Self-propelled chrysanthemum intelligent harvesting machine and automatic control system and method thereof

Abstract

The application provides a self-propelled chrysanthemum intelligent harvesting machine and an automatic control system and method thereof, and the device comprises a harvesting header and a four-wheel mobile chassis hinged with the harvesting header through a header fixed hinge device, a chrysanthemum collecting device is installed on the four-wheel mobile chassis, and the chrysanthemum collecting device is connected with a pneumatic conveying channel; the harvesting header comprises a lifting mechanism and a cutter mechanism installed on a side protection plate of the header; a power module is fixed on the lifting mechanism; the power module drives a roller driving module installed at both ends of a roller; a variable-spacing comb plate device and a fixed comb plate device are arranged on the roller; and binocular cameras are symmetrically arranged on both sides of the harvesting header shell. The application can ensure the chrysanthemum harvesting efficiency, improve the chrysanthemum picking rate and reduce the chrysanthemum breakage rate.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent harvesting of chrysanthemum, and particularly relates to a self-propelled chrysanthemum intelligent harvesting machine and an automatic control system and method thereof. BACKGROUND

[0002] Chrysanthemum is a perennial herb with high economic value. However, the chrysanthemum picking in the rural planting base is still manual picking, and even if a tool with comb teeth is used for picking, the manual picking has high labor cost and is heavy on the human body, and the chrysanthemum has a short flowering period, so the picking efficiency is low, and the chrysanthemum will wither, so the present application designs a self-propelled chrysanthemum intelligent harvesting machine and an automatic control system and method thereof to realize intelligent chrysanthemum picking. SUMMARY

[0003] In view of the deficiencies of the prior art, the present application provides a self-propelled chrysanthemum intelligent harvesting machine and an automatic control system and method thereof, which can improve the picking rate of chrysanthemum and reduce the damage rate of chrysanthemum while ensuring the harvesting efficiency of chrysanthemum.

[0004] To achieve the above object, the present application provides the following scheme:

[0005] A self-propelled chrysanthemum intelligent harvesting machine, comprising: a harvesting header and a four-wheel mobile chassis hinged to the harvesting header through a header fixing hinge device, wherein a chrysanthemum collecting device is installed on the four-wheel mobile chassis, and the chrysanthemum collecting device is connected with a pneumatic conveying channel.

[0006] The harvesting header comprises a lifting mechanism and a cutter mechanism installed on a header side guard plate.

[0007] A power module is fixed on the lifting mechanism.

[0008] The power module drives a roller driving module installed at both ends of the roller.

[0009] The roller is provided with a variable-pitch comb plate device and a fixed comb plate device.

[0010] Double cameras are symmetrically arranged on both sides of the harvesting header shell.

[0011] Preferably, the power module is used to provide power for the harvesting header, and comprises a motor mounting cover plate and a driving motor mounted on the motor mounting cover plate.

[0012] Preferably, the roller driving module comprises a roller driving cover plate, a driven gear and a driving gear.

[0013] The roller driving cover plate is fixedly installed on the driven gear, and the driven gear is driven by the driving gear.

[0014] The driving gear is driven by the driving motor mounted on the motor mounting cover plate;

[0015] Preferably, the lifting mechanism comprises a lifting base plate, a rolling slider, a lifting rail and a rail fixing block.

[0016] The lifting base plate is fixedly installed on the rolling slider, the rolling slider is installed on the lifting rail, the lifting rail is installed in the rail fixing block, and the lifting mechanism is installed on the side protection plate of the cutting table through two groups of rail fixing blocks.

[0017] Preferably, the cutting knife mechanism comprises a reciprocating cutting knife, an electric push rod and an electric push rod fixing block; and the cutting knife mechanism is installed on the side protection plate of the cutting table through the electric push rod fixing block.

[0018] The application also provides an automatic control system of the self-propelled chrysanthemum intelligent harvesting machine.

[0019] The image acquisition module is used for acquiring chrysanthemum images based on a binocular camera.

[0020] The image processing module is used for processing the chrysanthemum images.

[0021] The chrysanthemum recognition module is used for inputting the processed chrysanthemum images into a chrysanthemum recognition counting model to obtain the number of chrysanthemums meeting a preset pistil diameter.

[0022] The control parameter acquisition module is used for obtaining self-propelled chrysanthemum intelligent harvesting machine control parameters based on the number of chrysanthemums.

[0023] The control module is used for adjusting the self-propelled chrysanthemum intelligent harvesting machine based on the control parameters to complete chrysanthemum harvesting.

[0024] Preferably, the image processing module comprises a decomposition unit, a reflection component adjustment unit, an illumination component adjustment unit and an image enhancement unit.

[0025] The decomposition unit is used for decomposing the collected natural light chrysanthemum images to obtain reflection components, illumination components and brightness adjustment proportion maps of high-light chrysanthemum images and low-light chrysanthemum images respectively.

[0026] The reflection component adjustment unit is used for splicing the illumination components and the reflection components of the low-light chrysanthemum images, guiding the reflection components to be denoised by using the illumination components, and taking the reflection components of the high-light chrysanthemum images as reference images to obtain denoised reflection components.

[0027] The light component adjusting unit is configured to splice the light component of the low-light chrysanthemum image with the brightness adjustment proportion map, take the light component of the high-light chrysanthemum image as a reference image, and obtain an adjusted light component.

[0028] The image enhancement unit is configured to fuse the denoised reflection component and the adjusted light component to obtain an enhanced chrysanthemum image, thereby completing the processing of the chrysanthemum image.

[0029] Preferably, the chrysanthemum recognition module comprises a feature extraction unit, a feature processing unit, a recognition unit, a diameter calculation unit, and a counting unit.

[0030] The feature extraction unit is configured to use a VGG16 convolutional neural network as a feature extraction network to perform multi-scale feature extraction on the enhanced chrysanthemum image, thereby obtaining multi-scale information.

[0031] The feature processing unit is configured to process the multi-scale information based on a context perception layer to obtain context features.

[0032] The recognition unit is configured to obtain a chrysanthemum recognition result based on the context features.

[0033] The diameter calculation unit is configured to extract a chrysanthemum pistil contour based on the chrysanthemum recognition result using edge detection, and obtain a chrysanthemum pistil diameter based on the pixel size of the chrysanthemum pistil contour and the image resolution.

[0034] The counting unit is configured to filter chrysanthemums that meet a preset pistil diameter based on the pistil diameter of the recognized chrysanthemums, thereby completing the construction of a chrysanthemum recognition counting model.

[0035] Preferably, the regulation parameter acquisition module comprises an optimal rotating speed acquisition unit, a comb tooth spacing acquisition unit, and a cutting platform height acquisition unit.

[0036] The optimal rotating speed acquisition unit is configured to calculate a roller load based on the number of chrysanthemums, and obtain a comb tooth optimal rotating speed based on the roller load.

[0037] The comb tooth spacing acquisition unit is configured to calculate an average diameter of chrysanthemums that meet the preset pistil diameter, and obtain a comb tooth spacing based on the average diameter.

[0038] The cutting platform height acquisition unit is configured to obtain pixel coordinates of the highest point of chrysanthemums that meet the preset pistil diameter, and obtain chrysanthemum actual height information based on the pixel coordinates of the highest point, and obtain a cutting platform height adjustment parameter based on the chrysanthemum actual height information.

[0039] The application also provides an automatic regulation method of a self-propelled chrysanthemum intelligent harvesting machine.

[0040] Collecting a chrysanthemum image based on a binocular camera, and processing the chrysanthemum image;

[0041] Inputting the processed chrysanthemum image into a chrysanthemum recognition and counting model to obtain a number of chrysanthemums meeting a preset pistil diameter;

[0042] Based on the number of chrysanthemums, obtaining a self-propelled chrysanthemum intelligent harvester control parameter;

[0043] Based on the control parameter, adjusting the self-propelled chrysanthemum intelligent harvester to complete chrysanthemum harvesting.

[0044] Compared with the prior art, the beneficial effects of the present application are that the technical solution of the present application can better adjust the rotating speed of the comb teeth under the load that the rotating speed of the drum can withstand when the size and density of the chrysanthemums are different, and realize real-time adjustment of the header height of the chrysanthemum picking machine when the height of the chrysanthemum picking point is different, so as to get rid of the current situation that the rotating speed of the comb teeth and the header height can only be manually controlled. In the case of ensuring the chrysanthemum harvesting efficiency, the chrysanthemum picking rate is improved and the chrysanthemum breakage rate is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the present application, the following briefly introduces the drawings needed to be used in the embodiments, and obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0046] Figure 1 It is a structural schematic diagram of the self-propelled chrysanthemum intelligent harvester of the embodiment of the present application.

[0047] Figure 2 It is a structural schematic diagram of the chrysanthemum harvesting header of the embodiment of the present application.

[0048] Figure 3 It is a front view of the chrysanthemum harvesting header of the embodiment of the present application.

[0049] Figure 4 It is a two-dimensional schematic diagram of the key structure of the chrysanthemum harvesting header of the embodiment of the present application.

[0050] Figure 5 It is a local sectional view of the key structure of the chrysanthemum harvesting header of the embodiment of the present application.

[0051] Figure 6 It is a three-dimensional schematic diagram of the key structure of the chrysanthemum harvesting header of the embodiment of the present application.

[0052] Figure 7 It is a flow chart of the automatic control method of the self-propelled chrysanthemum intelligent harvester of the embodiment of the present application.

[0053] Figure 8 A chrysanthemum recognition counting model training flowchart for an embodiment of the present application;

[0054] Figure 9 A cutting table displacement distance acquisition diagram for an embodiment of the present application;

[0055] Figure 10 A working schematic diagram of a self-propelled chrysanthemum intelligent harvesting machine for an embodiment of the present application.

[0056] BRIEF DESCRIPTION OF DRAWINGS: 1 - power module, 101 - drive motor, 102 - motor mounting cover plate, 2 - roller drive module, 201 - rolling drive cover plate, 202 - driven gear, 203 - drive gear, 3 - harvesting header, 301 - header side guard, 302 - header fixed hinging device, 4 - lifting mechanism, 401 - lifting bottom plate, 402 - rolling slide block, 403 - lifting guide rail, 404 - guide rail fixed block, 5 - cutter mechanism, 501 - reciprocating cutter, 502 - electric push rod, 503 electric push rod fixed block, 6 - binocular camera, 7 - variable pitch comb plate device, 701 - gear shaft, 702 - rack, 703 - rack fixing frame, 704 - cam, 705 - small comb plate I, 706 - small comb plate II, 707 - thrust bearing, 708 - ring linear motor, 709 - variable pitch support frame, 710 - guide pipe seat, 711 - guide rod, 8 - fixed comb plate device, 9 - roller, 901 - roller rotating shaft, 10 - rolling brush, 11 - Roots blower, 12 - spiral flower discharging device, 13 - pneumatic conveying channel, 14 - four-wheel mobile chassis, 15 - chrysanthemum collecting device, h1 - inflorescence width; h2 - depth of working device penetrating inflorescence layer; h3 - flower height; h4 - header height. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the protection scope of the present application.

[0058] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0059] Embodiment one

[0060] As Figures 1-6As shown, a self-propelled chrysanthemum intelligent harvesting machine, comprising: a self-propelled chrysanthemum intelligent harvesting machine, comprising: a harvesting header 3 and a four-wheel mobile chassis 14 hinged with the harvesting header 3 through a header fixed hinge device 302, a chrysanthemum collecting device 15 is installed on the four-wheel mobile chassis 14, and the chrysanthemum collecting device 15 is connected with a pneumatic conveying channel 13; the harvesting header 3 is symmetrically arranged with double cameras 6 on both sides of the shell.

[0061] The harvesting header 3 comprises a lifting mechanism 4 installed on the header side guard plate 301 and a cutter mechanism 5;

[0062] The lifting mechanism 4 is fixed with a power module 1;

[0063] The power module 1 drives a roller driving module 2 installed at both ends of a roller 9;

[0064] The roller 9 is provided with a variable pitch comb plate device 7 and a fixed comb plate device 8.

[0065] Further embodiments are that the power module 1 is used to provide power for the harvesting header 3, comprising a motor mounting cover plate 102 and a driving motor 101 installed on the motor mounting cover plate 102.

[0066] Further embodiments are that the roller driving module 2 comprises a roller driving cover plate 201, a driven gear 202 and a driving gear 203;

[0067] The roller driving cover plate 201 is fixedly installed on the driven gear 202, and the driven gear 202 is driven by the driving gear 203;

[0068] The driving gear 203 is driven by the driving motor 101 installed on the motor mounting cover plate 102;

[0069] Further embodiments are that the lifting mechanism 4 comprises a lifting bottom plate 401, a rolling slide 402, a lifting guide rail 403 and a guide rail fixing block 404; the motor mounting cover plate 102 is fixed on the lifting bottom plate 401;

[0070] The lifting bottom plate 401 is fixedly installed on the rolling slide 402, the rolling slide 402 is installed on the lifting guide rail 403, the lifting guide rail 403 is installed in the guide rail fixing block 404, and the lifting mechanism 4 is installed on the header side guard plate 304 through the two groups of guide rail fixing blocks 404.

[0071] Further embodiments are that the cutter mechanism 5 comprises a reciprocating cutter 501, an electric push rod 502 and an electric push rod fixing block 503; the cutter mechanism 5 is installed on the header side guard plate 301 through the electric push rod fixing block 503.

[0072] The binocular camera 6 is symmetrically arranged on both sides of the outer shell of the harvesting header 3. The variable-pitch comb plate device 7 mainly comprises a gear shaft 701, a rack 702, a rack fixing frame 703, a cam 704, a small comb plate I 705, a small comb plate II 706, a thrust bearing 707, a ring linear motor 708, a variable-pitch support frame 709, a guide pipe base 710, and a guide rod 711.

[0073] In the embodiment, the four sets of variable-pitch comb plate devices 7 are installed on the roller 9 at intervals of 90°, and the roller driving cover plate 201 is arranged on both ends of the roller 9. The roller driving cover plate 201 is internally provided with the ring linear motor 708. One end of the ring linear motor 708 is provided with the thrust bearing 707. One end of the thrust bearing 707 is the variable-pitch support frame 709. The variable-pitch support frame 709 is provided with four sets of racks 702 at intervals of 90°. Each rack 702 is provided with 20 equidistant gear shafts 701. The tail end of each gear shaft 701 is provided with the cam 704. The cam 704 is arranged inside the two symmetrically-shaped small comb plate I 705 and the small comb plate II 706. The guide rod 711 is arranged in the end hole of the small comb plate I 705 and the small comb plate II 706. Both ends of the guide rod 711 are installed on the roller driving cover plate 201.

[0074] The harvesting header further comprises the rolling brush 10, the Roots blower 11, and the spiral flower arranging device 12, which are all installed on the header side guard plate 301. The pneumatic conveying channel 13 is connected to the Roots blower 11 and the chrysanthemum collecting device 15, respectively. The harvesting header 3 comprises the power module 1, the roller driving module 2, the lifting mechanism 4, the cutter mechanism 5, the variable-pitch comb plate device 7, the fixed comb plate device 8, and the roller 9. The harvesting header 3 is hinged to the four-wheel mobile chassis 14 through the header fixed hinge device 302. The chrysanthemum collecting device 15 is installed on the four-wheel mobile chassis 14.

[0075] A further embodiment is that when the two sets of binocular cameras 6 arranged on the harvesting header 3 identify that the height and growth density of the chrysanthemums in front change, the data is processed through the chrysanthemum recognition counting model, and corresponding parameters are output to control the lifting mechanism 4, such as Figure 10As shown, the outer diameter of the working comb tooth is ensured to be within the depth h2 interval penetrated by the working device to the layer of the inflorescence, and the length of the annular linear motor 708 is controlled to extend or retract, the thrust bearing 707 at one end of the annular linear motor 708 pushes the variable-pitch support frame 709, the rack 702 is moved, the fixed-axis gear shaft 701 is rotated by the rack 702, the cam 704 fixed at the tail end is rotated at the same time, thereby changing the distance between the pair of small comb tooth plates I 705 and the small comb tooth plates II 706, and the distance between the 20 sets of small comb tooth plates I 705 and the small comb tooth plates II 706 is changed at the same time, so that the adjustment of the distance between the adjacent comb teeth of each pair of comb tooth plates is realized. The speed of the driving motor 101 is controlled at the same time, and the adaptive control of the picking comb tooth speed in different chrysanthemum planting environments is realized.

[0076] Further embodiments are that the cutter mechanism 5 can cut the chrysanthemum stems after harvesting. The chrysanthemums harvested by the comb are pushed into the spiral flower arranging device 12 by the rotation of the rolling brush 10. Further, the chrysanthemums at the discharge port of the spiral flower arranging device 12 are sucked into the pneumatic conveying channel 13 by the Roots blower 11, and are transmitted to the chrysanthemum collecting device 15.

[0077] Embodiment two

[0078] As Figures 7-10 The application also provides an automatic control system of the self-propelled chrysanthemum intelligent harvesting machine, which is used for controlling the chrysanthemum harvesting machine and includes an image acquisition module, an image processing module, a chrysanthemum recognition module, a control parameter acquisition module, and a control module.

[0079] The image acquisition module is used for acquiring chrysanthemum images based on a binocular camera.

[0080] The image processing module is used for processing the chrysanthemum images.

[0081] Further embodiments are that the image processing module includes a decomposition unit, a reflection component adjustment unit, an illumination component adjustment unit, and an image enhancement unit.

[0082] The decomposition unit is used for decomposing the natural light chrysanthemum images acquired to obtain the reflection components, the illumination components, and the brightness adjustment proportion map corresponding to the high-light chrysanthemum images and the low-light chrysanthemum images respectively.

[0083] The reflection component adjustment unit is used for splicing the illumination components and the reflection components of the low-light chrysanthemum images, using the illumination components to guide the reflection components to be denoised, and using the reflection components of the high-light chrysanthemum images as reference images to obtain the denoised reflection components.

[0084] The illumination component adjustment unit is used for splicing the illumination components of the low-light chrysanthemum images with the brightness adjustment proportion map, using the illumination components of the high-light chrysanthemum images as reference images to obtain the adjusted illumination components.

[0085] The image enhancement unit is configured to fuse the de-noised reflection component and the adjusted illumination component to obtain an enhanced chrysanthemum image, thereby completing the processing of the chrysanthemum image.

[0086] The chrysanthemum recognition module is configured to input the processed chrysanthemum image into a chrysanthemum recognition counting model to obtain a number of chrysanthemums that satisfy a preset diameter of a pistil; in this embodiment, the preprocessed picture samples are divided into training samples and test samples according to a ratio of 9:1, and the training samples are used to train the chrysanthemum recognition counting model.

[0087] Further embodiments are directed to the chrysanthemum recognition module, which includes a feature extraction unit, a feature processing unit, a recognition unit, a diameter calculation unit, and a counting unit.

[0088] The feature extraction unit is configured to use a VGG16 convolutional neural network as a feature extraction network to perform multi-scale feature extraction on the enhanced chrysanthemum image to obtain multi-scale information; the VGG16 network is composed of multiple convolutional layers and pooling layers, which extract the features of the image layer by layer. The convolutional layer extracts the local features of the image through convolution operation, while the pooling layer reduces the size of the feature map through max-pooling operation and increases the robustness of the features. The chrysanthemum feature map output by the pooling layer is up-sampled to construct a four-layer spatial feature pyramid, and the different scale feature maps obtained are fused to obtain multi-scale information. The multi-scale information includes chrysanthemum edge, texture, and shape information.

[0089] In this embodiment, the fully connected layer and the softmax layer of the VGG16 convolutional neural network are removed, and an edge detection layer is added to extract the pistil contour described below.

[0090] The feature processing unit is configured to process the multi-scale information based on a context perception layer to obtain context features.

[0091] In this embodiment, the chrysanthemum feature map output by the pooling layer of the VGG16 convolutional neural network is subtracted from the multi-scale information output by the spatial feature pyramid to obtain new scale features; a preset context feature weight parameter is obtained, and the context perception layer is constructed based on the new scale features and the chrysanthemum feature map output by the pooling layer. The context feature acquisition formula is as follows:

[0092] ,

[0093] is the chrysanthemum feature map output by the pooling layer, is the preset context feature weight parameter, is the new scale feature.

[0094] The recognition unit is configured to obtain a chrysanthemum recognition result based on the context features; the chrysanthemum recognition result is a chrysanthemum density map.

[0095] a diameter calculation unit configured to, based on the chrysanthemum recognition result, extract a chrysanthemum pistil contour by edge detection, and obtain a chrysanthemum pistil diameter based on a pixel size of the chrysanthemum pistil contour and an image resolution;

[0096] In this embodiment, the Canny algorithm is used to perform contour monitoring on the chrysanthemum density map to obtain all closed contours, and the Hough transform is used to confirm the pistil contour in all closed contours. According to the resolution of the image (i.e. the number of pixels per inch or per centimeter, DPI or PPI), the pistil pixel size is converted into the actual size. For example, if the image resolution is 300 DPI and the measured pistil diameter is 100 pixels, the actual pistil diameter is about 1 / 3 inch or about 8.47 mm.

[0097] a counting unit configured to, based on the pistil diameter of the recognized chrysanthemum, screen chrysanthemums that meet a preset pistil diameter, and complete construction of a chrysanthemum recognition counting model.

[0098] In this embodiment, the improved manta ray foraging optimization algorithm is used to optimize the parameters of the chrysanthemum recognition counting model. First, in the individual initialization stage, the first half of the individuals are randomly initialized, and the second half of the individuals are initialized within the mean interval, which preserves the randomness of the population and avoids random distribution of the initialized individuals. Second, in the iterative optimization process of the optimization algorithm, the exponential weight coefficient is obtained based on the exponential change of the iteration number. According to the initialized population individuals and the obtained exponential weight coefficient, the iterative optimization of the model is performed to find the optimal recognition effect of the chrysanthemum recognition counting model, and the construction of the training weight of the chrysanthemum recognition counting model is completed.

[0099] The model is judged for convergence. If it converges, the constructed training weight is used for testing samples to test the accuracy of the chrysanthemum recognition counting model. If it does not converge, the parameters are adjusted and iteratively optimized until the model converges.

[0100] The model convergence refers to whether the convolutional neural network can learn the information of the chrysanthemum target features well and complete the learning. The loss can be used to feedback the standard of training model convergence. When the network validation loss and training loss change little and the difference is small, and tend to be stable, it indicates that the network has learned the features of the chrysanthemum target well. When the validation loss gradually becomes greater than the training loss, it indicates that the network has overfitting, and the optimal iteration number of network iteration is the point of separation.

[0101] The evaluation criteria of the constructed model are as follows: the constructed training weight is used for verification in the verification set, and the accuracy (P), recall rate (R), and average accuracy (mAP) are used as evaluation indexes, and the larger the value is, the better the model performance is. P is the proportion of the model prediction correct in the model prediction result; R is the proportion of the prediction correct in all true values; mAP is the proportion of all prediction correct and all prediction. The calculation formulas of P, R and mAP are as follows:

[0102]

[0103]

[0104]

[0105] Wherein, TP is the number of positive samples predicted as positive samples; FP is the number of negative samples predicted as positive samples; FN is the number of positive samples predicted as negative samples; N represents the number of types of chrysanthemums detected, only one type of chrysanthemum is studied, so N is equal to 1.

[0106] The constructed model can realize inputting picture samples, outputting the number of chrysanthemums with standard diameter of pistil in the limited area and the pixel coordinates of the feature points obtained when labeling, which are used for conversion and calculation of the roller load and the actual height of chrysanthemums at the next moment.

[0107] The control parameter acquisition module is used to obtain the self-propelled chrysanthemum intelligent harvesting machine control parameters based on the number of chrysanthemums.

[0108] Further, the control parameter acquisition module comprises: an optimal speed acquisition unit, a comb tooth spacing acquisition unit and a header height acquisition unit.

[0109] The optimal speed acquisition unit is used to calculate the roller load based on the number of chrysanthemums, and obtain the optimal speed of the comb tooth based on the roller load.

[0110] In this embodiment, the error threshold interval [min, max] of predicting the roller load at the next moment and the rated load of the roller is set;

[0111] According to the number of chrysanthemums of a certain size, the roller load at the next moment is predicted, and the error with the rated load of the roller is calculated. When the error is greater than max (i.e. the predicted roller load at the next moment is greater than the rated load), the roller speed is reduced until the error is in the interval [min, max]; when the error is less than min (i.e. the predicted roller load at the next moment is less than the rated load), the roller speed is increased until the error is in the interval [min, max]; when the error is in the interval [min, max], no adjustment is made.

[0112] The comb tooth spacing acquisition unit is configured to calculate an average diameter of the chrysanthemum satisfying the preset pistil diameter, and acquire the comb tooth spacing based on the average diameter.

[0113] The cutting table height acquisition unit is configured to acquire pixel coordinates of the highest point of the chrysanthemum satisfying the preset pistil diameter, and acquire the actual height information of the chrysanthemum based on the pixel coordinates of the highest point; and acquire the cutting table height adjustment parameter based on the actual height information of the chrysanthemum.

[0114] In the embodiment, the actual height of the chrysanthemum satisfying the preset pistil diameter, which is recognized by the chrysanthemum recognition counting model, is calculated; the threshold interval [h min , h max ] of the distance of the chrysanthemum picking point from the ground is preset according to the positional relationship between the chrysanthemum picking point and the actual height of the chrysanthemum by means of artificial design criteria;

[0115] The height of the cutting table from the ground is measured by means of an ultrasonic sensor; the height e of the comb tooth cutting point from the ground is calculated based on the comb tooth radius; and the cutting table displacement distance is calculated by means of the acquired cutting table height h4.

[0116] In the embodiment, according to the maximum and minimum height interval [h min , h max ] of the distance of the preset picking point from the ground, the height e of the comb tooth cutting point from the ground, when e is located in the interval [h min , h max ], the cutting table displacement distance d can be 0, if e<= h min , d=[(h min +h max ) / 2]-e>0, the cutting table is raised, if e>= h max , d=[(h min +h max ) / 2]-e<0, the cutting table is lowered. The cutting table is raised and lowered by means of the size of the error e, so that the cutting knife is as possible as in the midpoint of the picking interval [h min , h max ], and the cutting table height h4 is controlled in real time. As shown in FIG. Figure 10 The inflorescence width is h1; the depth of the working device penetrating the inflorescence layer is h2; the flower height is h3; and the cutting table height is h4.

[0117] The control module is configured to adjust the self-propelled chrysanthemum intelligent harvesting machine based on the control parameter, and complete the chrysanthemum harvesting. The adjustment module is equivalent to an industrial computer.

[0118] Embodiment three

[0119] The application further provides an automatic control method of the self-propelled chrysanthemum intelligent harvesting machine, which applies the automatic control system and comprises the following steps:

[0120] Based on binocular camera to collect chrysanthemum image, and process chrysanthemum image;

[0121] The processed chrysanthemum image is input into a chrysanthemum recognition counting model to obtain the number of chrysanthemums meeting the preset pistil diameter;

[0122] Based on the number of chrysanthemums, obtain the regulation parameters of the self-propelled chrysanthemum intelligent harvesting machine;

[0123] Based on the regulation parameters, adjust the self-propelled chrysanthemum intelligent harvesting machine to complete the chrysanthemum harvesting.

[0124] The above-described embodiments are only descriptions of the preferred modes of the present application and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. A self-propelled chrysanthemum intelligent harvesting machine, characterized in that, The application relates to a harvesting header and a four-wheel mobile chassis articulated with the harvesting header through a header fixing articulation device, wherein a chrysanthemum collecting device is installed on the four-wheel mobile chassis, and an air conveying channel is connected to the chrysanthemum collecting device. The harvesting header comprises a lifting mechanism and a cutter mechanism which are installed on a header side guard plate. A power module is fixed on the lifting mechanism. The power module drives a roller driving module installed on both ends of a roller. A variable-spacing comb plate device and a fixed comb plate device are installed on the roller. Double cameras are symmetrically arranged on both sides of a harvesting header shell. The variable-spacing comb plate device is composed of a gear shaft, a rack, a rack fixing frame, a cam, a small comb plate I, a small comb plate II, a thrust bearing, a ring linear motor, a variable-spacing support frame, a guide pipe base and a guide rod. Roller driving cover plates are arranged on both ends of the roller, and a ring linear motor is arranged in the roller driving cover plates. One end of the ring linear motor is provided with a thrust bearing, and one end of the thrust bearing is provided with a variable-spacing support frame. Four groups of racks which are spaced 90 degrees apart are arranged on the variable-spacing support frame.

2. The self-propelled chrysanthemum intelligent harvester according to claim 1, characterized in that, Each rack is provided with 20 equidistant gear shafts.

3. The self-propelled chrysanthemum intelligent harvester according to claim 2, characterized in that, The tail end of each gear shaft is provided with a cam. The cam is arranged in the small comb plate I and the small comb plate II. Guide rods are arranged in the end holes of the small comb plate I and the small comb plate II.

4. The self-propelled chrysanthemum intelligent harvester according to claim 2, characterized in that, The average diameter of chrysanthemums meeting a preset pistil diameter is calculated, and the comb spacing of the variable-spacing comb plate device is adjusted according to the average diameter. The power module is used for providing power for the harvesting header.

5. The self-propelled chrysanthemum intelligent harvester according to claim 2, characterized in that, The power module comprises a motor mounting cover plate and a driving motor installed on the motor mounting cover plate.

6. An automatic control system of a self-propelled chrysanthemum intelligent harvesting machine, for controlling the self-propelled chrysanthemum intelligent harvesting machine according to any one of claims 1-5, characterized in that, The roller driving module comprises a roller driving cover plate, a driven gear and a driving gear. The roller driving cover plate is fixedly installed on the driven gear. The driving gear is driven by the driving motor installed on the motor mounting cover plate. The lifting mechanism comprises a lifting bottom plate, a rolling slider, a lifting guide rail and a guide rail fixing block. The lifting bottom plate is fixedly installed on the rolling slider. The rolling slider is installed on the lifting guide rail. The lifting guide rail is installed in the guide rail fixing block. The lifting mechanism is installed on the header side guard plate through two groups of upper and lower guide rail fixing blocks. The cutter mechanism comprises a reciprocating cutter, an electric push rod and an electric push rod fixing block. The cutter mechanism is installed on the header side guard plate through the electric push rod fixing block. The application relates to a self-propelled chrysanthemum intelligent harvesting machine. The machine comprises an image acquisition module, an image processing module, a chrysanthemum recognition module, a control parameter acquisition module and a control module. The image acquisition module is used for acquiring chrysanthemum images based on the double cameras. The image processing module is used for processing the chrysanthemum images. The chrysanthemum recognition module is used for inputting the processed chrysanthemum images into a chrysanthemum recognition counting model to obtain the number of chrysanthemums meeting a preset pistil diameter. The control parameter acquisition module is used for obtaining self-propelled chrysanthemum intelligent harvesting machine control parameters based on the number of chrysanthemums. The control module is used for adjusting the self-propelled chrysanthemum intelligent harvesting machine based on the control parameters to complete chrysanthemum harvesting.

7. The automatic control system of the self-propelled chrysanthemum intelligent harvester according to claim 6, characterized in that, The image processing module comprises a decomposition unit, a reflection component adjusting unit, an illumination component adjusting unit and an image enhancement unit; The decomposition unit is configured to decompose the collected natural light chrysanthemum image to obtain a high-light chrysanthemum image and a low-light chrysanthemum image, and obtain a reflection component, an illumination component and a brightness adjustment ratio map corresponding to each of the high-light chrysanthemum image and the low-light chrysanthemum image; The reflection component adjusting unit is configured to splice the illumination component and the reflection component of the low-light chrysanthemum image, use the illumination component to guide the reflection component to be denoised, and use the reflection component of the high-light chrysanthemum image as a reference image to obtain a denoised reflection component; The illumination component adjusting unit is configured to splice the illumination component of the low-light chrysanthemum image with the brightness adjustment ratio map, and use the illumination component of the high-light chrysanthemum image as a reference image to obtain an adjusted illumination component; The image enhancement unit is configured to fuse the denoised reflection component and the adjusted illumination component to obtain an enhanced chrysanthemum image, thereby completing the processing of the chrysanthemum image.

8. The automatic control system of the self-propelled chrysanthemum intelligent harvester according to claim 7, characterized in that, The chrysanthemum recognition module comprises a feature extraction unit, a feature processing unit, a recognition unit, a diameter calculation unit and a counting unit; The feature extraction unit is configured to use a VGG16 convolutional neural network as a feature extraction network to perform multi-scale feature extraction on the enhanced chrysanthemum image to obtain multi-scale information; The feature processing unit is configured to process the multi-scale information based on a context perception layer to obtain context features; The recognition unit is configured to obtain a chrysanthemum recognition result based on the context features; The diameter calculation unit is configured to use edge detection to extract a chrysanthemum pistil contour based on the chrysanthemum recognition result, and obtain a chrysanthemum pistil diameter based on a pixel size of the chrysanthemum pistil contour and an image resolution; The counting unit is configured to filter chrysanthemums that meet a preset pistil diameter based on the recognized chrysanthemum pistil diameter, and complete the construction of a chrysanthemum recognition counting model.

9. The automatic control system of the self-propelled chrysanthemum intelligent harvester according to claim 6, characterized in that, The control parameter acquisition module comprises an optimal rotating speed acquisition unit, a comb tooth spacing acquisition unit and a cutting platform height acquisition unit; The optimal rotating speed acquisition unit is configured to calculate a roller load based on the number of chrysanthemums, and obtain a comb tooth optimal rotating speed based on the roller load; The comb tooth spacing acquisition unit is configured to calculate an average diameter of chrysanthemums that meet the preset pistil diameter, and obtain a comb tooth spacing based on the average diameter; The cutting platform height acquisition unit is configured to obtain a pixel coordinate of a highest point of chrysanthemums that meet the preset pistil diameter, and obtain chrysanthemum actual height information based on the pixel coordinate of the highest point; and obtain a cutting platform height adjustment parameter based on the chrysanthemum actual height information.

10. The automatic control method of the self-propelled chrysanthemum intelligent harvesting machine, the automatic control system of any one of claims 6-9 is applied, characterized in that, The method comprises the following steps: collecting a chrysanthemum image based on a binocular camera, and processing the chrysanthemum image; inputting the processed chrysanthemum image into a chrysanthemum recognition counting model to obtain a number of chrysanthemums that meet a preset pistil diameter; obtaining self-propelled chrysanthemum intelligent harvester control parameters based on the number of chrysanthemums; adjusting the self-propelled chrysanthemum intelligent harvester based on the control parameters to complete chrysanthemum harvesting.

Images