on the visualization screen before starting the shooting, allowing the user to decide whether to shoot and store the data.

Images