Fruit and vegetable harvesting method, device, electronic device and storage medium

Abstract

This invention provides a method, apparatus, electronic device, and storage medium for harvesting fruits and vegetables. The method is applied to a fruit and vegetable harvesting machine and includes: acquiring a regional image of the harvesting area; determining the center line of the planting row of fruits and vegetables to be harvested in the harvesting area based on the regional image; determining the harvesting area corresponding to the center line of the planting row in the harvesting area; predicting harvesting points based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area to obtain the target harvesting points of the fruits and vegetables to be harvested; generating a continuous dynamic harvesting trajectory based on the target harvesting points; and harvesting the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory. This achieves unmanned and precise control, effectively avoiding the efficiency reduction and high damage rate problems caused by frequent vertical lifting and blind operation of traditional machinery. While significantly improving the efficiency of automated fruit and vegetable harvesting, it reduces the harvesting damage rate to an extremely low level from the root cause.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, electronic device, and storage medium for harvesting fruits and vegetables. Background Technology

[0002] With the continuous and rapid development of modern agriculture, realizing unmanned and intelligent crop harvesting has become an important research direction in the field of smart agriculture.

[0003] However, traditional fruit and vegetable harvesting, such as the harvesting of root vegetables like carrots and radishes, relies heavily on conventional mechanical control equipment. In actual harvesting, manual operation and the operator's experience are still essential for aligning the machine with the rows and finding the optimal harvesting points. Each production area faces different soil textures, varietal variations, and uneven growth. Blindly operating the machine without scientific basis not only increases labor intensity and leads to low harvesting efficiency but also results in extremely high harvesting damage rates due to the mismatch between the machinery's structure and the crop's planting method. Summary of the Invention

[0004] This invention provides a method, apparatus, electronic device, and storage medium for harvesting fruits and vegetables, which solves the problems of inefficiency and high breakage caused by blind operation or human experience error in traditional machinery, significantly improving harvesting efficiency and reducing breakage rate.

[0005] This invention provides a method for harvesting fruits and vegetables, applied to a fruit and vegetable harvester, comprising: Acquire a regional image of the harvesting area, and determine the center line of the planting row of the fruits and vegetables to be harvested in the harvesting area based on the regional image; Determine the harvesting area corresponding to the center line of the planting row in the harvesting area, and predict the harvesting point based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area to obtain the target harvesting point of the fruits and vegetables to be harvested. A continuous dynamic harvesting trajectory is generated based on the target harvesting points, and the fruits and vegetables to be harvested in each harvesting area are harvested based on the dynamic harvesting trajectory.

[0006] According to a fruit and vegetable harvesting method provided by the present invention, the spatial planting distribution characteristics include planting location, growth direction and main axis direction; The harvesting point prediction is performed based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area to obtain the target harvesting points of the fruits and vegetables to be harvested; a continuous dynamic harvesting trajectory is generated based on the target harvesting points, including: Determine the spatial distribution image of the planting area for each harvesting zone; Color space conversion and feature extraction are performed on the planting spatial distribution image to obtain the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested in each harvesting area; Based on the planting location, growth direction, and main axis direction of the fruits and vegetables to be harvested, the harvesting point is predicted to obtain the target harvesting point of the fruits and vegetables to be harvested. Multiple target harvesting points obtained from continuous prediction are subjected to fuzzy prediction to obtain a continuous wave-shaped harvesting curve, and the wave-shaped harvesting curve is used as the dynamic harvesting trajectory.

[0007] According to a fruit and vegetable harvesting method provided by the present invention, the spatial planting distribution characteristics further include the average size of the fruits and vegetables to be harvested; The harvesting of fruits and vegetables in each harvesting area based on the dynamic harvesting trajectory also includes the following steps beforehand: Based on the standard size and average size of the fruits and vegetables to be harvested, a benchmark resistance threshold for distinguishing geological types is determined. The real-time soil resistance experienced by the soil loosening mechanism of the fruit and vegetable harvester in each harvesting area is compared with the benchmark resistance threshold, and the vibration frequency of the soil loosening mechanism in each harvesting area is adjusted based on the comparison results.

[0008] According to a fruit and vegetable harvesting method provided by the present invention, the step of harvesting the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory further includes: The machine posture data of the fruit and vegetable harvester and the terrain data of each harvesting area are obtained, and the terrain slope is extracted from the terrain data. Based on the terrain slope, the terrain compensation parameters are determined. Extract the current pitch angle and pitch angle change rate from the fuselage attitude data, and determine the basic adjustment parameters based on the current pitch angle and pitch angle change rate; Generate attitude adjustment commands based on the basic adjustment parameters and the terrain compensation parameters; Based on the attitude adjustment command, the suspension system of the fruit and vegetable harvester is adjusted to keep the harvester balanced.

[0009] According to a fruit and vegetable harvesting method provided by the present invention, the method further includes harvesting the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory, and then further comprising: Determine the current speed of the fruit and vegetable harvester; Based on the current driving speed, the multi-stage conveying mechanism in the fruit and vegetable harvester is controlled to convey the harvested and separated fruits and vegetables at a speed matched in stages. Acquire images of the fruits and vegetables to be harvested during the transportation process; Extract the fruit and vegetable boundary features from the fruit and vegetable image, and determine the root-stem junction position of the fruit and vegetable to be harvested based on the fruit and vegetable boundary features; Based on the root-stem junction position, the cutting height of the cutting mechanism in the fruit and vegetable harvester is adjusted, and the adjusted cutting mechanism is controlled to cut off the stems and leaves of the fruits and vegetables to be harvested.

[0010] According to a method for harvesting fruits and vegetables provided by the present invention, the controlled and adjusted cutting mechanism removes the stems and leaves of the fruits and vegetables to be harvested, and then further includes: Determine the stacking height of fruits and vegetables on the multi-stage conveying mechanism, and the load current of the motor driving the cutting mechanism in the fruit and vegetable harvester; Transmission blockage is determined based on the height of the fruit and vegetable stack and the load current. In the event of a known transmission blockage, the height abrupt change characteristics of the fruit and vegetable stacking height and the periodic fluctuation characteristics of the load current are extracted. Based on the highly abrupt change characteristics and the periodic fluctuation characteristics, the blockage type is identified to obtain the target blockage type; Based on the target congestion type, a congestion handling strategy is recommended.

[0011] According to a fruit and vegetable harvesting method provided by the present invention, the end of the multi-stage conveying mechanism is provided with a baffle structure with an inclination angle. The baffle structure is used to intercept the fruits and vegetables to be harvested conveyed by the multi-stage conveying mechanism at the end position, so as to change the falling posture of the fruits and vegetables to be harvested.

[0012] The present invention also provides a fruit and vegetable harvesting device, applied to a fruit and vegetable harvesting machine, comprising: The identification unit is used to acquire a regional image of the harvesting area and determine the center line of the planting row of the fruits and vegetables to be harvested in the harvesting area based on the regional image. The prediction unit is used to determine the harvesting area corresponding to the center line of the planting row in the harvesting area, and to predict the harvesting point based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, so as to obtain the target harvesting point of the fruits and vegetables to be harvested. The harvesting unit is used to generate a continuous dynamic harvesting trajectory based on the target harvesting point, and to harvest the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the fruit and vegetable harvesting method as described above.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the fruit and vegetable harvesting method as described above.

[0015] The fruit and vegetable harvesting method, apparatus, electronic equipment, and storage medium provided by this invention determine the center line of the planting rows through regional images of the harvesting area, realizing intelligent and autonomous row alignment of agricultural machinery and eliminating reliance on human driving experience. By predicting the target harvesting point through spatial planting distribution characteristics and generating a continuous dynamic harvesting trajectory accordingly, the actuator can make flexible and smooth dynamic adjustments based on the actual growth, thereby achieving unmanned and precise control. This effectively avoids the efficiency reduction and high damage rate problems caused by frequent vertical lifting and blind operation of traditional machinery. While significantly improving the efficiency of automated fruit and vegetable harvesting, it reduces the harvesting damage rate to an extremely low level from the root cause. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the fruit and vegetable harvesting method provided by the present invention; Figure 2 This is an example diagram of the anti-falling baffle structure provided by the present invention; Figure 3 This is a schematic diagram of the structure of the fruit and vegetable harvesting device provided by the present invention; Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0019] In many regions, root vegetables such as carrots and radishes are widely cultivated, and mechanization has been achieved in land preparation, sowing, and field management. However, the mechanization level in harvesting is severely insufficient, and the high cost of manual harvesting seriously restricts the sustainable development of the industry. Currently, the harvesting machinery used in the market also faces difficulties in promotion due to problems such as mismatch between mechanical structure and planting methods, high harvesting damage rate, and low efficiency.

[0020] Currently, some research has attempted to use machine learning or model control for fruit and vegetable harvesting. However, these technologies still have significant shortcomings in practical applications. Specifically, most current harvesting equipment requires manual driving and operation, resulting in high labor intensity. Especially during the harvesting process, key operations such as locating planting ridges and pinpointing the harvesting points heavily rely on the operator's experience. Furthermore, due to the differences in varieties and uneven growth in each production area, existing harvesting equipment cannot dynamically adjust in real time according to the spatial distribution and growth of fruits and vegetables, thus severely impacting harvesting efficiency and increasing the damage rate of fruits and vegetables.

[0021] In response, this invention provides a fruit and vegetable harvesting method that aims to automatically align rows by identifying the center line of planting rows through regional image recognition, and dynamically predict the target harvesting point based on the spatial planting distribution characteristics of fruits and vegetables, thereby generating a continuous and smooth dynamic harvesting trajectory to guide precise mechanical operations. This overcomes the shortcomings of traditional harvesting, which relies on human experience and cannot be adjusted in real time according to the growth status, and achieves unmanned and precise control of fruit and vegetable harvesting, resulting in a significant improvement in harvesting quality, increased harvesting efficiency, and a substantial reduction in damage rate.

[0022] Figure 1 This is a flowchart illustrating the fruit and vegetable harvesting method provided by the present invention. This method is applied to a fruit and vegetable harvesting machine, and the specific executing entity can be a central controller, on-board computing unit, unmanned control system, etc., mounted on the harvesting machine. It is understood that in practical applications, the fruit and vegetable harvesting machine can be any agricultural machinery used for harvesting root vegetables. For example, it can be an intelligent harvesting equipment modified from a traditional tracked fruit and vegetable (such as carrot) harvester by adding a positioning system, angle sensor, lidar, vision sensor, attitude perception sensor, electro-hydraulic unit, electronic handle, etc. Figure 1 As shown, the method includes: Step 110: Obtain a regional image of the harvesting area, and determine the center line of the planting rows of the fruits and vegetables to be harvested in the harvesting area based on the regional image; Step 120: Determine the harvesting area corresponding to the center line of the planting row in the harvesting area. Based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, predict the harvesting points to obtain the target harvesting points of the fruits and vegetables to be harvested. Step 130: Generate a continuous dynamic harvesting trajectory based on the target harvesting point, and harvest the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory.

[0023] Understandably, the primary task during fruit and vegetable harvesting is to ensure the harvester accurately enters and aligns with the furrows or rows where the harvested produce is located. Therefore, the first step is to determine the harvesting area and acquire an image of that area. The harvesting area here refers to the farmland or plot where the harvester will currently be harvesting the produce. The area image can be obtained through real-time visual image data captured by image acquisition devices mounted on the front of the harvester or at specific locations on its body, such as RGB (Red, Green, Blue) cameras, multispectral cameras, and vision sensors, capturing images of the ground, the planted produce, and the soil environment within the harvesting area.

[0024] After acquiring the regional image, image processing and analysis can be performed to extract the planting distribution pattern of the fruits and vegetables to be harvested on the ground, thereby determining the center line of the planting rows in the harvesting area. Here, the center line of the planting rows is a virtual directional reference line representing the linear or strip-shaped concentrated distribution of the fruits and vegetables to be harvested in the farmland or plot.

[0025] Specifically, since fruits and vegetables to be harvested, such as carrots, are usually planted in rows on ridges, visual algorithms, such as edge extraction algorithms and line fitting algorithms, can be used to identify the edge contour of the planting row containing the fruits and vegetables to be harvested. This allows for the calculation of the geometric centerline representing that planting row, i.e., the planting row centerline. Using this planting row centerline, the fruit and vegetable harvester can achieve intelligent automatic row alignment. By adjusting its steering mechanism, it can ensure that its operating direction remains consistent with the planting row centerline, thus replacing the traditional row alignment operation that relies on human experience and visual observation. This lays an accurate positional foundation for subsequent precise harvesting.

[0026] During the alignment process of the fruit and vegetable harvester along the center line of the planting row, it is necessary to further focus on the specific harvesting target, that is, to determine the specific harvesting area corresponding to the center line of the planting row. Here, the harvesting area refers to a local block around the center line of the planting row that is currently or will soon enter the harvesting operation range of the fruit and vegetable harvester. Since the fruits and vegetables to be harvested within the same planting row exhibit individual or cluster differences in growth, tilting posture, etc., this embodiment of the invention also requires extracting the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, such as the planting location, growth direction, and distribution pattern of branches, leaves, and roots.

[0027] After extracting the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, harvesting points can be predicted to calculate the optimal force application position for harvesting actions such as clamping, lifting, and shaking, i.e., the target harvesting point of the fruits and vegetables to be harvested. For example, for carrots, if a fixed mechanical height is used for clamping, problems such as missed harvesting or cutting may easily occur due to the varying heights of the carrots. In view of this, this embodiment of the invention introduces harvesting point prediction based on spatial planting distribution. The target harvesting point predicted in this way is the optimal clamping or cutting coordinate for the fruits and vegetables to be harvested in the corresponding harvesting area, such as precisely locking onto a node at a certain set distance above the root. This dynamic prediction of harvesting points completely changes the traditional rigid and fixed mechanical harvesting mode, realizing flexible decision-making that moves with the growth.

[0028] After obtaining precise target harvesting points, the fruit and vegetable harvester needs to control its end effectors, such as the harvesting arm and clamping device, to perform the harvesting action. However, to avoid abrupt and jarring vertical movements or large-scale frequent jumps between target harvesting points at different heights or positions, this embodiment of the invention also generates a continuous dynamic harvesting trajectory based on multiple continuously predicted target harvesting points. This dynamic harvesting trajectory is a smooth transitional motion path in space, and the fruit and vegetable harvester's execution architecture will follow this path servo-driven.

[0029] Specifically, based on the dynamic harvesting trajectory, the unmanned control system on the fruit and vegetable harvester can drive the harvesting arm and other actuators to smoothly and adaptively harvest the fruits and vegetables to be harvested in each harvesting area. This dynamic operation mode based on continuous trajectory makes the lifting or angle adjustment of the harvesting arm present a smooth and streamlined transition, rather than the traditional abrupt straight rise and fall.

[0030] The fruit and vegetable harvesting method provided by this invention determines the center line of the planting rows through regional images of the harvesting area, realizing intelligent and autonomous row alignment of agricultural machinery and eliminating reliance on human driving experience. By predicting the target harvesting point through spatial planting distribution characteristics, and generating a continuous dynamic harvesting trajectory accordingly, the actuator can make flexible and smooth dynamic adjustments based on the actual growth, thereby achieving unmanned and precise control. This effectively avoids the problems of reduced efficiency and high damage rate caused by frequent vertical lifting and blind operation of traditional machinery. While significantly improving the efficiency of automated fruit and vegetable harvesting, it reduces the harvesting damage rate to an extremely low level from the root cause.

[0031] Based on the above embodiments, the process of determining the center line of the planting row of fruits and vegetables to be harvested in the harvesting area may specifically include: First, visual algorithms, such as the Sobel operator (gradient in the X direction), can be used to extract edges from a region image or an image after grayscale processing, resulting in a binary edge map. That is, by calculating the gradient magnitude and direction of image pixels, the visual edges of the planting rows containing the fruits and vegetables to be harvested can be accurately extracted, thus obtaining a binary edge image.

[0032] Then, the planting row centerline can be determined based on each foreground point in the extracted binary edge image; specifically, this can be achieved using the Hough transform (accumulator voting method). That is, within a set discrete angle range for the gradient direction and a set step size, such as 0° to 180° and a step size of 1°, the coordinate system of each foreground point in the binary edge image is transformed to calculate the corresponding polar radius parameter. Subsequently, the corresponding position is found in the accumulator array constructed based on the polar radius parameter and the gradient direction for voting (i.e., the value at that position is incremented by 1).

[0033] After traversing all foreground points in the binary edge image, local peaks can be found in the accumulator array. Each peak corresponds to a possible straight line in the image space, and this straight line fitted by voting is the center line of the planting row.

[0034] It should be noted that the key parameters involved in this centerline detection process include the range of discrete angle values, the step size, and the minimum vote threshold for detecting local peaks. To further optimize the performance of the detection algorithm, in practical applications, prior knowledge, such as the approximate direction of planting rows in farmland, can be used to pre-constrain the angle search range. This not only filters out invalid interfering lines but also effectively improves the computational efficiency and accuracy of the detection.

[0035] Based on the above embodiments, the spatial planting distribution characteristics include planting location, growth direction, and main axis direction; step 120 includes: Determine the spatial distribution image of planting areas in each harvesting zone; Color space conversion and feature extraction were performed on the planting spatial distribution image to obtain the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested in each harvesting area; Based on the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested, the harvesting point is predicted to obtain the target harvesting point of the fruits and vegetables to be harvested. In step 130, a continuous dynamic harvesting trajectory is generated based on the target harvesting point, including... Fuzzy prediction is performed on multiple target harvesting points obtained from continuous prediction to obtain a continuous wavy harvesting curve, which is then used as the dynamic harvesting trajectory.

[0036] Specifically, to address the challenges of accurately locating harvesting points in complex farmland environments where individual fruits and vegetables exhibit varying growth postures, and the inefficiency and high damage caused by the frequent vertical lifting and lowering of traditional harvesting arms, the harvesting point prediction process in this embodiment of the invention may include: As the fruit and vegetable harvester operates along its route, it collects images of the planting spatial distribution in each harvesting area. In practical applications, the harvester can use vision sensors, such as RGB-D (Depth) cameras, mounted at a lower position on the harvesting arm to collect real-time visual and depth information of the entire row ahead, such as two rows of fruits and vegetables to be harvested. This allows it to obtain planting spatial distribution images that reflect the overall three-dimensional spatial distribution of the fruits and vegetables to be harvested in that area.

[0037] Because natural light in farmland environments varies drastically and can easily interfere with visual recognition, this embodiment of the invention also performs color space conversion and feature extraction after acquiring the planting spatial distribution image. Specifically, the planting spatial distribution image can be converted from the RGB color space to the HSV (Hue, Saturation, Value) or HSL (Hue, Saturation, Lightness) color space. Since the H (hue) channel in the HSV color space is not sensitive to changes in light, this characteristic can be used to accurately distinguish and segment representative fruits and vegetables of ready-to-be-harvested color (such as the orange of carrots) from the background environment (such as the brown of the soil and the green of the stems and leaves). After color segmentation, image processing algorithms, such as the minimum bounding rectangle method, can be further applied to extract features from the segmented fruit and vegetable outlines. After this processing, the center position of the circumscribed rectangle can represent the planting location of the fruits and vegetables to be harvested, while the direction of the long side and the angle of inclination of the rectangle approximately indicate the growth direction and main axis direction of the fruits and vegetables to be harvested above the ground, thereby accurately obtaining the specific growth posture (spatial planting distribution characteristics) of the fruits and vegetables to be harvested in each harvesting area.

[0038] After this, the harvesting point can be predicted based on the obtained planting location, growth direction, and main axis direction, thus obtaining the target harvesting point for the fruits and vegetables to be harvested. Here, based on the planting location, growth direction, and main axis direction, the optimal force application position for harvesting actions such as clamping, lifting, and shaking can be accurately calculated using the proportional method, which is the target harvesting point.

[0039] Furthermore, to make the harvesting arm's movements more flexible and efficient, this embodiment of the invention also requires fuzzy prediction of multiple continuously predicted target harvesting points to fit a continuous wave-shaped harvesting curve, i.e., a dynamic harvesting trajectory. Specifically, if the fruit and vegetable harvester only performs discrete mechanical tracking on isolated target harvesting points, it will inevitably lead to sudden, large-amplitude, and frequent vertical rises and falls of the harvesting arm at different heights. This not only severely slows down the harvesting efficiency but also easily tears the fruits and vegetables, significantly increasing the damage rate. In view of this, this embodiment of the invention introduces a fuzzy algorithm to smoothly fit and fuzzily predict multiple continuously predicted target harvesting points in the direction of travel to form a dynamic, continuous, undulating wave-shaped harvesting curve, thereby driving the harvesting arm to perform streamlined dynamic rises and falls and angle alignment along this curve.

[0040] In this embodiment of the invention, by performing color space conversion and feature extraction on the planting space distribution image, the interference of complex lighting and soil background in the field is effectively eliminated, and the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested are accurately obtained, providing reliable data support for determining the optimal harvesting point. More importantly, by generating a wave-shaped harvesting curve through fuzzy prediction of multiple target harvesting points obtained by continuous prediction, the streamlined dynamic tracking of the harvesting arm is realized, which completely changes the traditional rigid, vertical and frequent lifting and lowering operation mode of the harvesting arm. This not only greatly improves the continuity of continuous operation and harvesting efficiency, but also greatly reduces the damage rate of fruits and vegetables caused by physical pulling from the mechanical execution level.

[0041] Based on the above embodiments, the spatial planting distribution characteristics also include the average size of fruits and vegetables to be harvested; Based on dynamic harvesting trajectories, the fruits and vegetables awaiting harvest in each harvesting area are harvested. This previously included: Based on the standard and average size of fruits and vegetables to be harvested, a benchmark resistance threshold for distinguishing geological types was determined. The real-time soil resistance experienced by the soil loosening mechanism of the fruit and vegetable harvester in each harvesting area is compared with the benchmark resistance threshold, and the vibration frequency of the soil loosening mechanism in each harvesting area is adjusted based on the comparison results.

[0042] Specifically, in the process of harvesting fruits and vegetables, in addition to precise positioning and dynamic trajectory control of the harvesting arms and other actuators, the soil loosening operation before harvesting is equally important for reducing the damage rate of fruits and vegetables. If the working state of the soil loosening mechanism cannot adapt to the soil texture, it is very easy to miss or cut the roots of fruits and vegetables in hard soil, resulting in a high damage rate.

[0043] Based on this, before harvesting the fruits and vegetables to be harvested in each harvesting area according to the dynamic harvesting trajectory, adaptive soil loosening control can also be carried out. This process is as follows: First, a baseline resistance threshold for differentiating geological types needs to be determined based on the standard and average sizes of the fruits and vegetables to be harvested. It's understandable that fruits and vegetables of different growth stages and sizes have different perceptions and requirements regarding soil compaction. The average size, such as the average length of radishes, can be determined through analysis of planting spatial distribution images, while the standard size is the normal reference length for that variety of fruit or vegetable to be harvested, measured under standard experimental conditions. In actual calculations, a fixed threshold is not used; instead, a dynamic algorithm is introduced, based on a predetermined baseline threshold. Based on, combined with standard dimensions and average size and the set calibration coefficient This allows us to determine the specific benchmark resistance threshold for the corresponding harvesting area. .

[0044] Here, the reference resistance threshold can be calculated using the following formula: The benchmark resistance threshold serves as a dynamic scale for measuring resistance, and can accurately define the soil texture type of the corresponding harvesting area, such as soft sandy soil or hard clay.

[0045] Next, after obtaining the baseline resistance threshold, the real-time soil resistance experienced by the loosening mechanism of the fruit and vegetable harvester in each harvesting area can be compared with the corresponding baseline resistance threshold. In actual operation, if a force sensor is added to the top of the loosening shovel at the bottom of the fruit and vegetable harvester, the force sensor will continuously collect the resistance experienced by the loosening mechanism when it cuts into the soil as the harvester moves forward, i.e., the real-time soil resistance. This allows for the comparison of real-time soil resistance with the benchmark resistance threshold for the corresponding harvesting area.

[0046] Furthermore, if the real-time soil resistance is less than the corresponding benchmark resistance threshold, i.e. If the soil in the corresponding harvesting area is determined to have a high sand content or a soft texture, the soil loosening mechanism will be controlled to operate at a lower vibration frequency in that harvesting area.

[0047] Conversely, if the real-time soil resistance is greater than or equal to the corresponding benchmark resistance threshold, i.e. If the soil texture in the corresponding harvesting area is too hard, the soil loosening mechanism should be switched to a higher vibration frequency when operating in that harvesting area.

[0048] It is worth noting that while adjusting the vibration frequency, controlling the soil loosening depth also has a significant impact on the breakage rate. Loosening the soil too deeply not only hinders the smooth pulling of fruits and vegetables, easily leading to missed pulls, but also increases the mechanical running resistance, making it difficult for the machine to move and turn; while loosening the soil too shallowly can easily cut off the roots of fruits and vegetables, resulting in a high breakage rate. Therefore, in this embodiment of the invention, the soil loosening depth is precisely controlled according to the average size of the fruits and vegetables to be harvested, so as to completely change the original extensive mode that relied on the operator's visual estimation and human experience.

[0049] Specifically, in this embodiment of the invention, an ultrasonic sensor and an angle sensor are installed at the top of the soil loosening mechanism. During operation, the ultrasonic sensor detects and determines the current soil loosening depth, while the angle sensor identifies the angle of inclination between the harvesting arm and the ground. For example, for carrot harvesting, when the optimal clamping point is determined to be 15cm above the root, the optimal soil loosening depth is assumed to be... cm, the angle between the harvesting arm and the ground is The sensor is installed 30cm away from the optimal clamping / removing point, and this can be determined according to... Dynamic calculations are performed, and the current soil loosening depth is adjusted based on the obtained optimal soil loosening depth to keep the soil loosening depth and clamping height locked at the optimal state at all times.

[0050] In addition, the force sensors installed on the soil loosening mechanism not only collect real-time soil resistance to help determine soil texture and adjust vibration frequency, but also issue a safety alarm when encountering hard objects such as underground rocks that cause an abnormal surge in resistance, in order to protect the fruit and vegetable harvester from damage.

[0051] In this embodiment of the invention, the average size of fruits and vegetables is integrated with the soil stress conditions to construct a soil texture determination mechanism based on dynamic thresholds. By comparing real-time soil resistance and adaptively adjusting the vibration frequency of the loosening mechanism, the traditional harvesting equipment relies on manual experience to blindly loosen the soil. This not only effectively overcomes the defects of machine movement and turning difficulties in complex geological environments, but also significantly reduces the root breakage and damage of fruits and vegetables caused by excessive soil resistance or improper loosening from the execution source level. This further improves the closed loop of low-loss harvesting process and significantly enhances the adaptability of agricultural machinery in variable soil environments.

[0052] Based on the above embodiments, the harvesting of fruits and vegetables in each harvesting area is carried out based on the dynamic harvesting trajectory, and the preceding steps also include: Acquire the machine attitude data of the fruit and vegetable harvester, as well as the terrain data of each harvesting area; The terrain slope is extracted from the terrain data, and the terrain compensation parameters are determined based on the terrain slope. Extract the current pitch angle and pitch angle change rate from the fuselage attitude data, and determine the basic adjustment parameters based on the current pitch angle and pitch angle change rate; Generate attitude adjustment commands based on basic adjustment parameters and terrain compensation parameters; Based on attitude adjustment commands, the suspension system of the fruit and vegetable harvester is adjusted to keep the harvester balanced.

[0053] Specifically, during the process of the fruit and vegetable harvester traveling along the planned path and the harvesting arm harvesting, the unevenness, ridging, and varying soil texture of the farmland often cause the harvester to tilt. This tilt not only severely interferes with the accurate positioning of the visual sensors but also directly changes the downward angle of the harvesting arm, causing the originally predicted target harvesting point to deviate during harvesting, leading to problems such as missed harvesting or cutting of fruits and vegetables, resulting in high damage. Therefore, in this embodiment of the invention, the posture of the fruit and vegetable harvester can be adjusted before actual harvesting.

[0054] In detail, the first step is to obtain the machine's attitude data and the terrain data of each harvesting area. During actual operation, the harvester uses sensors such as LiDAR to scan the harvesting area ahead in real time, thereby obtaining terrain data containing three-dimensional spatial features. At the same time, it uses attitude sensing sensors on the machine, such as inertial measurement units (IMUs), to obtain the current machine attitude data in real time.

[0055] Furthermore, to achieve proactive terrain adaptation, this embodiment of the invention also utilizes environmental point cloud information from lidar to identify in advance whether there are slopes, potholes, etc., ahead, in order to extract the terrain slope representing the undulation of the terrain from the terrain data. Once a slope is detected ahead, predictive control is performed, that is, the terrain slope is multiplied by a preset feedforward gain coefficient to calculate in advance a terrain compensation parameter for actively responding to changes in terrain.

[0056] Meanwhile, in this embodiment of the invention, the current pitch angle and pitch angle change rate are extracted in real time from the fuselage attitude data, and the basic adjustment parameters are determined based on these two. Here, the current pitch angle represents the degree of tilt of the fruit and vegetable harvester in the longitudinal direction, while the pitch angle change rate reflects the speed and dynamic trend of the vehicle's tilt. Specifically, the tilt error can be obtained by subtracting the set target pitch angle (usually set to 0 degrees of absolute horizontal) from the current pitch angle using the PID (Proportion Integration Differentiation) controller on the fruit and vegetable harvester. Then, based on this tilt error and the pitch angle change rate, a joint calculation is performed to obtain the basic adjustment parameters used to correct the current vehicle attitude deviation.

[0057] After obtaining the adjustment bases for both feedforward and feedback, attitude adjustment commands can be generated based on these two. Specifically, the basic adjustment parameters for responding to the current vehicle body tilt are superimposed and fused with the terrain compensation parameters for predicting the terrain ahead, thus obtaining the attitude adjustment command. Furthermore, it should be noted that to prevent excessively abrupt adjustments that could damage the mechanical structure, this embodiment of the invention also incorporates amplitude limiting protection to ensure that the superimposed signal does not exceed a set maximum hydraulic pressure threshold, ultimately generating a safe and accurate attitude adjustment command.

[0058] Finally, the suspension system of the fruit and vegetable harvester can be adjusted according to attitude adjustment commands to maintain the harvester's balance. That is, the unmanned control system of the harvester will control the electro-hydraulic valves on the chassis in real time according to attitude adjustment commands to perform differential extension and retraction compensation on the suspension system, such as the hydraulic cylinders on both sides. Through this continuous active suspension repair, the harvester can maintain its balance regardless of the undulating terrain of the farmland.

[0059] In this embodiment of the invention, based on terrain data and fuselage attitude data, an active vehicle balance control mechanism combining PID feedback and terrain feedforward pre-compensation is introduced. This effectively overcomes the vehicle tilting problem caused by complex farmland terrain and eliminates the harvesting point offset caused by unstable vehicle attitude at the physical execution level. This not only ensures high-precision alignment between visual positioning and mechanical harvesting but also avoids hard breakage of fruits and vegetables caused by angle deviation of the harvesting arm, providing a stable and reliable guarantee for subsequent low-loss dynamic harvesting.

[0060] Based on the above embodiments, and based on the dynamic harvesting trajectory, the fruits and vegetables to be harvested in each harvesting area are harvested, followed by: Determine the current speed of the fruit and vegetable harvester; Based on the current driving speed, control the multi-stage conveying mechanism in the fruit and vegetable harvester to convey the harvested and detached fruits and vegetables at a speed matched in stages. Acquire images of fruits and vegetables during the transportation process; Extract the boundary features of fruits and vegetables from the fruit and vegetable images, and determine the root-stem junction of the fruits and vegetables to be harvested based on the boundary features; Based on the root-stem junction, adjust the cutting height of the cutting mechanism in the fruit and vegetable harvester, and control the adjusted cutting mechanism to cut off the stems and leaves of the fruits and vegetables to be harvested.

[0061] Specifically, after the fruit and vegetable harvester successfully harvests the fruits and vegetables from the field, the key to determining the final breakage rate lies in how to smoothly transport the harvested vegetables backward and accurately remove the stems and leaves. Traditional harvesting equipment often causes congestion and pulling of fruits and vegetables at this stage due to mismatched conveyor belt speeds, and relies heavily on physical baffles to forcibly align them during the top-cutting (i.e., stem and leaf removal) stage. This not only easily breaks the fruits and vegetables but also results in uneven cuts. In view of this, in this embodiment of the invention, a collaborative control mechanism of dynamic tiered conveying and visually precise top-cutting is introduced after harvesting.

[0062] In detail, during the transportation phase after successful harvesting, it is necessary to determine the current travel speed of the fruit and vegetable harvester. In actual operation, the current travel speed refers to the real-time physical linear velocity of the harvester chassis as it moves through the farmland. This speed will dynamically change with the terrain and the set operating mode. This speed can be accurately obtained in real time through the chassis encoder or positioning system.

[0063] Once the current travel speed S1 is determined, the operating speed of the multi-stage conveyor mechanism in the fruit and vegetable harvester can be determined accordingly, thereby controlling the multi-stage conveyor mechanism to transport the harvested fruits and vegetables at the corresponding operating speed. Here, the multi-stage conveyor mechanism typically includes multiple independent transmission modules such as a clamping conveyor belt responsible for clamping and transporting upwards, and a tassel-removing conveyor belt responsible for discharging waste stems and leaves backwards. To prevent the harvested fruits and vegetables from congesting and piling up at the junction of the conveyor belts or being pulled roughly due to excessive speed differences, a strict speed matching linkage relationship is constructed in this embodiment of the invention. For example, the operating speed S2 of the clamping conveyor belt and the operating speed S3 of the tassel-removing conveyor belt are set to satisfy a specific proportional relationship (such as S3=1.5S2=1.1S1).

[0064] Through control algorithms, multi-stage conveying mechanisms can be driven to operate at speeds that match each other in stages. This step-by-step acceleration conveying strategy can effectively ensure that fruits and vegetables that have just been removed from the soil are conveyed to the rear in a uniform, smooth and consistent manner.

[0065] To ensure smooth transport and prevent damage during topping removal, this embodiment of the invention also acquires real-time images of the harvested and detached fruits and vegetables on the conveyor belt, i.e., fruit and vegetable images. Specifically, image acquisition devices installed above or to the side of the conveyor belt continuously capture images of the fruits and vegetables during the clamping and transporting process, obtaining high-resolution fruit and vegetable images.

[0066] Subsequently, built-in deep learning object detection algorithms, such as the YOLO (You Only Look Once) v8 visual model, can be used to extract the boundary features of fruits and vegetables in the images, and based on this, determine the root-stem junction of the fruits and vegetables to be harvested. Specifically, the visual model can perform precise semantic segmentation or object bounding on the pixels in the fruit and vegetable images to extract boundary features reflecting the morphology of the fruits and vegetables, such as the boundary between the orange root outline of a carrot and the green stem and leaves. Using these features, the root-stem junction of each harvested and detached fruit and vegetable can be accurately identified and calculated without physical contact, i.e., the precise coordinates connecting the top of the fruit and the tassel.

[0067] Finally, the cutting height of the cutting mechanism in the fruit and vegetable harvester can be adjusted according to the root-stem junction position, and the adjusted cutting mechanism can be controlled to cut off the stems and leaves of the fruits and vegetables to be harvested. That is, after determining the precise coordinates, the cutting mechanism, such as a servo motor-controlled rotary cutter or a reciprocating cutter, can be driven in real time for height compensation and position fine-tuning. Since the cutting height is dynamically adjusted according to the actual root-stem junction position of each fruit and vegetable, when the fruit and vegetable arrive at the cutting position with the conveyor belt, the cutting mechanism can accurately cut off the stems and leaves at the set stem-leaf position.

[0068] In this embodiment of the invention, by implementing tiered matching speed control of the multi-stage conveying mechanism, the problem of pulling and breaking that easily occurs in the conveying process of traditional harvesting equipment is completely solved, ensuring the smooth flow of fruits and vegetables. At the same time, by extracting the boundary features of fruits and vegetables and accurately locating the root-stem junction to adjust the cutting mechanism in real time, the traditional mechanical baffle method that easily causes physical damage is abandoned. This not only effectively avoids cutting damage caused by pulling and leakage, but also ensures the high consistency of the stem length after the stems and leaves are cut off, comprehensively ensuring a low damage rate and greatly improving the appearance quality and value of fruits and vegetables after harvesting.

[0069] Based on the above embodiments, the adjusted cutting mechanism cuts off the stems and leaves of the fruits and vegetables to be harvested, and then further includes: Determine the stacking height of fruits and vegetables on the multi-stage conveying mechanism, and the load current of the motor driving the cutting mechanism in the fruit and vegetable harvester; Transmission blockage is determined based on the height of fruit and vegetable stacking and the load current. In the case of known transmission congestion, extract the height change characteristics of fruit and vegetable accumulation and the periodic fluctuation characteristics of load current. Based on the characteristics of high mutation and periodic fluctuation, the blockage type is identified to obtain the target blockage type; Based on the target congestion type, a congestion handling strategy is recommended.

[0070] Specifically, after the cutting mechanism in the fruit and vegetable harvester successfully removes the stems and leaves of the fruits and vegetables, the fruits and vegetables themselves and the cut stems and leaves continue to be conveyed backward on the conveyor belt. However, due to excessive fruit and vegetable yields, overly vigorous foliage, and poor mechanical connections, blockages and accumulations can easily occur during the conveying process. If these issues are not detected and addressed in time, newly harvested fruits and vegetables will continue to accumulate, eventually leading to serious mechanical failures or large-scale crushing and damage. Therefore, this embodiment of the invention introduces a blockage monitoring and diagnostic early warning mechanism after the stems and leaves are removed.

[0071] In detail, the height of the fruits and vegetables on the conveyor belt, i.e., the stacking height, and the load current of the motor driving the cutting mechanism are monitored in real time during the fruit and vegetable conveying process. In actual hardware deployment, infrared or ultrasonic sensors installed above the conveyor belt can detect and collect the height of the material stacking above the conveyor belt in real time, i.e., the stacking height of fruits and vegetables; at the same time, current sensors installed on the motor circuit can collect the load current of the motor in real time, and this current value can directly reflect the actual resistance encountered by the cutter when cutting the stems and leaves of fruits and vegetables.

[0072] Next, transmission blockage can be determined based on the height of the fruit and vegetable accumulation and the load current. That is, to avoid false alarms caused by normal fluctuations in fruit and vegetable volume, this embodiment of the invention presets normal operating condition thresholds, such as a height threshold and a maximum current threshold, which have been determined through experiments. When the height of the fruit and vegetable accumulation exceeds the set height threshold and continues to exceed the set time window, while the load current also continues to be higher than the maximum current threshold under normal operating conditions, it can be determined that not only has fruit and vegetable accumulation occurred, but the cutting power system has also encountered abnormal resistance, thus leading to the conclusion that transmission blockage has occurred.

[0073] Furthermore, in order to accurately identify the root cause of transmission blockage in order to address the problem effectively, this embodiment of the invention will also conduct in-depth analysis of the fruit and vegetable stacking height and load current. For example, by analyzing the historical time series data of the sensor, the height abrupt change characteristics reflecting the stacking evolution process can be extracted. If the fruit and vegetable stacking height shows a steep step-like increase in a very short period of time, and the periodic fluctuation characteristics reflecting the current change, such as the regular, high-amplitude abnormal peaks appearing on the current waveform of the motor, this means that the cutting mechanism is repeatedly stuck by the fruit and vegetables and tough leaves.

[0074] Subsequently, the blockage type can be identified based on the height abrupt change characteristics and periodic fluctuation characteristics to obtain the target blockage type. That is, the state machine or diagnostic algorithm in the fruit and vegetable harvester will combine these two extracted abnormal features for pattern matching. For example, when both the height abrupt change and the current periodic spikes are present, the current target blockage type can be accurately determined to be a blockage at the end of the clamping conveyor belt (the cause is usually that the cutting mechanism is jammed by foreign objects, causing the fruits and vegetables that cannot be cut in time to quickly backflow and accumulate). In contrast, if the height of the fruit and vegetable accumulation continues to rise slowly without obvious current spikes, it may be identified as other types of blockage, such as accumulation caused by insufficient power of the tassel removal conveyor belt.

[0075] Finally, a blockage handling strategy can be recommended based on the identified target blockage type. That is, once the target blockage type is identified, a corresponding blockage handling strategy can be matched to that specific blockage type. For example, a tiered alarm can be issued to the operator on the control panel, recommending an intervention strategy of "immediately stopping the machine and prompting manual removal of foreign objects from the cutter," or an automatic control strategy can be adopted to slow down the running speed of the front-end clamping conveyor belt to alleviate the accumulation pressure.

[0076] In this embodiment of the invention, accurate transmission blockage judgment and type identification are achieved by using the fruit and vegetable stacking height and load current. This not only enables real-time and accurate monitoring of abnormal blockages during the conveying process, but also allows for in-depth analysis of the causes of blockages and recommendations for targeted handling strategies. This effectively avoids long-term downtime caused by severe blockages, significantly reduces the time cost of troubleshooting, and reduces damage to fruits and vegetables caused by large-area compression and stacking on the conveyor belt. This further ensures the efficient, stable, and low-loss operation of the fruit and vegetable harvester.

[0077] Based on the above embodiments, route planning is crucial in the unmanned operation of fruit and vegetable harvesters, serving as the core of the entire unmanned control system. This system achieves precise control over actions such as walking, turning, and U-turns through path planning. The complete path planning process includes full-coverage path planning for the harvesting area and turning / U-turn planning for non-harvesting areas. The specific path planning logic is as follows: When performing full-coverage path planning, the Boustrophedon cell decomposition method is used as the algorithm basis. The algorithm scans the polygonal work area along a specific direction (such as the y-axis). When the topology of the scan line changes, such as encountering obstacle vertices or polygon boundary vertices that cause changes in the number or order of intersections, key points are created at those locations. By connecting the line segments between adjacent key points, the entire work area can be divided into multiple sub-regions along the scan direction (each sub-region's boundary consists of two line segments parallel to the scan direction at the top and bottom, and vertical boundaries on the left and right sides), thus decomposing the entire work area into multiple trapezoidal cells. Subsequently, a standard bow-shaped path is automatically generated within each trapezoidal cell, where the row spacing between paths is equal to the effective working width of the fruit and vegetable harvester.

[0078] To ensure that the planned path achieves the most efficient operation, this embodiment of the invention sets the objective of minimizing the total path length, namely: in, It is the first Effective work path length within a trapezoidal unit It is in the The length of the transition turning path within and between each trapezoidal unit.

[0079] Therefore, a global optimization objective function can be constructed, namely: in, This represents the total travel distance. For the number of turns, This refers to the length of the empty (non-operational) route. , and The weighting coefficients are set to flexibly reflect the priority of different optimization objectives such as fuel consumption, operation time, and mechanical wear, in order to obtain the coverage route with the lowest overall cost.

[0080] When planning turns and U-turns, when the fruit and vegetable harvester ends each work row and needs to turn around to transition to the next work row, the Dubins path or Reeds-Shepp path planning algorithm is invoked to calculate the shortest smooth path connecting two parallel work lines while satisfying the minimum turning radius constraint of the fruit and vegetable harvester.

[0081] Taking Dubins path calculation as an example, its turning path consists of a combination of straight line segments and circular arc segments with a fixed radius. When calculating the specific turning path length, the given starting point pose, target point pose, and minimum turning radius are input into the Dubins path calculator, which can then quickly solve for the most efficient shortest path type and its corresponding length.

[0082] However, in actual turning strategy selection, at the global planning level, the implementation costs of different turning geometries, such as Ω-shaped, pear-shaped, and semi-circular, are comprehensively evaluated to select the optimal strategy that minimizes the turning path length or the turning time. This strategy is then rigorously verified before outputting control commands to ensure absolute safety and collision-free operation of the entire turning area.

[0083] Based on the above embodiments, the end of the multi-stage conveying mechanism is provided with a baffle structure with an inclination angle; the baffle structure is used to intercept the fruits and vegetables to be harvested conveyed by the multi-stage conveying mechanism at the end position, so as to change the falling posture of the fruits and vegetables to be harvested.

[0084] Specifically, after successfully harvesting fruits and vegetables, the harvested and detached produce is conveyed backward through a multi-stage conveyor system, eventually falling into a collection box or packaging bag. In traditional harvesting equipment, when fruits and vegetables detach from the end of the conveyor belt, they often fall straight down vertically onto the collection device below. This unprotected drop easily causes the tops of the fruits and vegetables to crack or break, a significant blind spot contributing to the high harvesting breakage rate. Therefore, in this embodiment of the invention, a baffle structure to prevent direct drop is installed at the end of the multi-stage conveyor system.

[0085] Figure 2 This is an example diagram of the anti-falling baffle structure provided by the present invention, such as... Figure 2 As shown, the baffle structure has a certain angle of inclination. Specifically, this angled baffle structure can be a U-shaped anti-fall baffle added to the end of the conveyor belt. To achieve the best force buffering and guiding effect, the installation angle of this baffle structure can be a specific flexible angle with the horizontal line of the conveyor belt, such as approximately 30 degrees.

[0086] In actual continuous harvesting and conveying, the baffle structure can intercept the fruits and vegetables conveyed by the conveyor belt at the end position to change the falling posture of the fruits and vegetables to be harvested, so that the fruits and vegetables to be harvested fall into the flexible buffer collection structure in a gentle posture. That is, when the fruits and vegetables with their stems and leaves removed are conveyed to the end of the conveyor belt and are about to fall, they will first touch and impact the baffle structure with a 30-degree inclination. The baffle structure plays a key role in interception and guidance. It can force the fruits and vegetables to slide along the inclined baffle surface, thereby changing the falling posture of the fruits and vegetables. It adjusts the original vertical falling state, which is very vulnerable to damage, to a horizontal lying state with a larger force-bearing area. Subsequently, the fruits and vegetables will fall smoothly into the collection box or ton bag below in a gentle posture with the impact force greatly weakened.

[0087] In addition, it is worth noting that, in order to minimize end-point damage, in this embodiment of the invention, the material of the conveyor belt can be changed from traditional hard plastic to soft cloth-like rubber material, utilizing the elasticity of the rubber material itself to absorb impact kinetic energy, thereby further reducing losses.

[0088] In this embodiment of the invention, by adding an angled baffle structure at the end of the multi-stage conveying mechanism, the dangerous vertical drop posture of fruits and vegetables during the final collection stage is changed. By using the interception and guidance of the angled baffle, the vertical drop of fruits and vegetables is transformed into a flat landing. With the assistance of soft rubber material for stress relief, the problem of the top of fruits and vegetables cracking and breaking at the end of the harvest is solved, and a flexible closed loop of the entire harvesting process is realized. This provides a decisive hardware guarantee for achieving the goal of extremely low damage rate (less than 1%) of the whole machine harvesting.

[0089] The fruit and vegetable harvesting device provided by the present invention is described below. The fruit and vegetable harvesting device described below can be referred to in correspondence with the fruit and vegetable harvesting method described above.

[0090] Figure 3 This is a schematic diagram of the fruit and vegetable harvesting device provided by the present invention, as shown below. Figure 3 As shown, this device is used in a fruit and vegetable harvester, and the device includes: The identification unit 310 is used to acquire a regional image of the harvesting area and determine the center line of the planting row of the fruits and vegetables to be harvested in the harvesting area based on the regional image. Prediction unit 320 is used to determine the harvesting area corresponding to the center line of the planting row in the harvesting area, and to predict the harvesting point based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, so as to obtain the target harvesting point of the fruits and vegetables to be harvested. Harvesting unit 330 is used to generate a continuous dynamic harvesting trajectory based on the target harvesting point, and to harvest the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory.

[0091] The fruit and vegetable harvesting device provided by this invention determines the center line of the planting rows through regional images of the harvesting area, realizing intelligent and autonomous row alignment of agricultural machinery and eliminating reliance on human driving experience. It predicts the target harvesting point through spatial planting distribution characteristics and generates a continuous dynamic harvesting trajectory accordingly, enabling the actuator to make flexible and smooth dynamic adjustments based on the actual growth, thereby achieving unmanned and precise control. This effectively avoids the problems of reduced efficiency and high damage rate caused by frequent vertical lifting and blind operation of traditional machinery. While significantly improving the efficiency of automated fruit and vegetable harvesting, it reduces the harvesting damage rate to an extremely low level from the root cause.

[0092] Based on the above embodiments, the spatial planting distribution characteristics include planting location, growth direction, and main axis direction; Prediction unit 320 is used for: Determine the spatial distribution image of the planting area for each harvesting zone; Color space conversion and feature extraction are performed on the planting spatial distribution image to obtain the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested in each harvesting area; Based on the planting location, growth direction, and main axis direction of the fruits and vegetables to be harvested, the harvesting point is predicted to obtain the target harvesting point of the fruits and vegetables to be harvested. Harvesting unit 330 is used for: Multiple target harvesting points obtained from continuous prediction are subjected to fuzzy prediction to obtain a continuous wave-shaped harvesting curve, and the wave-shaped harvesting curve is used as the dynamic harvesting trajectory.

[0093] Based on the above embodiments, the spatial planting distribution characteristics also include the average size of the fruits and vegetables to be harvested; the device further includes a frequency adjustment unit, used for: Based on the standard size and average size of the fruits and vegetables to be harvested, a benchmark resistance threshold for distinguishing geological types is determined. The real-time soil resistance experienced by the soil loosening mechanism of the fruit and vegetable harvester in each harvesting area is compared with the benchmark resistance threshold, and the vibration frequency of the soil loosening mechanism in each harvesting area is adjusted based on the comparison results.

[0094] Based on the above embodiments, the device further includes an attitude adjustment unit, used for: The machine posture data of the fruit and vegetable harvester and the terrain data of each harvesting area are obtained, and the terrain slope is extracted from the terrain data. Based on the terrain slope, the terrain compensation parameters are determined. Extract the current pitch angle and pitch angle change rate from the fuselage attitude data, and determine the basic adjustment parameters based on the current pitch angle and pitch angle change rate; Generate attitude adjustment commands based on the basic adjustment parameters and the terrain compensation parameters; Based on the attitude adjustment command, the suspension system of the fruit and vegetable harvester is adjusted to keep the harvester balanced.

[0095] Based on the above embodiments, the device further includes a height adjustment unit, used for: Determine the current speed of the fruit and vegetable harvester; Based on the current driving speed, the multi-stage conveying mechanism in the fruit and vegetable harvester is controlled to convey the harvested and separated fruits and vegetables at a speed matched in stages. Acquire images of the fruits and vegetables to be harvested during the transportation process; Extract the fruit and vegetable boundary features from the fruit and vegetable image, and determine the root-stem junction position of the fruit and vegetable to be harvested based on the fruit and vegetable boundary features; Based on the root-stem junction position, the cutting height of the cutting mechanism in the fruit and vegetable harvester is adjusted, and the adjusted cutting mechanism is controlled to cut off the stems and leaves of the fruits and vegetables to be harvested.

[0096] Based on the above embodiments, the device further includes a blockage identification unit, used for: Determine the stacking height of fruits and vegetables on the multi-stage conveying mechanism, and the load current of the motor driving the cutting mechanism in the fruit and vegetable harvester; Transmission blockage is determined based on the height of the fruit and vegetable stack and the load current. In the event of a known transmission blockage, the height abrupt change characteristics of the fruit and vegetable stacking height and the periodic fluctuation characteristics of the load current are extracted. Based on the highly abrupt change characteristics and the periodic fluctuation characteristics, the blockage type is identified to obtain the target blockage type; Based on the target congestion type, a congestion handling strategy is recommended.

[0097] Based on the above embodiments, the end of the multi-stage conveying mechanism is provided with a baffle structure with an inclination angle; The baffle structure is used to intercept the fruits and vegetables to be harvested conveyed by the multi-stage conveying mechanism at the end position, so as to change the falling posture of the fruits and vegetables to be harvested.

[0098] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communication interface 420, a memory 430, and a communication bus 440. The processor 410, communication interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a fruit and vegetable harvesting method. This method is applied to a fruit and vegetable harvesting machine and includes: acquiring a regional image of the harvesting area; determining the center line of the planting rows of fruits and vegetables to be harvested in the harvesting area based on the regional image; determining the harvesting zones corresponding to the center lines of the planting rows in the harvesting area; predicting harvesting points based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting zone to obtain the target harvesting points of the fruits and vegetables to be harvested; generating a continuous dynamic harvesting trajectory based on the target harvesting points; and harvesting the fruits and vegetables to be harvested in each harvesting zone based on the dynamic harvesting trajectory.

[0099] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] On the other hand, the present invention also provides a computer program product, the computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions, which, when executed by a computer, enable the computer to execute the fruit and vegetable harvesting methods provided by the above methods. This method is applied to a fruit and vegetable harvesting machine and includes: acquiring a regional image of the harvesting area; determining the center line of the planting rows of fruits and vegetables to be harvested in the harvesting area based on the regional image; determining the harvesting zones corresponding to the center lines of the planting rows in the harvesting area; predicting harvesting points based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting zone; generating a continuous dynamic harvesting trajectory based on the target harvesting points; and harvesting the fruits and vegetables to be harvested in each harvesting zone based on the dynamic harvesting trajectory.

[0101] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the fruit and vegetable harvesting methods provided by the above methods. This method is applied to a fruit and vegetable harvesting machine and includes: acquiring a regional image of a harvesting area; determining, based on the regional image, the center line of the planting rows of fruits and vegetables to be harvested in the harvesting area; determining the harvesting zones corresponding to the center lines of the planting rows in the harvesting area; predicting harvesting points based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting zone; generating a continuous dynamic harvesting trajectory based on the target harvesting points; and harvesting the fruits and vegetables to be harvested in each harvesting zone based on the dynamic harvesting trajectory.

[0102] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0103] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for harvesting fruits and vegetables, characterized in that, Applications in fruit and vegetable harvesters include: Acquire a regional image of the harvesting area, and determine the center line of the planting row of the fruits and vegetables to be harvested in the harvesting area based on the regional image; Determine the harvesting area corresponding to the center line of the planting row in the harvesting area, and predict the harvesting point based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area to obtain the target harvesting point of the fruits and vegetables to be harvested. A continuous dynamic harvesting trajectory is generated based on the target harvesting points, and the fruits and vegetables to be harvested in each harvesting area are harvested based on the dynamic harvesting trajectory.

2. The fruit and vegetable harvesting method according to claim 1, characterized in that, The spatial planting distribution characteristics include planting location, growth direction, and main axis direction; The harvesting point is predicted based on the spatial planting distribution characteristics of fruits and vegetables to be harvested in each harvesting area, and the target harvesting point of the fruits and vegetables to be harvested is obtained. Generate a continuous dynamic harvesting trajectory based on the target harvesting point, including: Determine the spatial distribution image of the planting area for each harvesting zone; Color space conversion and feature extraction are performed on the planting spatial distribution image to obtain the planting location, growth direction and main axis direction of the fruits and vegetables to be harvested in each harvesting area; Based on the planting location, growth direction, and main axis direction of the fruits and vegetables to be harvested, the harvesting point is predicted to obtain the target harvesting point of the fruits and vegetables to be harvested. Multiple target harvesting points obtained from continuous prediction are subjected to fuzzy prediction to obtain a continuous wave-shaped harvesting curve, and the wave-shaped harvesting curve is used as the dynamic harvesting trajectory.

3. The fruit and vegetable harvesting method according to claim 2, characterized in that, The spatial planting distribution characteristics also include the average size of the fruits and vegetables to be harvested; The harvesting of fruits and vegetables in each harvesting area based on the dynamic harvesting trajectory also includes the following steps beforehand: Based on the standard size and average size of the fruits and vegetables to be harvested, a benchmark resistance threshold for distinguishing geological types is determined. The real-time soil resistance experienced by the soil loosening mechanism of the fruit and vegetable harvester in each harvesting area is compared with the benchmark resistance threshold, and the vibration frequency of the soil loosening mechanism in each harvesting area is adjusted based on the comparison results.

4. The fruit and vegetable harvesting method according to any one of claims 1 to 3, characterized in that, The harvesting of fruits and vegetables in each harvesting area based on the dynamic harvesting trajectory also includes the following steps beforehand: The machine posture data of the fruit and vegetable harvester and the terrain data of each harvesting area are obtained, and the terrain slope is extracted from the terrain data. Based on the terrain slope, the terrain compensation parameters are determined. Extract the current pitch angle and pitch angle change rate from the fuselage attitude data, and determine the basic adjustment parameters based on the current pitch angle and pitch angle change rate; Generate attitude adjustment commands based on the basic adjustment parameters and the terrain compensation parameters; Based on the attitude adjustment command, the suspension system of the fruit and vegetable harvester is adjusted to keep the harvester balanced.

5. The method for harvesting fruits and vegetables according to any one of claims 1 to 3, characterized in that, Based on the dynamic harvesting trajectory, the harvesting of fruits and vegetables to be harvested in each harvesting area is then included, followed by: Determine the current speed of the fruit and vegetable harvester; Based on the current driving speed, the multi-stage conveying mechanism in the fruit and vegetable harvester is controlled to convey the harvested and separated fruits and vegetables at a speed matched in stages. Acquire images of the fruits and vegetables to be harvested during the transportation process; Extract the fruit and vegetable boundary features from the fruit and vegetable image, and determine the root-stem junction position of the fruit and vegetable to be harvested based on the fruit and vegetable boundary features; Based on the root-stem junction position, the cutting height of the cutting mechanism in the fruit and vegetable harvester is adjusted, and the adjusted cutting mechanism is controlled to cut off the stems and leaves of the fruits and vegetables to be harvested.

6. The fruit and vegetable harvesting method according to claim 5, characterized in that, The controlled and adjusted cutting mechanism cuts off the stems and leaves of the fruits and vegetables to be harvested, and then further includes: Determine the stacking height of fruits and vegetables on the multi-stage conveying mechanism, and the load current of the motor driving the cutting mechanism in the fruit and vegetable harvester; Transmission blockage is determined based on the height of the fruit and vegetable stack and the load current. In the event of a known transmission blockage, the height abrupt change characteristics of the fruit and vegetable stacking height and the periodic fluctuation characteristics of the load current are extracted. Based on the highly abrupt change characteristics and the periodic fluctuation characteristics, the blockage type is identified to obtain the target blockage type; Based on the target congestion type, a congestion handling strategy is recommended.

7. The fruit and vegetable harvesting method according to claim 5, characterized in that, The end of the multi-stage conveying mechanism is provided with a baffle structure with an inclination angle; The baffle structure is used to intercept the fruits and vegetables to be harvested conveyed by the multi-stage conveying mechanism at the end position, so as to change the falling posture of the fruits and vegetables to be harvested.

8. A fruit and vegetable harvesting device, characterized in that, Applications in fruit and vegetable harvesters include: The identification unit is used to acquire a regional image of the harvesting area and determine the center line of the planting row of the fruits and vegetables to be harvested in the harvesting area based on the regional image. The prediction unit is used to determine the harvesting area corresponding to the center line of the planting row in the harvesting area, and to predict the harvesting point based on the spatial planting distribution characteristics of the fruits and vegetables to be harvested in each harvesting area, so as to obtain the target harvesting point of the fruits and vegetables to be harvested. The harvesting unit is used to generate a continuous dynamic harvesting trajectory based on the target harvesting point, and to harvest the fruits and vegetables to be harvested in each harvesting area based on the dynamic harvesting trajectory.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the fruit and vegetable harvesting method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the fruit and vegetable harvesting method as described in any one of claims 1 to 7.

Images