Pneumatic pepper plant harvester and method of fruit separation thereof
By combining image recognition and negative pressure sensors, pixel-level precision separation of fruits and light impurities is achieved in chili harvesters, solving the problem of insufficient separation accuracy in existing technologies and reducing fruit damage and impurity residue.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUIZHOU INST OF MOUNTAIN AGRI MACHINERY
- Filing Date
- 2026-02-10
- Publication Date
- 2026-06-26
Smart Images
Figure CN122271128A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of harvester technology, and in particular to a pneumatic chili plant harvester and a method for separating the fruit from the harvester. Background Technology
[0002] Mechanized harvesting of chili peppers is a key step in improving agricultural production efficiency. The core challenge lies in how to efficiently and cleanly separate the fruit from light impurities such as plants, leaves, and stems, while avoiding mechanical damage to the delicate fruit. Pneumatic separation technology is widely used due to its non-contact and wide adaptability.
[0003] Existing pneumatic harvesting and separation methods mainly fall into two categories. One is pure airflow sieving, which uses the airflow field generated by a blower to sort fruits and impurities based on differences in weight, size, or aerodynamic characteristics. However, chili pepper fruits and lightweight impurities often have highly overlapping physical properties, and the state of materials in the field is complex and variable, resulting in limited separation accuracy for this method. Either there is a lot of impurity residue, or the fruits are blown away together, leading to a high loss rate. The other method attempts to introduce negative pressure adsorption as a fixation method, first adsorbing the mixture onto a conveyor belt, and then supplementing it with airflow purging. However, existing solutions usually treat negative pressure adsorption as a simple, globally uniform fixation action. This leads to subsequent airflow purging, which is usually global or simply zoned, lacking precise guidance. It is still in a general attack mode, making it difficult to ensure that the fruits remain firmly in place while blowing away impurities. The separation selectivity is poor, and the problems of fruit damage and impurity residue are not fundamentally solved. Therefore, how to achieve pixel-level intelligent fruit separation of chili pepper plants based on the microscopic differences in the stable state of airflow chamber adsorption has become a challenge for the industry. Summary of the Invention
[0004] Based on this, this application provides a pneumatic chili harvester and its fruit separation method that performs pixel-level intelligent fruit separation of chili plants based on the microscopic differences in the stable adsorption state of the airflow chamber.
[0005] In a first aspect, this application provides a method for separating fruits in a pneumatic chili harvester, comprising the following steps: Image information of the mixture spread on the porous negative pressure adsorption belt inside the chili plant harvester was collected; Visual recognition of chili pepper fruits and light impurities is performed on the image information to obtain the outlines of chili pepper fruits and light impurities and their corresponding two-dimensional positions. At the same time, the real-time negative pressure of the corresponding independent airflow chamber is monitored by the air pressure sensor set in each independent airflow chamber below the adsorption zone to obtain the negative pressure stability of each independent airflow chamber. Based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, as well as the negative pressure stability of each independent airflow chamber, the areas that are visually identified as chili pepper fruits and have corresponding real-time stable negative pressure are marked as fruit retention areas, and the areas that are visually identified as light impurities or have corresponding real-time abnormal negative pressure fluctuations are marked as impurity removal areas. A depressurization-brief positive pressure backflushing sequence is performed on the independent airflow chamber corresponding to the impurity removal zone to remove light impurities, while the current adsorption state is maintained in the independent airflow chamber corresponding to the fruit holding zone. After impurities are removed from the impurity removal area, all independent airflow chambers are subjected to controlled gradient depressurization until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected.
[0006] In some embodiments, visually recognizing chili pepper fruits and light impurities from the image information to obtain the outlines of the chili pepper fruits and light impurities and their corresponding two-dimensional positions specifically includes: The image information is preprocessed to enhance its features; The preprocessed image information is input into a pre-trained semantic segmentation model, which outputs pixel-level classification results. Based on the pixel-level classification results, the outlines of chili pepper fruits and light impurities were extracted; The image coordinates corresponding to the outlines of the chili fruit and light impurities are converted into two-dimensional position coordinates relative to the plane of the porous negative pressure adsorption zone.
[0007] In some embodiments, the real-time negative pressure of each independent airflow chamber is monitored by a pressure sensor disposed in each independent airflow chamber below the adsorption zone, and the negative pressure stability of each independent airflow chamber is obtained specifically by: The real-time negative pressure value sequence of each independent airflow chamber is obtained by the air pressure sensor set in each independent airflow chamber below the adsorption zone at a fixed sampling frequency. A fixed-time sequence of real-time negative pressure values is extracted for each independent airflow chamber as an analysis window; The standard deviation of the real-time negative pressure value sequence within each analysis window is used as the quantification value of the negative pressure stability of the corresponding independent airflow chamber.
[0008] In some embodiments, based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, and the negative pressure stability of each independent airflow chamber, areas visually identified as chili pepper fruits and corresponding to stable real-time negative pressure are marked as fruit retention areas, and areas visually identified as light impurities or corresponding to abnormal fluctuations in real-time negative pressure are marked as impurity removal areas. Specifically, this includes: Based on the two-dimensional position coordinates corresponding to the outlines of the chili fruit and the light impurities, the independent airflow chambers covered by each outline are determined. For the outline of each chili pepper fruit, query the negative pressure stability of all the individual airflow chambers it covers; If the negative pressure stability of all independent airflow chambers is lower than the preset stability threshold, the outline area of the chili fruit is marked as the fruit retention area; otherwise, the outline area of the chili fruit is marked as the impurity removal area. For each light impurity's outline, the outline area of the light impurity is directly marked as the impurity to be removed area.
[0009] In some embodiments, performing a depressurization-brief positive pressure backflushing sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities specifically includes: Obtain the corresponding independent airflow chamber number based on the area information marked as the impurity removal zone; Send a control command to the airflow valve corresponding to the independent airflow chamber to switch it to the depressurization state in order to release the negative pressure adsorption; After confirming that the corresponding independent airflow chamber has been depressurized, a control command is sent to the airflow valve to switch it to connect to a positive pressure air source and maintain this connection for a short, set time to generate an upward pulse airflow that blows away light impurities.
[0010] In some embodiments, maintaining the current adsorption state of the independent airflow chamber corresponding to the fruit holding area specifically includes: The corresponding independent airflow chamber number is obtained based on the area information marked as the fruit preservation area; Send control commands to the airflow valves corresponding to the independent airflow chambers to keep them connected to the negative pressure source.
[0011] In some embodiments, after impurity removal is completed in the impurity removal area, all independent airflow chambers are subjected to controlled gradient depressurization until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected. Specifically, this includes: Determine whether all independent airflow chambers marked as areas to be cleaned of impurities have completed the depressurization-brief positive pressure backflush sequence; If completed, a gradient pressure relief control command is sent to the airflow valves corresponding to all independent airflow chambers, so that they are slowly connected to the atmosphere at a controlled rate until the negative pressure in each chamber drops to zero. After the negative pressure is completely released, collect the chili peppers that have gently fallen off the porous negative pressure adsorption belt.
[0012] Secondly, this application provides a pneumatic chili plant harvester, which includes a fruit separation unit, the fruit separation unit comprising: The acquisition module is used to acquire image information of the mixture spread on the porous negative pressure adsorption belt inside the chili plant harvester; The processing module is used to visually identify the chili pepper fruit and light impurities in the image information, obtain the outline of the chili pepper fruit and light impurities and their corresponding two-dimensional positions. At the same time, the real-time negative pressure of the corresponding independent airflow chamber is monitored by the air pressure sensor set in each independent airflow chamber below the adsorption zone, and the negative pressure stability of each independent airflow chamber is obtained. The processing module is also used to mark the area that is visually identified as a chili fruit and has a stable real-time negative pressure as the fruit retention area, and the area that is visually identified as a light impurity or has an abnormal real-time negative pressure fluctuation as the impurity removal area, based on the outline and corresponding two-dimensional position of the chili fruit and light impurities, as well as the negative pressure stability of each independent airflow chamber. The processing module is also used to perform a depressurization-brief positive pressure backflushing operation sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities and maintain the current adsorption state of the independent airflow chamber corresponding to the fruit holding area. The execution module is used to perform controlled gradient depressurization of all independent airflow chambers after the impurities are removed in the impurity removal area, until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected.
[0013] Thirdly, this application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described fruit separation method of a pneumatic chili plant harvester.
[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described fruit separation method for a pneumatic chili plant harvester.
[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The pneumatic chili harvester and its fruit separation method provided in this application firstly acquire image information of the mixed material surface laid flat on the porous negative pressure adsorption belt inside the chili harvester. This step enables stable acquisition of high-definition, motion-blur-free material surface images, providing a reliable data foundation for subsequent accurate visual recognition, thereby improving the input quality and reliability of the entire separation system's sensing link. Secondly, the image information is used for visual recognition of chili fruits and light impurities to obtain the outlines of the chili fruits and light impurities and their corresponding two-dimensional positions. Simultaneously, air pressure sensors installed in each independent airflow chamber below the adsorption belt monitor the corresponding independent airflow chambers. The real-time negative pressure is used to obtain the negative pressure stability of each independent airflow chamber. This step enables the simultaneous perception and quantification of multimodal information on the material surface morphology and adsorption microstate, combining what is seen with whether the adsorption is stable, thus providing a more comprehensive judgment basis beyond single vision for subsequent accurate decision-making. Subsequently, based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, as well as the negative pressure stability of each independent airflow chamber, the areas visually identified as chili pepper fruits and corresponding to stable real-time negative pressure are marked as fruit retention areas, and the areas visually identified as light impurities or corresponding to abnormal fluctuations in real-time negative pressure stability are marked as impurity removal areas. This step can be based on multiple... The refined spatial decision-making based on source information fusion accurately distinguishes between reliable fruits that need protection and impurities or abnormal areas that need to be removed, generating an action map to guide execution. This elevates subsequent airflow action from coarse zoning to pixel-level targeted planning. Then, a depressurization-brief positive pressure backflushing sequence is executed on the independent airflow chambers corresponding to the impurity removal area to remove light impurities. The current adsorption state is maintained in the independent airflow chambers corresponding to the fruit retention area. This step can be precisely executed based on the differences in the action map, applying a directional airflow that first loosens and then blows away impurity areas while maintaining stable adsorption in the fruit area. This efficiently removes impurities while minimizing airflow disturbance. The flow disturbs and damages the fruit, achieving selective separation. Finally, after impurity removal is completed in the impurity removal area, all independent airflow chambers are subjected to controlled gradient depressurization until the negative pressure is completely released, allowing the pepper fruits on the adsorption belt to fall off smoothly and be collected. This step enables flexible and controlled release of the fruit's adsorption force, avoids impact collisions caused by sudden pressure changes through gradient depressurization, and completes collection in a buffered manner, thereby significantly reducing mechanical damage to the delicate pepper fruits during harvesting and ensuring the final commercial quality of the fruit. In summary, the solution of this application can perform pixel-level intelligent fruit separation of pepper plants based on the microscopic differences in the stable adsorption state of the airflow chambers. Attached Figure Description
[0016] Figure 1 This is an exemplary flowchart of a fruit separation method for a pneumatic chili plant harvester according to some embodiments of this application; Figure 2 This is a schematic diagram illustrating an application scenario of a fruit separation data processing system according to some embodiments of this application; Figure 3 This is a schematic flowchart illustrating the process of determining negative pressure stability according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a fruit separation unit according to some embodiments of this application; Figure 5 This is a schematic diagram of the structure of a computer device for implementing a fruit separation method of a pneumatic chili plant harvester according to some embodiments of this application. Detailed Implementation
[0017] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.
[0018] refer to Figure 1 The figure is an exemplary flowchart of a fruit separation method for a pneumatic chili harvester according to some embodiments of this application. The fruit separation method of the pneumatic chili harvester mainly includes the following steps: In step 101, image information of the mixture spread on the porous negative pressure adsorption belt inside the chili plant harvester is acquired.
[0019] In specific implementation, the image information of the mixture spread on the porous negative pressure adsorption belt inside the chili harvester can be achieved in the following way: Inside the frame of the chili harvester, directly above the active and driven rollers of the porous negative pressure adsorption belt, spanning the entire width of the belt, a rigid mounting beam is fixedly installed; on this rigid mounting beam, at least two industrial color area array cameras are evenly spaced along the width direction of the porous negative pressure adsorption belt, forming a machine vision unit covering the full width of the belt; simultaneously, a strip-shaped LED white light source synchronized with the shooting sequence of the industrial color area array cameras is integrated on the mounting beam to provide uniform, stable, and shadowless illumination for the surface of the porous negative pressure adsorption belt, eliminating interference caused by changes in ambient light inside the harvester and the shadows cast by the belt itself; the imaging plane of the industrial color area array camera is aligned with the porous negative pressure adsorption belt... The bearing plane of the belt is parallel, and a fixed focal length and aperture are set to ensure that the depth of field of the image within the entire field of view is sufficient to clearly image the surface of the mixed material. The triggering of image acquisition is synchronized with the continuous running speed of the porous negative pressure adsorption belt. Specifically, a pulse signal is emitted by a rotary encoder installed on the drive shaft of the porous negative pressure adsorption belt to control the industrial color area array camera to perform fixed-distance trigger shooting. That is, the porous negative pressure adsorption belt acquires one frame of image every time it travels a fixed distance, such as 10 mm, ensuring that the acquired images are seamlessly stitched and will not be blurred due to movement, thereby obtaining the image information of the surface of the mixed material. It should be noted that the image information of the surface of the mixed material can be transmitted in real time through the high-speed data interface of the industrial color area array camera to the memory of the central processing unit of the chili plant harvester for caching, so as to be called and processed in subsequent steps. Other methods can also be used in other embodiments, which are not limited here.
[0020] It should be noted that the above steps can achieve stable acquisition of high-definition, motion-blur-free material surface images, providing a reliable data foundation for subsequent accurate visual recognition, thereby improving the input quality and reliability of the entire separation system's sensing link.
[0021] In some embodiments, reference Figure 2 As shown in the figure, this figure is a schematic diagram of the application scenario of the fruit separation data processing system shown in some embodiments of this application. The figure includes three main components: a data acquisition device, a server, and a data storage device. The data acquisition device is responsible for collecting image information of the mixed material surface laid flat on the porous negative pressure adsorption belt inside the chili plant harvester, and sending the collected image information of the mixed material surface to the server through a communication network. The fruit separation data processing system runs in the server, and the server stores the processing results in the data storage device and visualizes them.
[0022] In step 102, the image information is visually identified to obtain the outlines of the chili peppers and light impurities and their corresponding two-dimensional positions. At the same time, the real-time negative pressure of each independent airflow chamber is monitored by the air pressure sensor set in each independent airflow chamber below the adsorption zone to obtain the negative pressure stability of each independent airflow chamber.
[0023] In some embodiments, visual recognition of chili pepper fruits and light impurities from the image information to obtain the outlines of the chili pepper fruits and light impurities and their corresponding two-dimensional positions can be achieved by the following steps: The image information is preprocessed to enhance its features; The preprocessed image information is input into a pre-trained semantic segmentation model, which outputs pixel-level classification results. Based on the pixel-level classification results, the outlines of chili pepper fruits and light impurities were extracted; The image coordinates corresponding to the outlines of the chili fruit and light impurities are converted into two-dimensional position coordinates relative to the plane of the porous negative pressure adsorption zone.
[0024] It should be noted that the preprocessing in this application refers to a series of optimization operations applied to the original image information before visual recognition. Its purpose is to improve image quality, highlight target features, and suppress interference, so as to ensure that the subsequent recognition algorithm can obtain clear, stable and feature-distinctive input data, thereby improving recognition accuracy and reliability.
[0025] In specific implementation, the image information is preprocessed to enhance features, which can be achieved in the following way: After reading the image information acquired by the industrial color area array camera, color space conversion is first performed to convert the acquired RGB image into the HSV color space, which is more suitable for color segmentation, to highlight the hue difference between the chili fruit and the light impurities and background; then, the V channel is subjected to contrast-limited adaptive histogram equalization to compensate for uneven illumination and enhance the overall contrast of the image; next, Gaussian filtering is performed on the image to suppress random noise that may be introduced during transmission, while maintaining the clarity of object edges; finally, according to the fixed position of the porous negative pressure adsorption strip in the image, region of interest clipping is performed to retain only the strip-shaped area containing the mixed material, and the useless frame background on both sides of the image is removed, thereby reducing the amount of subsequent calculation and focusing on effective information, thus completing the preprocessing of the image information to enhance features. Other methods can also be used in other embodiments, which are not limited here.
[0026] In specific implementation, the preprocessed image information is input into a pre-trained semantic segmentation model, and the pixel-level classification result can be output in the following way: The preprocessed image information is used as a three-dimensional array, namely height, width, and color channels, and input into a pre-trained semantic segmentation model. The semantic segmentation model runs in a feedforward inference manner, classifying the input image pixel by pixel; the model outputs a feature map of the same size as the input image, where each pixel contains a three-dimensional vector, representing the probability that the pixel belongs to the three categories of chili pepper fruit, light impurities, and background; by taking the category corresponding to the maximum probability value at each pixel position, the final category label of the pixel can be determined, thereby generating a pixel-level classification result map, in which each pixel is clearly labeled as chili pepper fruit, light impurities, or background. Other methods can also be used in other embodiments, which are not limited here.
[0027] In specific implementation, the semantic segmentation model can be pre-trained in the following manner: In the offline stage, a large number of sample images of chili plant materials under different lighting conditions, varieties, and impurity coverage are pre-collected on the porous negative pressure adsorption belt. Human annotation experts accurately annotate the chili fruits, leaves, stems, and other lightweight impurities in each sample image at the pixel level, forming a training dataset. A convolutional neural network with an encoder-decoder structure, such as U-Net or its variants, is used as the model architecture, with the sample images as input and the corresponding manually annotated images as supervision signals. Iterative training is performed on a graphics processing unit cluster using stochastic gradient descent or the Adam optimizer, with the cross-entropy loss function as the optimization objective, until the model's classification accuracy on an independent validation dataset reaches a preset threshold, such as mIoU greater than 0.85. This completes the pre-training of the semantic segmentation model. The trained model parameters are then solidified and deployed to the central processing unit of the chili plant harvester. Other methods can also be used in other embodiments, which are not limited here.
[0028] In specific implementation, the extraction of the outlines of chili pepper fruits and light impurities based on the pixel-level classification results can be achieved in the following way: After obtaining the pixel-level classification result image, the two categories of chili pepper fruits and light impurities are processed separately; for each category, morphological opening operation is first used, erosion followed by dilation to remove small isolated pixels caused by noise in the classification result, and morphological closing operation is used, dilation followed by erosion to fill the small holes inside the category caused by misclassification, thereby obtaining a purer connected region; then, an edge detection algorithm, such as the Canny algorithm or the connected component labeling algorithm, is used to find all connected pixel regions belonging to the category from the processed binary category image; finally, for each connected region, its outermost pixel sequence is extracted, which constitutes a closed polygon outline, which is the outline of an independent chili pepper fruit or a piece of light impurity. Other methods can also be used in other embodiments, which are not limited here.
[0029] In specific implementation, converting the image coordinates corresponding to the outlines of the chili pepper fruit and light impurities into two-dimensional position coordinates relative to the plane of the porous negative pressure adsorption belt can be achieved in the following way: After the harvester is assembled and before its first operation, the machine vision unit needs to be calibrated. This involves using a known-sized checkerboard calibration plate laid flat on the bearing plane of the porous negative pressure adsorption belt, and having the industrial color area array camera capture multiple images of the calibration plate at different positions. The camera's internal and external parameters are then solved using the Zhang Zhengyou calibration method. The internal parameters are the focal length and principal point, and the external parameters are the rotation and translation matrices relative to the plane of the porous negative pressure adsorption belt. An image pixel coordinate system is then established. The mapping relationship between the plane of the porous negative pressure adsorption belt and the world coordinate system; during real-time conversion, for each extracted contour, the center point of its smallest bounding rectangle or the centroid of the contour is taken as its representative point image coordinates in the frame image; using the homography matrix or camera parameters obtained by calibration, the point is transformed from image coordinates to coordinates in a two-dimensional mechanical coordinate system based on the plane of the porous negative pressure adsorption belt through perspective transformation formula. The coordinates are in millimeters, and the origin can be set at a fixed corner point of the porous negative pressure adsorption belt, thereby obtaining the two-dimensional position coordinates of each chili fruit and light impurity contour on the plane of the porous negative pressure adsorption belt. Other methods can also be used in other embodiments, which are not limited here.
[0030] In some embodiments, reference Figure 3 As shown in the figure, this is a schematic flowchart of determining negative pressure stability in some embodiments of this application. In this embodiment, the real-time negative pressure of each independent airflow chamber is monitored by a pressure sensor installed in each independent airflow chamber below the adsorption zone. The negative pressure stability of each independent airflow chamber can be obtained by the following steps: In step 1031, the real-time negative pressure value sequence of each independent airflow chamber is obtained by the air pressure sensor set in each independent airflow chamber below the adsorption zone at a fixed sampling frequency. In step 1032, a real-time negative pressure value sequence of a fixed time length is extracted for each independent airflow chamber as an analysis window; In step 1033, the standard deviation of the real-time negative pressure value sequence within each analysis window is used as the quantification value of the negative pressure stability of the corresponding independent airflow chamber.
[0031] It should be noted that the negative pressure stability in this application is a quantitative indicator that characterizes the stability of the negative pressure adsorption capacity of an independent airflow chamber within a specific time period. Its function is to reflect in real time whether the adsorption state of the material above the chamber is stable and whether there are any abnormal fluctuations, providing key operating condition basis for judging whether the material is a fruit that has been reliably adsorbed or loosely attached impurities.
[0032] In specific implementation, the real-time negative pressure value sequence of each independent airflow chamber, which is installed in a pressure sensor located below the adsorption zone and sampled at a fixed frequency, can be obtained in the following way: A pressure sensor installed inside each independent airflow chamber, such as a digital output MEMS pressure sensor, is connected to the I / O module of the central processing unit (CPU) via an I2C or SPI bus; the CPU is configured with a high-precision timer to periodically trigger synchronous data reading commands for all pressure sensors at a fixed and sufficiently high sampling frequency, such as 1 kHz; after each reading command is issued, the pressure sensor returns the measured current chamber pressure value through a digital interface, which is usually the difference relative to atmospheric pressure, with a negative value indicating negative pressure; the CPU stores the read pressure value, along with the number of the independent airflow chamber to which the value belongs and a precise timestamp, into a first-in-first-out (FIFO) circular data buffer, thereby continuously generating a real-time negative pressure value sequence arranged in chronological order for each independent airflow chamber. Other methods can also be used in other embodiments, which are not limited here.
[0033] In specific implementation, the analysis window for each independent airflow chamber can be achieved by extracting a real-time negative pressure value sequence of a fixed time length. This can be done as follows: When the central processing unit needs to calculate negative pressure stability, for example, after the porous negative pressure adsorption belt has traveled a fixed distance, it extracts from the circulating data buffer a newly stored continuous real-time negative pressure value sequence for each independent airflow chamber. This sequence has a time length exactly equal to the preset window duration, for example, 0.5 seconds. Since the sampling frequency is fixed, the 0.5-second window corresponds to 500 consecutive real-time negative pressure value data points. This subsequence containing a fixed number of data points constitutes the analysis window used for the current calculation. It should be noted that the analysis window will slide and update as the porous negative pressure adsorption belt operates and data is continuously collected, ensuring that each calculation is based on the latest operating conditions. Other methods can also be used in other embodiments, and are not limited here.
[0034] In specific implementation, the standard deviation of the real-time negative pressure value sequence within each analysis window can be used as the quantification value of the negative pressure stability of the corresponding independent airflow chamber. This can be achieved in the following way: For each analysis window, firstly, calculate the arithmetic mean of all real-time negative pressure values within the analysis window; then, calculate the difference between each real-time negative pressure value and the arithmetic mean, square each difference, sum them up, and then divide by the total number of real-time negative pressure value data points within the analysis window. Finally, take the square root of the result. The resulting value is the standard deviation of the real-time negative pressure value sequence within the analysis window. This standard deviation value is directly used as the quantification value of the negative pressure stability of the corresponding independent airflow chamber at the current moment. It should be noted that the smaller the quantification value, the smaller the negative pressure fluctuation and the higher the stability of the independent airflow chamber within the analysis window period; conversely, the larger the quantification value, the larger the fluctuation and the worse the stability of the independent airflow chamber within the analysis window period, i.e., abnormal negative pressure stability fluctuation. Other methods can also be used in other embodiments, which are not limited here.
[0035] It should be noted that the above steps can realize the simultaneous perception and quantification of multimodal information on the surface morphology of materials and the microscopic state of adsorption, combining what is seen with whether the adsorption is stable, thus providing a more comprehensive judgment basis that goes beyond a single vision for subsequent accurate decision-making.
[0036] In step 103, based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, as well as the negative pressure stability of each independent airflow chamber, the areas visually identified as chili pepper fruits and corresponding to stable real-time negative pressure are marked as fruit retention areas, and the areas visually identified as light impurities or corresponding to abnormal fluctuations in real-time negative pressure are marked as impurity removal areas.
[0037] In some embodiments, based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, and the negative pressure stability of each independent airflow chamber, areas visually identified as chili pepper fruits and corresponding to areas with stable real-time negative pressure are marked as fruit retention areas, and areas visually identified as light impurities or corresponding to areas with abnormal fluctuations in real-time negative pressure are marked as impurity removal areas. This can be achieved through the following steps: Based on the two-dimensional position coordinates corresponding to the outlines of the chili fruit and the light impurities, the independent airflow chambers covered by each outline are determined. For the outline of each chili pepper fruit, query the negative pressure stability of all the individual airflow chambers it covers; If the negative pressure stability of all independent airflow chambers is lower than the preset stability threshold, the outline area of the chili fruit is marked as the fruit retention area; otherwise, the outline area of the chili fruit is marked as the impurity removal area. For each light impurity's outline, the outline area of the light impurity is directly marked as the impurity to be removed area.
[0038] It should be noted that the stability threshold in this application is a preset critical value used to determine whether the negative pressure stability is qualified. Its function is to serve as a decision-making benchmark, transforming the continuously changing stability quantification value into a binary judgment of stability and abnormality, thereby achieving accurate classification and labeling of different adsorption state regions.
[0039] In specific implementation, the determination of the independent airflow chambers covered by each contour, based on the two-dimensional position coordinates corresponding to the contours of the chili fruit and light impurities, can be achieved in the following way: Obtain the precise boundary range of each independent airflow chamber in the two-dimensional mechanical coordinate system of the porous negative pressure adsorption zone plane, which is pre-stored during the initialization phase. That is, each chamber corresponds to a rectangular area with known length, width, and center point coordinates. For a given contour, its two-dimensional position coordinates are usually represented by the coordinate set of several feature points on the contour, such as the centroid or the vertices of the minimum circumscribed rectangle. Determine which independent airflow chambers the contour covers, i.e., perform spatial geometric calculations to check whether the coordinate set of the feature points of the contour falls within the rectangular boundary range of a certain independent airflow chamber, or calculate whether there is an intersection between the minimum circumscribed rectangle of the contour and the rectangular areas of each independent airflow chamber. If there is an intersection or inclusion relationship, it is determined that the contour covers the independent airflow chamber. Finally, output a list of the numbers of the independent airflow chambers covered by each contour of the chili fruit and light impurities. Other methods can also be used in other embodiments, which are not limited here.
[0040] Additionally, it should be noted that a region marker list can be generated and maintained in real time. The data structure of this region marker list explicitly records the number of each region marked as an impurity removal area and the number of one or more independent airflow chambers corresponding to each region marked as a fruit retention area. Other methods can also be used in other embodiments, which are not limited here.
[0041] In specific implementation, the stability threshold can be preset in the following way: before the chili harvester leaves the factory or before operation for a specific variety, a threshold calibration experiment needs to be conducted; in the calibration experiment, the harvester is run unloaded under typical working conditions, and multiple sets of negative pressure stability quantification values of all the independent airflow chambers are recorded when only clean chili fruits are adsorbed, i.e., without impurities mixed in. The statistical distribution of these quantification values is calculated, and the average value plus several times the standard deviation, such as 2 times the standard deviation, is taken as the initial reference stability threshold; subsequently, the working conditions of slight blockage or impurity adhesion are simulated on the test bench, and the critical value of negative pressure fluctuation that causes obvious instability in fruit adsorption is observed and recorded, and the initial reference threshold is fine-tuned and verified; finally, the specific value that has been verified and can reliably distinguish between stable adsorption and abnormal fluctuation is used as the preset stability threshold, which is solidified and written into the configuration file of the central processing unit and called in real-time operation. Other methods can also be used in other embodiments, which are not limited here.
[0042] It should be noted that the above steps can be based on refined spatial decision-making through multi-source information fusion, accurately distinguishing between reliable fruits that need to be protected and impurities or abnormal areas that need to be removed, generating action maps to guide execution, thereby elevating the subsequent airflow effect from coarse zoning to pixel-level targeted planning.
[0043] In step 104, a depressurization-brief positive pressure backflushing operation sequence is performed on the independent airflow chamber corresponding to the impurity removal zone to remove light impurities, while maintaining the current adsorption state of the independent airflow chamber corresponding to the fruit holding zone.
[0044] In some embodiments, performing a depressurization-brief positive pressure backflushing sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities can be achieved by the following steps: Obtain the corresponding independent airflow chamber number based on the area information marked as the impurity removal zone; Send a control command to the airflow valve corresponding to the independent airflow chamber to switch it to the depressurization state in order to release the negative pressure adsorption; After confirming that the corresponding independent airflow chamber has been depressurized, a control command is sent to the airflow valve to switch it to connect to a positive pressure air source and maintain this connection for a short, set time to generate an upward pulse airflow that blows away light impurities.
[0045] It should be noted that the depressurization-brief positive pressure backflushing operation sequence in this application refers to a set of airflow control actions with a fixed sequence performed on an independent airflow chamber for a specific area, such as the impurity removal area. Its function is to first quickly release the adsorption and fixation force of the material in the area, making it loose, and then use a brief, directional upward airflow to efficiently blow the loosened light impurities away from the separation surface, thereby achieving selective removal.
[0046] In specific implementation, the acquisition of the corresponding independent airflow chamber number based on the area information marked as the impurity to be removed area can be achieved in the following way: the acquisition operation is to extract all entries marked as impurity to be removed areas from the data structure of the area marker list generated and maintained in real time in step S103, and read out the number field of the independent airflow chamber to form a set of numbers of the independent airflow chamber to be operated. Other methods can also be used in other embodiments, which are not limited here.
[0047] In specific implementation, sending a control command to the airflow valve corresponding to the independent airflow chamber to switch it to the depressurization state to release the negative pressure adsorption can be achieved in the following way: The central processing unit sends a specific digital signal to the drive circuit of the two-position three-way solenoid valve corresponding to the independent airflow chamber through its digital output module or fieldbus, such as CAN bus. This digital signal causes the valve core of the solenoid valve to actuate, cutting off the connection between the independent airflow chamber and the negative pressure fan pipeline, and at the same time switching its vent to connect with the atmospheric environment. At this time, external air enters the chamber rapidly through the pores of the porous negative pressure adsorption strip, causing its internal pressure to rise to the ambient atmospheric pressure within tens of milliseconds, thereby releasing the negative pressure adsorption force on the material in the area. Other methods can also be used in other embodiments, which are not limited here.
[0048] In specific implementation, after confirming that the corresponding independent airflow chamber has completed depressurization, a control command is sent to the airflow valve to switch it to connect to a positive pressure air source and maintain this connection for a set short period of time to generate an upward pulse airflow that blows away light impurities. This can be achieved in the following ways: confirming that depressurization is complete can be done using a fixed delay method or a pressure feedback method; wherein, the fixed delay method means that after the central processing unit issues the depressurization command, it waits for a preset fixed time, sufficient for the chamber pressure to balance, for example, 80 milliseconds; the pressure feedback method reads the real-time data from the pressure sensor in the corresponding independent airflow chamber, and determines that depressurization is complete when the pressure value stabilizes in a range close to atmospheric pressure; After the recognition is completed, the central processing unit immediately sends another digital control signal to the same solenoid valve through the digital output module, driving the valve core to act again, switching the air inlet of the corresponding independent airflow chamber to connect to a low-pressure positive pressure fan pipeline; at the same time, the central processing unit starts a high-precision timer to control the positive pressure connection state to continue for a preset short time, such as 150 milliseconds. During this period, low-pressure airflow, such as 0.8 kPa, continuously flows into the independent airflow chamber and sprays upward from the pores of the porous negative pressure adsorption belt, forming a brief pulse airflow that blows the loosened light impurities off the surface. Other methods can also be used in other embodiments, which are not limited here.
[0049] In some embodiments, maintaining the current adsorption state of the independent airflow chamber corresponding to the fruit holding area can be achieved by the following steps: The corresponding independent airflow chamber number is obtained based on the area information marked as the fruit preservation area; Send control commands to the airflow valves corresponding to the independent airflow chambers to keep them connected to the negative pressure source.
[0050] In specific implementation, obtaining the corresponding independent airflow chamber number based on the area information marked as the fruit preservation area can be achieved in the following way: corresponding to the method of "obtaining the corresponding independent airflow chamber number based on the area information marked as the impurity to be removed area", all entries marked as fruit preservation areas can be extracted from the same area marking list, and the independent airflow chamber numbers recorded in these entries can be read to form a set of numbers of independent airflow chambers that need to maintain adsorption. Other methods can also be used in other embodiments, which are not limited here.
[0051] In specific implementation, sending control commands to the airflow valves corresponding to the independent airflow chambers to maintain their connection to the negative pressure source can be achieved in the following way: For each of the independent airflow chamber number sets corresponding to the fruit holding area, the central processing unit sends a constant control signal to the drive circuit of the corresponding solenoid valve through its digital output module within the same control cycle to maintain the current path state; this control signal keeps the valve core of the solenoid valve in the initial position connected to the negative pressure fan pipeline, thereby ensuring that the corresponding independent airflow chamber continuously provides a stable negative pressure adsorption force during the entire separation operation, so that the material identified as pepper fruit is firmly attached to the porous negative pressure adsorption belt. Other methods can also be used in other embodiments, which are not limited here.
[0052] It should be noted that the above steps can be precisely executed according to the differences in the action map. The directional airflow that first loosens and then blows away the impurities is applied only to the impurity area, while maintaining stable adsorption on the fruit area. This achieves selective separation by minimizing the disturbance and damage to the fruit caused by the airflow while efficiently removing impurities.
[0053] In step 105, after the impurities are removed in the impurity removal area, a controlled gradient depressurization is performed on all independent airflow chambers until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can fall off smoothly and be collected.
[0054] In some embodiments, after impurity removal is completed in the impurity removal area, a controlled gradient depressurization is performed on all independent airflow chambers until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected. This can be achieved by the following steps: Determine whether all independent airflow chambers marked as areas to be cleaned of impurities have completed the depressurization-brief positive pressure backflush sequence; If completed, a gradient pressure relief control command is sent to the airflow valves corresponding to all independent airflow chambers, so that they are slowly connected to the atmosphere at a controlled rate until the negative pressure in each chamber drops to zero. After the negative pressure is completely released, collect the chili peppers that have gently fallen off the porous negative pressure adsorption belt.
[0055] It should be noted that the gradient pressure relief in this application refers to the process of controlling all independent airflow chambers to release negative pressure at a controlled and gradual speed during the fruit collection stage until it is completely released. Its purpose is to avoid the fruit from violently bouncing or colliding due to inertia caused by the instantaneous disappearance of the adsorption force, thereby achieving a flexible and low-damage release of the chili fruit and ensuring the quality of the finally harvested fruit.
[0056] In specific implementation, determining whether all independent airflow chambers marked as impurity-to-be-removed areas have completed the depressurization-brief positive pressure backflushing operation sequence can be achieved in the following way: maintain a status flag bit for each independent airflow chamber; when a depressurization command and a brief positive pressure backflushing command are sent to an independent airflow chamber marked as an impurity-to-be-removed area, and the backflushing phase is confirmed by its corresponding timer after completion, set the status flag bit of that chamber to "completed"; simultaneously, continuously check the status flag bits of all independent airflow chambers marked as impurity-to-be-removed areas in the area marking list; if and only if the status flag bits of all independent airflow chambers marked as impurity-to-be-removed areas are set to "completed", it is determined that all independent airflow chambers marked as impurity-to-be-removed areas have completed the depressurization-brief positive pressure backflushing operation sequence. Other methods can also be used in other embodiments, which are not limited here.
[0057] In practical implementation, if this is already completed, a gradient pressure relief control command is sent to the airflow valves corresponding to all independent airflow chambers, causing them to slowly connect to the atmosphere at a controlled rate until the negative pressure drop in each chamber reaches zero. This can be achieved as follows: the central processing unit simultaneously sends specific control signals to the airflow valves corresponding to all independent airflow chambers, such as proportional valves or high-speed switching valves, through its digital or analog output module. For proportional valves, this control signal is an analog voltage or current quantity, controlling the valve opening to slowly and linearly increase from a closed state to a fully open state. The closed state maintains negative pressure, and the fully open state connects to the atmosphere. This process lasts for a set total time. For example, 2 seconds; for high-speed switching valves, pulse width modulation control is used to rapidly open and close the valve with a fixed high-frequency cycle, and the duty cycle of the valve opening time within each cycle is controlled to linearly increase from 0% to 100%, achieving a smooth increase in flow rate through this averaging effect; during this process, the negative pressure in all the independent airflow chambers is slowly and uniformly released through the pores of the porous negative pressure adsorption belt, and the pressure value smoothly decreases from the working negative pressure to the normal pressure; it should be noted that this process can be closed-loop controlled by the feedback of the pressure sensor to ensure that the pressure relief curve conforms to the preset gradient. Other methods can also be used in other embodiments, which are not limited here.
[0058] In practice, after the negative pressure is completely released, the collection of chili peppers that have gently fallen from the porous negative pressure adsorption belt can be achieved as follows: Once the pressure sensors in all independent airflow chambers confirm that the pressure within the chambers has stabilized at normal atmospheric pressure, the negative pressure is considered completely released. At this point, the chili peppers, having lost their adsorption force, gently fall vertically downwards under the influence of gravity. A fruit collection conveyor belt is installed directly below the porous negative pressure adsorption belt, its running direction being perpendicular to or in the same direction as the porous negative pressure adsorption belt. The fallen chili peppers fall directly onto the fruit collection conveyor belt. The fruit collecting conveyor belt typically uses a soft-surfaced material, such as rubber or canvas, and has cushioning supports underneath to further reduce the impact of the fruit falling. The fruit collecting conveyor belt transports the cleaned pepper fruits to the fruit collection box at the rear of the pepper plant harvester for collection. At the same time, a buffer guide plate can be set in the drop space between the end of the porous negative pressure adsorption belt and the fruit collecting conveyor belt to guide the fruit to fall more accurately into the center of the conveyor belt. Other methods can also be used in other embodiments, which are not limited here.
[0059] It should be noted that the above steps enable flexible and controlled release of the fruit's adsorption force, avoid impact and collision caused by sudden pressure changes in the fruit through gradient pressure relief, and complete the collection in a buffered manner, thereby significantly reducing mechanical damage to the delicate pepper fruit during harvesting and ensuring the final commercial quality of the fruit.
[0060] In another aspect, in some embodiments, this application provides a pneumatic chili plant harvester, which includes a fruit separation unit, see reference. Figure 4 The figure is a schematic diagram of the structure of a fruit separation unit according to some embodiments of this application. The fruit separation unit includes: a collection module 401, a processing module 402, and an execution module 403, which are described below: The acquisition module 401 in this application is mainly used to acquire image information of the mixture on the surface of the porous negative pressure adsorption belt inside the chili plant harvester. Processing module 402, in this application, is mainly used to perform visual recognition of chili pepper fruits and light impurities on the image information, to obtain the outlines of chili pepper fruits and light impurities and their corresponding two-dimensional positions. At the same time, by monitoring the real-time negative pressure of the corresponding independent airflow chambers through the air pressure sensors set in each independent airflow chamber below the adsorption zone, the negative pressure stability of each independent airflow chamber is obtained. The processing module 402 described in this application is further used to mark the area that is visually identified as a chili fruit and has a corresponding real-time stable negative pressure as a fruit retention area, and the area that is visually identified as a light impurity or has an abnormal real-time negative pressure fluctuation as an impurity to be removed area, based on the outline of the chili fruit and the light impurity and the corresponding two-dimensional position, as well as the negative pressure stability of each independent airflow chamber. The processing module 402 described in this application is also used to perform a depressurization-brief positive pressure backflushing operation sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities and maintain the current adsorption state of the independent airflow chamber corresponding to the fruit holding area. The execution module 403 in this application is mainly used to perform controlled gradient depressurization on all independent airflow chambers after the impurities are removed in the impurity removal area, until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can fall off smoothly and be collected.
[0061] Each module in the aforementioned fruit separation unit can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.
[0062] In another embodiment, this application provides a computer device, which may be a server, and its internal structure diagram may be as follows. Figure 5 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores fruit separation data from a pneumatic chili harvester. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a fruit separation method for a pneumatic chili harvester.
[0063] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0064] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above embodiment of the fruit separation method of the pneumatic chili plant harvester.
[0065] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described embodiment of the fruit separation method for a pneumatic chili plant harvester.
[0066] In one embodiment, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps in the above-described embodiment of the fruit separation method for a pneumatic chili plant harvester.
[0067] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.
[0068] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0069] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for separating fruits in a pneumatic chili harvester, characterized in that, Includes the following steps: Image information of the mixture spread on the porous negative pressure adsorption belt inside the chili plant harvester was collected; Visual recognition of chili pepper fruits and light impurities is performed on the image information to obtain the outlines of chili pepper fruits and light impurities and their corresponding two-dimensional positions. At the same time, the real-time negative pressure of the corresponding independent airflow chamber is monitored by the air pressure sensor set in each independent airflow chamber below the adsorption zone to obtain the negative pressure stability of each independent airflow chamber. Based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, as well as the negative pressure stability of each independent airflow chamber, the areas that are visually identified as chili pepper fruits and have corresponding real-time stable negative pressure are marked as fruit retention areas, and the areas that are visually identified as light impurities or have corresponding real-time abnormal negative pressure fluctuations are marked as impurity removal areas. A depressurization-brief positive pressure backflushing sequence is performed on the independent airflow chamber corresponding to the impurity removal zone to remove light impurities, while the current adsorption state is maintained in the independent airflow chamber corresponding to the fruit holding zone. After impurities are removed from the impurity removal area, all independent airflow chambers are subjected to controlled gradient depressurization until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected.
2. The method as described in claim 1, characterized in that, Visual recognition of chili pepper fruits and light impurities from the image information to obtain the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities specifically includes: The image information is preprocessed to enhance its features; The preprocessed image information is input into a pre-trained semantic segmentation model, which outputs pixel-level classification results. Based on the pixel-level classification results, the outlines of chili pepper fruits and light impurities were extracted; The image coordinates corresponding to the outlines of the chili fruit and light impurities are converted into two-dimensional position coordinates relative to the plane of the porous negative pressure adsorption zone.
3. The method as described in claim 1, characterized in that, By monitoring the real-time negative pressure of each independent airflow chamber through pressure sensors installed below the adsorption zone, the specific negative pressure stability of each independent airflow chamber is obtained, including: The real-time negative pressure value sequence of each independent airflow chamber is obtained by the air pressure sensor set in each independent airflow chamber below the adsorption zone at a fixed sampling frequency. A fixed-time sequence of real-time negative pressure values is extracted for each independent airflow chamber as an analysis window; The standard deviation of the real-time negative pressure value sequence within each analysis window is used as the quantification value of the negative pressure stability of the corresponding independent airflow chamber.
4. The method as described in claim 1, characterized in that, Based on the outlines and corresponding two-dimensional positions of the chili pepper fruits and light impurities, as well as the negative pressure stability of each independent airflow chamber, the areas visually identified as chili pepper fruits and corresponding to stable real-time negative pressure are marked as fruit retention areas, and the areas visually identified as light impurities or corresponding to abnormal fluctuations in real-time negative pressure are marked as impurity removal areas. Specifically, these include: Based on the two-dimensional position coordinates corresponding to the outlines of the chili fruit and the light impurities, the independent airflow chambers covered by each outline are determined. For the outline of each chili pepper fruit, query the negative pressure stability of all the individual airflow chambers it covers; If the negative pressure stability of all independent airflow chambers is lower than the preset stability threshold, the outline area of the chili fruit is marked as the fruit retention area; otherwise, the outline area of the chili fruit is marked as the impurity removal area. For each light impurity's outline, the outline area of the light impurity is directly marked as the impurity to be removed area.
5. The method as described in claim 1, characterized in that, Perform a depressurization-brief positive pressure backflushing sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities. Specifically, this includes: Obtain the corresponding independent airflow chamber number based on the area information marked as the impurity removal zone; Send a control command to the airflow valve corresponding to the independent airflow chamber to switch it to the depressurization state in order to release the negative pressure adsorption; After confirming that the corresponding independent airflow chamber has been depressurized, a control command is sent to the airflow valve to switch it to connect to a positive pressure air source and maintain this connection for a short, set time to generate an upward pulse airflow that blows away light impurities.
6. The method as described in claim 1, characterized in that, Maintaining the current adsorption state in the independent airflow chambers corresponding to the fruit retention zone specifically includes: The corresponding independent airflow chamber number is obtained based on the area information marked as the fruit preservation area; Send control commands to the airflow valves corresponding to the independent airflow chambers to keep them connected to the negative pressure source.
7. The method as described in claim 1, characterized in that, After impurity removal is completed in the impurity removal area, a controlled gradient depressurization is performed on all independent airflow chambers until the negative pressure is completely removed, so that the pepper fruits on the adsorption belt can be gently detached and collected. Specifically, this includes: Determine whether all independent airflow chambers marked as areas to be cleaned of impurities have completed the depressurization-brief positive pressure backflush sequence; If completed, a gradient pressure relief control command is sent to the airflow valves corresponding to all independent airflow chambers, so that they are slowly connected to the atmosphere at a controlled rate until the negative pressure in each chamber drops to zero. After the negative pressure is completely released, collect the chili peppers that have gently fallen off the porous negative pressure adsorption belt.
8. A pneumatic chili plant harvester, comprising a fruit separation unit, characterized in that, The fruit separation unit includes: The acquisition module is used to acquire image information of the mixture spread on the porous negative pressure adsorption belt inside the chili plant harvester; The processing module is used to visually identify the chili pepper fruit and light impurities in the image information, obtain the outline of the chili pepper fruit and light impurities and their corresponding two-dimensional positions. At the same time, the real-time negative pressure of the corresponding independent airflow chamber is monitored by the air pressure sensor set in each independent airflow chamber below the adsorption zone, and the negative pressure stability of each independent airflow chamber is obtained. The processing module is also used to mark the area that is visually identified as a chili fruit and has a stable real-time negative pressure as the fruit retention area, and the area that is visually identified as a light impurity or has an abnormal real-time negative pressure fluctuation as the impurity removal area, based on the outline and corresponding two-dimensional position of the chili fruit and light impurities, as well as the negative pressure stability of each independent airflow chamber. The processing module is also used to perform a depressurization-brief positive pressure backflushing operation sequence on the independent airflow chamber corresponding to the impurity removal area to remove light impurities and maintain the current adsorption state of the independent airflow chamber corresponding to the fruit holding area. The execution module is used to perform controlled gradient depressurization of all independent airflow chambers after the impurities are removed in the impurity removal area, until the negative pressure is completely released, so that the pepper fruits on the adsorption belt can be gently detached and collected.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the fruit separation method of the pneumatic chili plant harvester according to any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the fruit separation method of the pneumatic chili plant harvester as described in any one of claims 1 to 7.