Depth control system and method for seeder in hilly and mountainous areas

Abstract

The invention relates to the technical field of automatic control, in particular to a depth control system and method for a seeder in a hilly and mountainous area, and the system comprises a visual module which is configured to collect a visual image of a to-be-sown area according to a first preset frequency; a data processing module configured to: convert the visual image into a depth map using an image processing model; performing gridding processing on the visual image; for each column of grids, determining at least one piece of target surface elevation data corresponding to the grid based on the depth map; determining at least one first adjustment instruction based on the preset seeding depth and the at least one piece of target surface elevation data; and the control instruction output module is configured to issue at least one first adjusting instruction corresponding to each column of grids to the corresponding depth limiting adjusting mechanism. According to the system, independent and accurate control over the ditching depth of each ditcher can be achieved through the visual preview and intelligent control technology, so that the seedling emergence uniformity and the crop yield are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automatic control, and particularly relates to a hill and mountain area seeding machine depth control system and method. BACKGROUND

[0002] In agricultural production, the seeding depth has a crucial influence on the seed germination rate, seedling uniformity and growth and development of crops in later period. In the hill and mountain area, the terrain is complex and changeable, and the landform has large fluctuation. Most of the seeding machines on the market adopt mechanical depth limiting structure to roughly determine the seeding depth through depth limiting wheels. However, when the seeding machine is operated in the hill and mountain area, the pressure on the depth limiting wheels is different at different positions due to the uneven ground, which leads to large difference in the seeding depth. For example, at the low-lying place, the depth limiting wheels sink more, the seeding depth is too deep, and the seed may be difficult to germinate due to lack of oxygen; at the higher place, the pressure of the depth limiting wheels on the ground is small, the seeding is too shallow, the seed is easily affected by the external environment, and the water loss is fast, which is also not conducive to germination.

[0003] It can be seen that the traditional seeding machine is difficult to ensure the consistency of the seeding depth in the complex terrain such as the hill and mountain area. SUMMARY

[0004] The embodiments of the present application provide a hill and mountain area seeding machine depth control system and method to solve the problem that the seeding machine is difficult to ensure the consistency of the seeding depth in the complex terrain such as the hill and mountain area.

[0005] In a first aspect, the embodiments of the present application provide a hill and mountain area seeding machine depth control system, which is applied to a seeding machine. The seeding machine comprises a plurality of furrow openers arranged side by side and a depth limiting wheel assembly corresponding to each of the furrow openers. The depth limiting wheel assembly comprises at least a depth limiting wheel and a depth limiting adjustment mechanism. The depth limiting wheel is used to roll along the ground surface to cooperate with the furrow opener to limit the furrowing depth. The depth limiting adjustment mechanism is used to adjust the installation height of the depth limiting wheel and thus adjust the furrowing depth of the furrow opener. The system comprises a visual module configured to collect a visual image of a to-be-seeded area at a first preset frequency. The visual image covers at least the furrowing range of the furrow opener. A data processing module is configured to convert the visual image into a depth map by using an image processing model. The depth map comprises ground elevation data. The visual image is subjected to grid processing. The size of the grid is determined according to the number of the furrow openers. The lateral size of the grid is equal to the number of the furrow openers. For each column of grid, at least one target ground elevation data corresponding to the column of grid is determined based on the depth map. At least one first adjustment instruction is determined based on a preset seeding depth and the at least one target ground elevation data. The first adjustment instruction comprises an installation height adjustment value. A control instruction output module is configured to send the at least one first adjustment instruction corresponding to each column of grid to the corresponding depth limiting adjustment mechanism, so that the depth limiting adjustment mechanism adjusts the installation height of the depth limiting wheel based on the first adjustment instruction, and thus adjusts the furrowing depth of the furrow opener.

[0006] In a possible implementation, the visual image further comprises two preset markers arranged at the lateral edges of the seeding machine. The data processing module is specifically configured to determine two preset marker points in the visual image. The preset marker points are imaging points of the preset markers in the visual image. The pixel coordinates of the two preset marker points and a preset geometric constraint function are used to determine two auxiliary marker points in the visual image. The two preset marker points and the two auxiliary marker points are used to form a trapezoid. The connecting line of the two preset marker points forms the first base of the trapezoid, and the connecting line of the two auxiliary marker points forms the second base of the trapezoid. The first base is the long base and the second base is the short base. A target analysis area is determined in the visual image. The target analysis area is a trapezoidal area enclosed by the two preset marker points and the two auxiliary marker points. The two preset marker points and the two auxiliary marker points are used as the vertices of the grid to perform grid processing on the target analysis area.

[0007] In a possible implementation, the data processing module is further configured to: determine, by using the depth map, first ground elevation data corresponding to each grid cell in each column grid in the target analysis area; determine, for each column grid, a first average value and a first standard deviation of the first ground elevation data corresponding to the column grid; and determine, for each column grid, N target grid cells by traversing each grid cell corresponding to the column grid, wherein the first ground elevation data of the target grid cells satisfy the following first formula: | x + t - μ | > σ; wherein x represents the first ground elevation data, t represents an adjustment parameter, t > 0, μ represents the first average value, and σ represents the first standard deviation.

[0008] In a possible implementation, the data processing module is further configured to: in a case where N = 0, divide all grid cells in a current column grid into a same cell group for each column grid; in a case where N = 1, if a sequence index P of a target grid cell is located in an interval [L-K, L], divide the current column grid into two cell groups, and a division point is located before an index position of the target grid cell, for each column grid; wherein L is a sequence length of the current column grid, and K is a preset positive integer; otherwise, divide the current column grid into three cell groups, the first cell group includes the first to the P-M-1th grid cells, the second cell group includes the P-Mth to the P+Mth grid cells, and the third cell group includes the P+M+1th to the Lth grid cells, wherein M is a preset positive integer.

[0009] In a possible implementation, the data processing module is further configured to: in a case where N > 1, determine whether there is a continuous target grid cell for each column grid; wherein an index difference between any adjacent target grid cells in the continuous target grid cell is not greater than a first preset threshold; in a case where there is no continuous target grid cell, divide all grid cells in the current column grid into a same cell group; in a case where there is a continuous target grid cell, divide each continuous target grid cell and grid cells covered by the continuous target grid cell into a same cell group, divide continuous grid cells not covered by the continuous target grid cell into a same cell group, and divide independent grid cells not covered by the continuous target grid cell and adjacent target grid cells of the independent grid cells into a same cell group.

[0010] In a possible implementation, the system further comprises: a height data collection module configured to collect the real-time installation height of each depth limiting wheel at a first preset frequency; a speed data collection module configured to collect the real-time running speed of the seeding machine at the first preset frequency; the data processing module is further configured to: for each grid, calculate the average value of each first ground elevation data corresponding to at least one unit group to obtain at least one target ground elevation data; for each grid, determine at least one installation height adjustment value based on the latest collected real-time installation height, the at least one target ground elevation data and / or the preset seeding depth; for each grid, determine the instruction issuing time stamp corresponding to each unit group based on the latest collected real-time running speed, the current system time and / or the starting index of the unit group; and generate at least one first adjustment instruction based on the at least one installation height adjustment value and the instruction issuing time stamp.

[0011] In a possible implementation, the first adjustment instruction further comprises the instruction issuing time stamp; and the control instruction output module is specifically configured to: receive the at least one first adjustment instruction; and for each grid, issue the at least one first adjustment instruction to the corresponding depth limiting adjustment mechanism based on the instruction issuing time stamp.

[0012] In a possible implementation, the seeding machine further comprises a pressure sensor installed on each depth limiting wheel assembly; and the system further comprises: a pressure data collection module configured to collect first pressure data of each pressure sensor at a second preset frequency; and the data processing module is further configured to: for each pressure sensor, periodically aggregate a plurality of first pressure data at a third preset frequency to form a data set; wherein the third preset frequency is less than the second preset frequency; for each pressure sensor, calculate the second standard deviation of the data set corresponding thereto; and generate an alarm information in a case where the second standard deviation is greater than a second preset threshold.

[0013] In a second aspect, the embodiments of the present application provide a hill and mountain area seeding machine depth control method, applied to a seeding machine, the seeding machine comprising: a plurality of furrowers arranged side by side and a depth limiting wheel assembly corresponding to each furrower, the depth limiting wheel assembly comprising at least a depth limiting wheel and a depth limiting adjusting mechanism, the depth limiting wheel being used to roll along the ground surface to cooperate with the furrower to limit the furrowing depth, and the depth limiting adjusting mechanism being used to adjust the installation height of the depth limiting wheel and further adjust the furrowing depth of the furrowers; the method comprising: acquiring a visual image of a region to be seeded according to a first preset frequency; wherein the field of view of the visual image covers at least the furrowing range of the furrowers; converting the visual image into a depth map by using an image processing model; wherein the depth map comprises ground elevation data; performing grid processing on the visual image; wherein the size of the grid is determined according to the number of furrowers, and the lateral size of the grid is equal to the number of furrowers; for each column of grids, determining at least one target ground elevation data corresponding thereto based on the depth map; and determining at least one first adjusting instruction based on the preset seeding depth and the at least one target ground elevation data, the first adjusting instruction comprising an installation height adjustment value; and delivering the at least one first adjusting instruction corresponding to each column of grids to the corresponding depth limiting adjusting mechanism, so that the depth limiting adjusting mechanism adjusts the installation height of the depth limiting wheel based on the first adjusting instruction, and further adjusts the furrowing depth of the furrowers.

[0014] In a third aspect, the embodiments of the present application provide an electronic device, comprising: one or more processors; and a memory configured to store one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the hill and mountain area seeding machine depth control method in the foregoing second aspect and various implementation manners thereof.

[0015] From the above, the embodiment of the present application provides a hilly and mountainous area seeding machine depth control system and method. The system comprises: a visual module configured to collect visual images of a to-be-seeded area according to a first preset frequency; wherein the visual image capturing range covers at least the furrowing range of the furrower; a data processing module configured to convert the visual image into a depth map by using an image processing model; wherein the depth map comprises ground elevation data; the visual image is subjected to grid processing; wherein the size of the grid is determined according to the number of the furrower, and the lateral size of the grid is equal to the number of the furrower; for each column of grid, at least one target ground elevation data corresponding thereto is determined based on the depth map; and based on the preset seeding depth and the at least one target ground elevation data, at least one first adjustment instruction is determined, and the first adjustment instruction comprises an installation height adjustment value; a control instruction output module configured to issue the at least one first adjustment instruction corresponding to each column of grid to the corresponding depth limiting adjustment mechanism, so that the depth limiting adjustment mechanism adjusts the installation height of the depth limiting wheel based on the first adjustment instruction, and further adjusts the furrowing depth of the furrower. The system can realize independent, accurate and adaptive adjustment of the furrowing depth of each furrower by using visual pre-look and intelligent control technology, effectively overcome the adverse effects of hilly and mountainous terrain undulations on the uniformity of seeding depth, create good conditions for seed germination and crop growth, and ultimately achieve the technical effect of improving the uniformity of emergence and crop yield. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A structure schematic diagram of a seeding machine provided by the embodiment of the present application is provided. Figure 2 A structure schematic diagram of a hilly and mountainous area seeding machine depth control system provided by the embodiment of the present application is provided. Figure 3 A working schematic diagram of a hilly and mountainous area seeding machine depth control system provided by the embodiment of the present application is provided. Figure 4 A schematic diagram of a visual image provided by the embodiment of the present application is provided. Figure 5 A schematic diagram of grid processing of a visual image provided by the embodiment of the present application is provided. Figure 6 A schematic diagram of grid division provided by the embodiment of the present application is provided. Figure 7 A flow schematic diagram of a hilly and mountainous area seeding machine depth control method provided by the embodiment of the present application is provided. Figure 8 A structure schematic diagram of an electronic device provided by the embodiment of the present application is provided.

[0017] Wherein, 10-processor; 20-memory; 100-gutter opener; 200-depth limiting wheel assembly; 201-depth limiting wheel; 202-depth limiting adjustment mechanism; 2021-adjusting motor; 2022-depth limiting wheel support; 2023-depth adjusting screw; 2024-depth limiting block; 300-pressure sensor; 4001-visual module; 4002-data processing module; 4003-control instruction output module; 4004-height data acquisition module; 4005-speed data acquisition module; 4006-pressure data acquisition module. DETAILED DESCRIPTION

[0018] In order for those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should fall within the scope of protection of the present application.

[0019] In order to solve the problem that the traditional seeding machine is difficult to ensure the consistency of the seeding depth in complex terrains such as hilly and mountainous areas, the present application provides a hilly and mountainous area seeding machine depth control system and method. The system and method can determine the front ground elevation data in real time through visual perception, realize the dynamic preview adjustment of the depth limiting wheel through the intelligent analysis and prediction of the data processing module, complete the terrain self-adaptation and seeding depth control without manual intervention, fundamentally avoid the problem of over-deep or over-shallow seeding caused by the terrain fluctuation of the traditional mechanical depth limiting method, and ensure the uniformity of the seeding depth and the uniformity of the emergence.

[0020] Figure 1 The structure diagram of the seeding machine provided in the embodiments of the present application is shown.

[0021] As shown in Figure 1 The present application provides a seeding machine, which at least comprises a plurality of gutter openers 100 arranged side by side and a depth limiting wheel assembly 200 corresponding to each of the gutter openers 100. The gutter opener 100 can have a sharp soil-entering end such as a wedge shape. When working, the gutter opener 100 cuts into the soil under the action of the traction force of the seeding machine, and extrudes the soil to both sides through the soil-entering end, thereby forming a seed furrow (or ground ridge) in the field.

[0022] Further, the depth limiting wheel assembly 200 at least comprises a depth limiting wheel 201 and a depth limiting adjusting mechanism 202. The depth limiting adjusting mechanism 202 can comprise an adjusting motor 2021, a depth limiting wheel support 2022, a depth adjusting screw 2023 and a depth limiting block 2024; the depth limiting wheel 201 is connected with the depth limiting wheel support 2022, the depth limiting block 2024 is sleeved on the depth adjusting screw 2023 and forms a threaded pair with the depth adjusting screw 2023. When the depth limiting wheel 201 contacts with the ground, the depth limiting wheel 201 can abut against the depth limiting block 2024, at this time, the depth limiting block 2024 can play a limiting role. When the adjusting motor 2021 drives the depth adjusting screw 2023 to rotate, the relative position of the depth adjusting screw 2023 and the depth limiting block 2024 changes, then, the relative height of the abutting position of the depth limiting wheel 201 and the depth limiting block 2024 changes. In this way, the effect of adjusting the installation height of the depth limiting wheel 201 can be achieved, and the installation height here refers to the vertical distance between the contact point of the depth limiting wheel 201 and the fixed frame of the seeding machine or the depth limiting wheel support 2022.

[0023] It can be seen that, in the embodiment of the present application, the depth limiting wheel 201 is used to roll along the ground surface to cooperate with the furrow opener 100 to limit the furrowing depth, and the depth limiting adjusting mechanism 202 is used to adjust the installation height of the depth limiting wheel 201 to further adjust the furrowing depth of the furrow opener 100.

[0024] In addition, the seeding machine further comprises a pressure sensor 300 installed on each depth limiting wheel assembly. In actual application, the pressure sensor 300 can be installed at a predetermined position of the depth limiting block 2024 by means of a mounting clamp or by screw connection and the like. When the depth limiting block 2024 plays a limiting role on the depth limiting wheel 201, the pressure sensor 300 can collect the contact force between the depth limiting wheel 201 and the depth limiting block 2024, and further obtain the pressure data of the contact between the depth limiting wheel 201 and the ground.

[0025] In some implementations, the seeding machine further comprises a traction device or a suspension mechanism, and a seed sowing device corresponding to the furrow opener 100. The traction device or the suspension mechanism is used to reliably connect with a power equipment such as a tractor, and further provides forward power for the seeding machine. The seed sowing device is used to carry seeds and accurately and quantitatively guide the seeds into the seed furrow opened by the furrow opener 100.

[0026] Figure 2 A structural schematic diagram of a depth control system of a hilly and mountainous area seeding machine provided by the embodiment of the present application.

[0027] Figure 3 A working schematic diagram of a depth control system of a hilly and mountainous area seeding machine provided by the embodiment of the present application.

[0028] As Figure 2 and Figure 3As shown, the embodiment of the present application provides a hilly and mountainous area seeding machine depth control system, which is applied to a seeding machine. The system can include a visual module 4001, a data processing module 4002, a control instruction output module 4003, a height data acquisition module 4004, a speed data acquisition module 4005, and a pressure data acquisition module 4006. Among them, the visual module 4001, the control instruction output module 4003, the height data acquisition module 4004, the speed data acquisition module 4005, and the pressure data acquisition module 4006 can be in communication connection with the data processing module 4002. The visual module 4001 is configured to perform the following step S100.

[0029] S100: Collect visual images of the area to be seeded according to a first preset frequency; wherein the visual image range covers at least the range of the furrow opener 100.

[0030] Among them, the first preset frequency can be 5Hz to 10Hz, that is, 5 to 10 times per second, to adapt to the working speed of the seeding machine.

[0031] Further, as shown in (a) of Figure 3 The area that has not been seeded or furrowed can be referred to as the area to be seeded or the area to be furrowed, and the area that has been seeded or furrowed can be referred to as the seeded area or the furrowed area. The embodiment of the present application can use binocular cameras or structured light depth cameras and other imaging devices to collect visual images, as shown in (b) of Figure 3 The imaging device range can cover an area 2 to 5 meters in front of the seeding machine to ensure enough preview distance to budget the installation height of the depth wheel 201.

[0032] The data processing module 4002 is configured to perform the following steps S200-S400.

[0033] S200: Convert the visual image into a depth map using an image processing model; wherein the depth map includes ground elevation data.

[0034] In the embodiment of the present application, the image processing model can be a monocular depth estimation model based on deep learning, such as the MiDaS model, and the image processing model can be pre-trained for agricultural scene adaptation. Further, the ground elevation data of the depth map is in units of meters (m) or centimeters (cm), which is used to accurately reflect the terrain undulations of the area to be seeded.

[0035] Specifically, the ground elevation data can represent the vertical height value of a certain point on the ground relative to a fixed reference surface (such as the lower surface of the fixed frame of the seeding machine, the current coordinate system origin of the seeding machine). In the depth map, each pixel or grid cell corresponds to an elevation value, and the larger the value, the higher the ground at that point, and the smaller the value, the lower the ground, which quantifies the terrain undulation. It can be understood that the embodiments of the present application can pre-train the image processing model based on the preset visual image, the preset depth map and the corresponding ground elevation data of the depth map, so that the image processing model can directly convert the real-time visual image into a depth map, while outputting the ground elevation data corresponding to the depth map.

[0036] In some implementations, the depth map can reflect the elevation value of the ground in terms of pixel color brightness or depth: the darker the color, the lower the elevation value (terrain depression), and the lighter the color, the higher the elevation value (terrain protrusion). Through this intuitive visual mapping relationship, the system can quickly identify the terrain undulation features. For example, the embodiments of the present application can use a gray scale or pseudo-color representation, in which the gray scale uses different gray scales to represent elevation changes, and the pseudo-color map more intuitively highlights the micro-terrain differences through a blue-green-yellow-red color gradient spectrum.

[0037] S300: performing grid processing on the visual image.

[0038] The size of the grid is determined according to the number of furrowers 100, and the lateral size of the grid is equal to the number of furrowers 100. That is, each column of grid corresponds to a furrower 100.

[0039] In some implementations, the longitudinal size of the grid can be greater than or equal to the number of furrowers 100, because along the direction of travel of the seeding machine, the area to be seeded appears as an extended strip in the viewfinder of the imaging device. This strip-shaped terrain feature requires the longitudinal size of the grid to cover enough of the area to be worked to ensure that the system can capture the terrain undulation information of the area to be reached by the furrower 100 in advance, and to reserve enough response time for the depth limiting adjustment mechanism 202; at the same time, a larger longitudinal size can also more completely reflect the continuous change trend of the ground elevation, avoiding the fragmentation of terrain features caused by a too short longitudinal grid, thereby improving the accuracy of the target ground elevation data.

[0040] For example, the number of furrowers 100 of a small seeding machine can be 2-6, and the number of grids in the lateral direction and the longitudinal direction can also be 2-6. The number of furrowers 100 of a medium-sized seeding machine can be 6-12, and the number of grids in the lateral direction and the longitudinal direction can also be 6-12. The number of furrowers 100 of a large-sized seeding machine can be 12-48, and the number of grids in the lateral direction can be consistent with the number of furrowers 100, and the number of grids in the longitudinal direction can be greater than the number of furrowers 100. For example, when the number of furrowers 100 is 20, the number of grids in the lateral direction can be 20, and the number of grids in the longitudinal direction can be 20, 25, or 30.

[0041] S400: For each column of grids, determining at least one target ground elevation data corresponding thereto based on the depth map; and determining at least one first adjustment instruction based on the preset seeding depth and the at least one target ground elevation data, the first adjustment instruction comprising an installation height adjustment value.

[0042] When working on complex terrains such as soil slopes, it is necessary to keep the relative height difference between the depth-limiting wheel 201 and the furrower 100 constant, and this height difference should be equal to the target furrowing depth (i.e., the preset seeding depth). If the soil slope is uphill, in order to ensure consistent furrowing depth, the depth-limiting wheel 201 needs to be appropriately raised relative to the furrower 100 (moved upward relative to the frame), that is, the installation height of the depth-limiting wheel 201 is increased, that is, the installation height adjustment value should be a negative value, and the value needs to match the ground height variation corresponding to the uphill slope, to ensure that the furrower 100 cuts into the soil at a consistent depth. If the soil slope is downhill, the depth-limiting wheel 201 needs to be appropriately lowered relative to the furrower 100 (moved downward relative to the frame), that is, the installation height of the depth-limiting wheel 201 is decreased, that is, the installation height adjustment value should be a positive value, and the value needs to be adapted to the ground height drop corresponding to the downhill slope, to compensate for the terrain difference through the sinking of the depth-limiting wheel 201, and to avoid over-deep or over-shallow furrowing of the furrower 100 due to terrain inclination.

[0043] In the embodiments of the present application, each column of grids corresponds to a specific furrower 100, and each furrower 100 corresponds to a specific depth-limiting wheel assembly 200, so that the depth-limiting adjustment mechanism 202 in each depth-limiting wheel assembly 200 has a corresponding relationship with the first adjustment instruction. This corresponding relationship ensures that the control instruction can accurately act on the target execution unit.

[0044] The step of determining at least one first adjustment instruction based on at least one target ground elevation data will be described in detail below, and will not be described here.

[0045] Further, the embodiment of the present application can establish a certain quantitative relationship between the installation height adjustment value (Δh) and the rotation number (N) of the adjusting motor 2021 through the mechanical parameters of the depth adjusting screw 2023. Moreover, the rotation number N of the adjusting motor 2021 required to be rotated can be carried in the first adjusting instruction.

[0046] It can be understood that the installation height adjustment value (Δh) can indicate the direction of adjustment through positive and negative signs. For example, when Δh is positive, it indicates that the depth limiting wheel 201 needs to be lowered; when Δh is negative, it indicates that the depth limiting wheel 201 needs to be raised. Correspondingly, the positive and negative signs of the rotation number (N) are used to control the rotation direction of the adjusting motor 2021. For example, when N is positive, the motor is controlled to rotate forward to raise the depth limiting wheel 201; when N is negative, the motor is controlled to rotate reversely to lower the depth limiting wheel 201.

[0047] The specific quantitative relationship formula is as follows: N = -Δh / O; Wherein, N represents the rotation number, Δh represents the installation height adjustment value, and O represents the pitch of the depth adjusting screw 2023.

[0048] For example, if the pitch O = 5 mm, and the depth limiting wheel 201 needs to be raised by 15 mm, then the first adjusting instruction can require the adjusting motor to rotate forward N times, N = 15 / 5 = 3, i.e. 3 times.

[0049] It should be further pointed out that the preset sowing depth can be 20-60 mm, and can have a preset error range of ±5 mm, which can be determined based on factors such as geology and seed type, and the embodiment of the present application does not make specific limitations.

[0050] Further, the control instruction output module 4003 is configured to perform the following step S500.

[0051] S500: The at least one first adjusting instruction corresponding to each column of grids is sent to the corresponding depth limiting adjusting mechanism 202, so that the depth limiting adjusting mechanism 202 adjusts the installation height of the depth limiting wheel 201 based on the first adjusting instruction, and further adjusts the furrowing depth of the furrower 100.

[0052] In the embodiments of the present application, the corresponding first adjustment instructions can be executed for each column of the grid (each opener 100 or each trench), and the number of first adjustment instructions for each column of the grid is not unique, and specifically, a plurality of continuous adjustment instructions can be issued to adapt to the terrain changes during the seeding machine running, to realize the whole process fine control of a single seed trench. For example, when it is detected that there is a bump in front and a depression behind, the system will first generate a first adjustment instruction for lifting, and immediately issue a first adjustment instruction for lowering after passing the bump, so as to ensure that the depth limiting wheel assembly 200 and the opener 100 can smoothly "follow" the terrain contour change, maintain a constant seeding depth in the length direction of the whole seed trench, and thus realize all-round accurate depth control in the complex operation environment of hilly areas.

[0053] It can be understood that the trenching depth here can refer to the vertical distance of the earth-penetrating tip of the opener 100 relative to the ground reference line (or ground reference surface). In complex terrains such as mountains and hills, the ground is rugged and uneven. In order to ensure the consistency of the seeding depth, the embodiments of the present application take the lowest part in the sowing area as the reference to construct a virtual ground reference line. Then, after the installation height of the depth limiting wheel 201 is adjusted, the depth of the earth-penetrating tip of the opener 100 relative to the ground reference line changes accordingly, so as to ensure that each opener 100 can form a seed trench with consistent depth even in the undulating terrain, guarantee the consistency of the seeding depth, and improve the seed germination rate and the uniformity of emergence.

[0054] It should be further emphasized that the core control strategy of the present system is distributed independent control, that is, each opener 100 and its corresponding depth limiting wheel assembly 200 form an independent control execution unit. The control instruction output module 4003 can independently and asynchronously issue the first adjustment instruction calculated for each column of the grid to the depth limiting adjustment mechanism 202 corresponding to the column of the grid. This makes different depth limiting wheels 201 can perform different lifting actions according to the terrain undulation in front of each, so as to jointly guarantee the depth consistency of all seeding rows on the rugged hilly area ground.

[0055] From the above, the embodiment of the present application provides a depth control system for a seeding machine in hilly and mountainous areas, which comprises: a visual module 4001 configured to collect a visual image of a region to be seeded at a first preset frequency; wherein the field of view of the visual image covers at least the furrowing range of the furrower 100; a data processing module 4002 configured to convert the visual image into a depth map using an image processing model; wherein the depth map comprises ground elevation data; the visual image is subjected to grid processing; wherein the size of the grid is determined according to the number of furrowers 100, and the lateral size of the grid is equal to the number of furrowers 100; for each column of grid, at least one target ground elevation data corresponding thereto is determined based on the depth map; and based on the preset seeding depth and the at least one target ground elevation data, at least one first adjustment instruction is determined, which comprises an installation height adjustment value; a control instruction output module 4003 configured to issue the at least one first adjustment instruction corresponding to each column of grid to the corresponding depth limiting adjustment mechanism 202, so that the depth limiting adjustment mechanism 202 adjusts the installation height of the depth limiting wheel 201 based on the first adjustment instruction, thereby adjusting the furrowing depth of the furrower 100. The system can realize independent, accurate and adaptive adjustment of the furrowing depth of each furrower 100 through visual preview and intelligent control technology, effectively overcome the adverse effects of the undulating terrain in hilly and mountainous areas on the uniformity of the seeding depth, create good conditions for seed germination and crop growth, and ultimately achieve the technical effects of improving the uniformity of emergence and crop yield.

[0056] Further, the field of view of the visual image also includes two preset calibration objects, which are arranged on the mechanical components at the lateral edges of the seeding machine. In actual application, the preset calibration objects can be colors that form high contrast with the field environment, such as bright fluorescent orange and bright yellow, to improve the imaging effect of the preset calibration objects in the visual image.

[0057] It should be further noted that the preset calibration objects should have a certain height to ensure that they have a suitable imaging position in the visual image. For example, the preset calibration objects can be reference rods installed on the two sides of the frame of the seeding machine, and the rod heads are provided with spherical objects. In this way, the preset calibration objects can directly and accurately calibrate the actual working width of the seeding machine, i.e., the physical distance between the two preset calibration objects is equal to the actual working width of the seeding machine, which provides an accurate reference for subsequent image calibration.

[0058] Further, the data processing module 4002 is specifically configured to perform the following steps S301-S304.

[0059] S301: Determine two preset calibration points in the visual image; wherein the preset calibration points are the imaging points of the preset calibration objects in the visual image.

[0060] Figure 4 A schematic diagram of a visual image provided for an embodiment of the present application.

[0061] As shown in Figure 3 and Figure 4 , due to the existence of perspective distortion and geometric distortion, the land of the sowing area appears in the visual image in a perspective convergent form of wide near and narrow far. The embodiment of the present application can determine the pixel position (pixel coordinates) of the preset calibration points in the visual image, and then establish the mapping relationship between the image and the actual physical space through the two preset calibration points.

[0062] S302: determining two auxiliary calibration points in the visual image by using the pixel coordinates of the two preset calibration points and a preset geometric constraint function; wherein the two preset calibration points and the two auxiliary calibration points are used to form a trapezoid, the connecting line of the two preset calibration points forms the first bottom side of the trapezoid, the connecting line of the two auxiliary calibration points forms the second bottom side of the trapezoid, and the first bottom side is the long bottom side and the second bottom side is the short bottom side.

[0063] Continuing to refer to Figure 4 , the embodiment of the present application can use the preset geometric constraint function to infer the two auxiliary calibration points far away from the two preset calibration points as a reference, so that the four calibration points enclose a trapezoidal area which is proportional to the actual terrain, and the area is the effective working area to be analyzed. Through this calibration method, a reliable spatial reference is provided for subsequent extraction of ground elevation data and generation of adjustment instructions.

[0064] The preset geometric constraint function can be a mathematical model based on the principle of perspective projection and homographic transformation, and the core is to calculate the pixel coordinates of the corresponding points far away from the known physical size (actual working width of the seeding machine) near the image.

[0065] The execution logic of the preset geometric constraint function includes: determining a predetermined front target distance D; based on the perspective projection relationship, mapping the known physical width W of the seeding machine to the image coordinate system according to the imaging device parameters (f, H, θ) and the target distance D through the perspective projection model, so as to calculate the corresponding auxiliary calibration point pixel coordinates. It should be noted that in the imaging device parameters, "f" represents the focal length of the imaging device, "H" represents the installation height of the imaging device, and "θ" represents the angle between the optical axis of the imaging device and the horizontal plane.

[0066] Specifically, the function is established as follows: The physical width of the seeding machine is W, the focal length of the imaging device is f, the installation height of the imaging device is H, and the angle between the optical axis of the imaging device and the horizontal plane is θ. According to the perspective projection relationship, the pixel width w(D) of the physical width at a horizontal distance D from the imaging device in the image satisfies: w(D) = (f·W) / (D·cosθ + H·sinθ); wherein w(D) represents the pixel width of the physical width at a horizontal distance D from the imaging device in the image, D is the imaging distance, f is the imaging focal length, W is the physical width of the seeding machine, H is the installation height of the imaging device, and θ is the angle between the optical axis of the imaging device and the horizontal plane.

[0067] In actual application, the physical width W of the seeding machine and the value of D can be pre-calibrated, for example, the physical width W of the seeding machine is 2.4 meters, and the value of D is 2 meters, 3 meters or 5 meters, and the specific value depends on the range of the imaging device. Alternatively, the physical width W can be calculated based on the visual measurement principle through the pixel coordinates of the two preset calibration points.

[0068] S303: determining a target analysis region in the visual image; wherein the target analysis region is a trapezoidal region enclosed by the two preset calibration points and the two auxiliary calibration points.

[0069] Continuing to refer to Figure 4 , it should be noted that the specific form of the target analysis region (trapezoidal region) is determined by the relative pose of the imaging device and the seeding machine and the terrain condition, and the target analysis region can be an isosceles trapezoid or a general trapezoid.

[0070] When the optical axis of the imaging device is aligned with the forward direction of the seeding machine, and the seeding machine maintains a horizontal state while working on the horizontal ground, the two preset calibration points are symmetrically distributed on both sides of the axis in the image, and at this time the target analysis region constructed according to the perspective projection principle presents an isosceles trapezoid. When the seeding machine produces significant lateral inclination during operation, or there is a horizontal deviation in the installation of the imaging device, the two preset calibration points are asymmetrically distributed in the image, and at this time the target analysis region constructed is a general trapezoid.

[0071] S304: performing grid processing on the target analysis region by taking the two preset calibration points and the two auxiliary calibration points as grid vertices.

[0072] Figure 5 A schematic diagram of the grid processing of the visual image provided by the embodiment of the present application is shown.

[0073] As Figure 5 shown, after the grid processing of the target analysis region, a plurality of columns of grids can be formed, Figure 5An exemplary illustration of each column grid and the grid cells contained in the column grid is shown.

[0074] In practical applications, the size of the grid can be adjusted adaptively based on actual conditions, which is not specifically limited in the embodiments of the present application.

[0075] Further, the data processing module 4002 is specifically configured to perform the following steps S305-S307.

[0076] S305: Determine the first ground elevation data corresponding to each grid cell in each column grid in the target analysis area by using the depth map.

[0077] It can be understood that the depth map contains depth information of each pixel, based on which the ground elevation data corresponding to each pixel can be determined, and based on this, the first ground elevation data corresponding to each grid cell in each column grid in the target analysis area can be further determined. Each grid cell can be composed of multiple pixels, and the average value of the ground elevation data of each pixel corresponding to the grid cell can be calculated as the first ground elevation data of the grid cell in the embodiments of the present application.

[0078] Further, in the depth map, the brighter the pixel color, the higher the ground elevation of the pixel, and the darker the pixel color, the lower the ground elevation of the pixel. The embodiments of the present application can determine the brightness value of each pixel corresponding to the target analysis area in the depth map and convert it into an elevation value d.

[0079] After that, the embodiments of the present application can convert each pixel point S(u, v) into a three-dimensional point S'(X, Y, Z) with the seeding machine body as the reference by coordinate transformation based on the elevation value d of each pixel and the imaging device intrinsic parameter, wherein "Z" is the absolute ground elevation of the pixel point. Further, the first ground elevation data of the grid cell can be determined through post-processing steps such as arithmetic mean or spatial interpolation algorithm.

[0080] S306: For each column grid, determine the first average value and the first standard deviation of each first ground elevation data corresponding thereto.

[0081] It can be understood that the first average value (μ) represents the overall altitude level of the corresponding ridge terrain of the column grid, and the first standard deviation (σ) is used to quantify the fluctuation degree of the ridge terrain. The larger the standard deviation, the more complex the terrain, and vice versa, the flatter the terrain.

[0082] S307: For each column grid, traverse each grid cell corresponding thereto to determine N target grid cells; wherein the first ground elevation data of the target grid cell satisfies the following first formula: |x + t - μ| > σ; wherein, x represents the first ground elevation data, t represents the adjustment parameter, t > 0, μ represents the first average value, and σ represents the first standard deviation.

[0083] In the embodiment of the present application, the first ground elevation data of the grid cell conforms to the first formula, which means that the first ground elevation data of the grid cell significantly deviates from the overall elevation level of the current column grid, and the grid cell can correspond to a significant undulating area such as a protrusion, a depression or a steep slope in the terrain. Such grid cells are marked as "target grid cells", which provide a core basis for subsequent cell group division, to ensure that the system can specifically process the seeding depth adjustment of the terrain mutation area.

[0084] In some implementations, the value range of t can be [0.5σ, 1.5σ], and σ represents the first standard deviation, or the value of t can be 10 millimeters or 20 millimeters, which is not limited in the embodiment of the present application.

[0085] Figure 6 The schematic diagram of grid division provided in the embodiment of the present application.

[0086] Further, as shown in Figure 6 , in the embodiment of the present application, the data processing module 4002 is further configured to perform the following steps S308-S310.

[0087] S308: For each column grid, in the case of N = 0, all grid cells in the current column grid are divided into the same cell group.

[0088] As shown in Figure 6 , in the case where there is no target grid cell, it indicates that the entire landform corresponding to the column grid is flat and has little undulation, and the embodiment of the present application can divide all grid cells of the column grid into the same cell group.

[0089] S309: For each column grid, in the case of N = 1, if the sequence index P of the target grid cell is located in the interval [L-K, L], the current column grid is divided into two cell groups, and the division point is located before the index position of the target grid cell; wherein, L is the sequence length of the current column grid, and K is a preset positive integer.

[0090] In the embodiment of the present application, the specific division point can be before the P-Kth grid cell. The value of K can be 2, 3 or 5, which is not limited in the embodiment of the present application.

[0091] As shown in Figure 6As shown in (b), K can be equal to 2, when the only abnormal point (target grid cell P) is located at the end of the column grid, embodiments of the present application can simplify it into two control stages: maintaining the original state before the seeding machine reaches the location corresponding to the target grid cell P, and starting local adjustment for the abnormal terrain in the end area.

[0092] S310: Otherwise, divide the current column grid into three cell groups, the first cell group includes the 1st to the (P-M-1)th grid cells, the second cell group includes the (P-M)th to the (P+M)th grid cells, and the third cell group includes the (P+M+1)th to the Lth grid cells, wherein M is a preset positive integer.

[0093] In embodiments of the present application, if the sequence index P of the target grid cell is not located in the interval [L-K, L], the current column grid can be divided into three cell groups. For example, the value of M can be less than or equal to the value of K, and the value of M can be 1 or 2, which can be dynamically adjusted according to the seeding machine speed, soil hardness and motor response time, and embodiments of the present application do not limit this.

[0094] As shown in (b), K can be equal to 2, when the only abnormal point (target grid cell P) is located at the end of the column grid, embodiments of the present application can simplify it into two control stages: maintaining the original state before the seeding machine reaches the location corresponding to the target grid cell P, and starting local adjustment for the abnormal terrain in the end area. Figure 6 As shown in (c), the value of M can be 2, when the only abnormal point (target grid cell P) is located in the middle of the column, embodiments of the present application divide it into three control intervals: "front section-middle section-rear section". The first cell group (including the 1st to the (P-M-1)th grid cells): This cell group corresponds to normal terrain. The second cell group (including the (P-M)th to the (P+M)th grid cells): This cell group corresponds to the abnormal protrusion or depression and its transition zone. The third cell group (including the (P+M+1)th to the Lth grid cells): This cell group corresponds to the normal terrain after the abnormal area.

[0095] Further, continuing to refer to Figure 6 The data processing module 4002 is further configured to perform the following steps S311-S313.

[0096] S311: For each column grid, in the case of N>1, determine whether there is a continuous target grid cell; wherein the index difference of any adjacent target grid cells in the continuous target grid cell is not greater than a first preset threshold.

[0097] It is worth noting that the continuous target grid cell can also be referred to as a "maximum continuous target grid cell group", so that multiple abnormal points located adjacent to each other can be identified as a continuous terrain feature (such as a continuous bump or ditch), rather than multiple isolated points, thereby generating a unified control instruction to avoid frequent starting and stopping of the actuator.

[0098] For example, the first preset threshold can be 2, 3 or 5, and this application embodiment does not specifically limit it.

[0099] S312: In the absence of continuous target grid cells, divide all grid cells in the current column into the same cell group.

[0100] like Figure 6 As shown in (d), when the number of target grid cells is relatively large (>1) and the target grid cells are discontinuous, the embodiments of this application can divide the column of grid cells into the same cell group. This ensures the smooth operation of the seeder.

[0101] For example, a grid column is set to have 15 grid cells (L=15), with sequence indices from 1 to 15, and a first preset threshold R=2. Through step S307, the system identifies 3 target grid cells in the column, with sequence indices of 2, 5, and 8 respectively.

[0102] Two target mesh cells with indices 2 and 5: index difference = 5 - 2 = 3 > 2, they are not contiguous; Two target grid cells with indices 5 and 8: index difference = 8 - 5 = 3 > 2, they are not contiguous.

[0103] In actual terrain, the above three target grid cells may correspond to small clods of soil, grass roots, etc.

[0104] S313: In the presence of continuous target grid cells, each continuous target grid cell and the grid cells it covers are divided into the same cell group, continuous grid cells not covered by continuous target grid cells are divided into the same cell group, and independent grid cells not covered by continuous target grid cells are divided into the same cell group with their adjacent target grid cells.

[0105] In some implementations, when the index of the target mesh cell is within [0, K] or [LK, L], embodiments of this application can group the 1st to Kth mesh cells and their adjacent target mesh cells into the same cell group, and group the LKth to Lth mesh cells and their adjacent target mesh cells into the same cell group. That is, although the 1st to Kth mesh cells and the LKth to Lth mesh cells are not covered by the target mesh cell, they can still be grouped into the same cell group as the target mesh cell. This reduces the number of cell groups and improves operational efficiency.

[0106] like Figure 6As shown in (e), a column of grids is set with 15 grid cells (L=15) and the sequence indexes are set from 1 to 15, and a first preset threshold R=2 is set. By the step S307, the system identifies 5 target grid cells in the column, and the sequence indexes are 2, 4, 6, 11, and 14 respectively.

[0107] The embodiment of the present application traverses the target grid cells [2, 4, 6, 11, 14] in the order of the indexes, and checks the index difference of the adjacent cells. The two target grid cells with indexes 2 and 4: index difference = 4-2 = 2 ≤ 2, and they are continuous. The two target grid cells with indexes 4 and 6: index difference = 6-4 = 2 ≤ 2, and they are continuous. The two target grid cells with indexes 6 and 11: index difference = 11-6 = 5 > 2, and they are not continuous. The two target grid cells with indexes 11 and 14: index difference = 14-11 = 3 > 2, and they are not continuous.

[0108] Then, the embodiment of the present application can determine a continuous target grid cell group, and the indexes of the continuous target grid cell group are [2, 4, 6], i.e., the continuous target grid cell group contains 3 target grid cells, so that a continuous abnormal area can be formed. In the actual terrain, the continuous abnormal area can correspond to a longer ridge or a wide depression.

[0109] In addition, the embodiment of the present application can determine two independent target grid cells, and the indexes are 11 and 14 respectively. In the actual terrain, the independent abnormal area can correspond to two isolated stones or independent small pits.

[0110] Further, the embodiment of the present application can divide the continuous group [2, 4, 6] and the grid cells covered thereby (i.e., the grid cells with indexes 3 and 5) into a cell group. All the remaining continuous grid cells not covered (such as the grid cells corresponding to [7, 8, 9, 10, 11, 12, 13, 14, 15]) are divided into a cell group. For the independent grid cell not covered (such as the grid cell corresponding to [1]), it can be classified into the same cell group as the adjacent target grid cell. That is, the grid cell with index [1] can be divided into the cell group corresponding to the continuous group [2, 4, 6].

[0111] It can be understood that the embodiment of the present application can traverse each column of grids, and perform the steps S305-S313 on each column of grids to determine one or more cell groups corresponding to each column of grids.

[0112] Further, in the embodiments of the present application, the height data acquisition module 4004 is configured to perform the following step S401.

[0113] S401: Collect the real-time installation height of each depth limiting wheel 201 according to the first preset frequency.

[0114] The real-time installation height can be collected by a laser ranging sensor or an ultrasonic distance sensor. The distance sensor can be arranged on the fixed frame of the seeding machine or the depth limiting wheel support 2022 to collect the vertical distance between the grounding point of the depth limiting wheel 201 and the fixed frame of the seeding machine or the depth limiting wheel support 2022, as the real-time installation height data.

[0115] Further, the speed data acquisition module 4005 can be configured to perform the following step S402.

[0116] S402: Collect the real-time running speed of the seeding machine according to the first preset frequency.

[0117] Further, the data processing module 4002 is further configured to perform the following steps S403-S406.

[0118] S403: For each column of grids, calculate the average value of each first ground elevation data corresponding to at least one group of unit groups, to obtain at least one target ground elevation data.

[0119] In the embodiments of the present application, for each unit group, the arithmetic mean of all grid unit ground elevation data contained therein can be calculated to generate the target ground elevation data of the unit group.

[0120] S404: For each column of grids, determine at least one installation height adjustment value based on the corresponding real-time installation height, at least one target ground elevation data and / or the preset seeding depth.

[0121] In the embodiments of the present application, the installation height adjustment value can be calculated based on the following formula.

[0122] Δh = (H_target - H_set) - H_curr; Where: Δh represents the installation height adjustment value, H_curr represents the real-time installation height of the depth limiting wheel 201, H_target represents the target ground elevation data, and H_set represents the preset seeding depth.

[0123] It can be understood that when (H_target - H_set) > H_curr, i.e., Δh > 0, it indicates that the current height of the depth limiting wheel 201 is too high, and needs to be lowered to compensate for the ground surface drop to avoid shallow trenching. (H_target - H_set) < H_curr, i.e., Δh < 0, indicates that the current height of the depth limiting wheel 201 is too low, and needs to be lowered to match the target ground surface to ensure that the trenching depth meets the standard. When Δh = 0, the height of the depth limiting wheel 201 matches the target state, and no adjustment is needed.

[0124] It should be noted that the above formula is only an example, and can be adjusted according to actual conditions. The embodiments of the present application do not make specific limitations.

[0125] S405: For each column grid, based on the latest collected real-time travel speed, the current system time and / or the starting index of the cell group, determine the instruction issuing time stamp corresponding to each cell group.

[0126] For example, when the number of cell groups is one, the value of the instruction issuing time stamp can be the sum of the current system time and the preset response time, which can be equal to 500 milliseconds. The embodiments of the present application do not make specific limitations.

[0127] When the number of cell groups is multiple, the instruction issuing time stamp can be calculated according to the following formula: T_stamp(i) = T_current + (S_index(i) × L_cell) / V_current + T_delay; Wherein: T_stamp(i) represents the instruction issuing time stamp of the i-th cell group, 0 < i ≤ V, V is the total number of cell groups, T_current represents the current system time, S_index(i) represents the starting index of the i-th cell group, L_cell represents the actual physical length corresponding to a single grid cell (i.e. land length), V_current represents the real-time travel speed of the seeding machine, and T_delay represents the system response delay.

[0128] It should be noted that L_cell can be calculated based on the following formula: L_cell = L_physical / R; Wherein, L_cell is the actual physical length corresponding to a single grid cell, R represents the grid longitudinal size (i.e. the number of grid longitudinal directions), and L_physical represents the estimated physical length, which is a pre-estimated actual physical length corresponding to the target analysis area, for example, equal to 3 meters.

[0129] For example, if L_physical = 3.0 meters, R = 30, then L_cell = 3.0 / 30 = 0.1 meters, i.e., each grid cell corresponds to a real terrain length of 10 centimeters.

[0130] In some implementations, the specific value of the system response delay can be determined based on actual conditions, for example, can be equal to 100 milliseconds, and the embodiments of the present application do not make specific limitations.

[0131] S406: Based on the at least one installation height adjustment value and the instruction issuance timestamp, generate at least one first adjustment instruction.

[0132] It can be understood that for each column of grids, the number of corresponding first adjustment instructions is equal to the number of cell groups, and the first adjustment instruction can carry the installation height adjustment value and the instruction issuance timestamp. In addition, the first adjustment instruction can also carry the motor rotation number of turns.

[0133] The control instruction output module 4003 can be specifically configured to perform the following steps S501-S502.

[0134] S501: Receive at least one first adjustment instruction.

[0135] It can be understood that the data processing module 4002 can be in communication connection with the control instruction output module 4003, so as to issue the first adjustment instruction through the control instruction output module 4003.

[0136] S502: For each column of grids, based on the instruction issuance timestamp, issue at least one first adjustment instruction to the corresponding depth limiting adjustment mechanism 202.

[0137] In the embodiments of the present application, the first adjustment instruction can also carry the instruction issuance timestamp. Then, the control instruction output module 4003 can determine the specific instruction issuance time based on the instruction issuance timestamp. Further, the first adjustment instruction adopts a structured data format, including fields such as instruction header, target mechanism address (corresponding to target trencher), installation height adjustment value, rotation number of turns, rotation direction, and / or instruction issuance timestamp.

[0138] Further, in the case of one first adjustment instruction, the first adjustment instruction can be directly issued to the corresponding depth limiting adjustment mechanism 202 based on the instruction issuance timestamp. In the case of multiple first adjustment instructions, in the embodiments of the present application, all first adjustment instructions can be arranged in ascending order according to the instruction issuance timestamps carried thereby, to form a time-sequenced instruction execution sequence, so as to accurately issue each first adjustment instruction. In this way, when the issuance time corresponding to the instruction issuance timestamp arrives, the corresponding first adjustment instruction can be issued to the corresponding depth limiting adjustment mechanism 202.

[0139] It can be seen that the embodiment of the application can establish a high-precision timing mechanism based on the system clock, trigger the sending of the instruction immediately when the instruction timestamp arrives, and realize the precise management of the trencher 100.

[0140] Further, the pressure data acquisition module 4006 can be configured to perform the following step S601.

[0141] S601: According to a second preset frequency, collect the first pressure data of each pressure sensor 300.

[0142] The second preset frequency can be 10Hz to 50Hz, that is, 10 to 50 times of pressure data collection per second, to ensure that the instantaneous impact force or high-frequency force fluctuation that may occur when the depth wheel 201 rolls on rugged terrain can be captured.

[0143] Further, the data processing module 4002 is further configured to perform the following steps S602-S604.

[0144] S602: For each pressure sensor 300, periodically collect a plurality of first pressure data at a third preset frequency to form a data set; wherein the third preset frequency is less than the second preset frequency.

[0145] For example, the third preset frequency can be 1Hz to 2Hz, that is, the embodiment of the application can collect first pressure data once every 1 second or 0.5 seconds. In the case of a second preset frequency of 10Hz and a third preset frequency of 1Hz, for a first pressure sensor 300, 10 pressure data can be collected in a collection period.

[0146] S603: For each pressure sensor 300, calculate the second standard deviation of the corresponding data set.

[0147] It can be understood that the second standard deviation is a quantitative indicator of the dispersion of all pressure values in the data set from the average value. The size of the second standard deviation directly reflects the degree of fluctuation of the pressure acting on the depth wheel 201 in the corresponding time window.

[0148] S604: In the case where the second standard deviation is greater than a second preset threshold, an alarm information is generated.

[0149] For example, the second preset threshold can be equal to 75N to 100N, which is not limited in the embodiment of the application.

[0150] It is worth noting that the sharp fluctuation of pressure data (represented by the second standard deviation being too large) is a key indicator of abnormality of the seeder state, indicating that the current terrain complexity has exceeded the normal adjustment range of the system, there may be continuous rugged road, or the depth limiting wheel assembly 200 encounters obstacles (such as buried stones) or gets stuck, etc. At this time, the embodiment of the application can generate an alarm information, and based on the alarm information, sound and light alarms and other warnings can be issued. The sound and light alarms can be set in the cab to remind the driver of the occurrence of abnormal conditions.

[0151] Figure 7 A flowchart of a hilly and mountainous area seeder depth control method provided by the embodiment of the application.

[0152] As Figure 7 shown, the embodiment of the application also provides a hilly and mountainous area seeder depth control method, which can be applied to the aforementioned seeder. The seeder includes a plurality of side-by-side arranged furrowers 100 and a depth limiting wheel assembly 200 corresponding to each furrower 100. The depth limiting wheel assembly 200 at least includes a depth limiting wheel 201 and a depth limiting adjustment mechanism 202. The depth limiting wheel 201 is used to roll along the ground surface to cooperate with the furrower 100 to limit the furrowing depth. The depth limiting adjustment mechanism 202 is used to adjust the installation height of the depth limiting wheel 201 and thus adjust the furrowing depth of the furrower 100. The method can include the following steps S701-S705.

[0153] S701: According to a first preset frequency, a visual image of a to-be-seeded area is obtained; wherein the visual image covers at least the furrowing range of the furrower 100; S702: The visual image is converted into a depth map using an image processing model; wherein the depth map includes ground elevation data; S703: The visual image is subjected to grid processing; wherein the size of the grid is determined according to the number of furrowers 100, and the lateral size of the grid is equal to the number of furrowers 100; S704: For each column of grid, at least one target ground elevation data corresponding thereto is determined based on the depth map; and based on the preset seeding depth and the at least one target ground elevation data, at least one first adjustment instruction is determined, the first adjustment instruction including an installation height adjustment value; S705: At least one first adjustment instruction corresponding to each column of grid is issued to the corresponding depth limiting adjustment mechanism, so that the depth limiting adjustment mechanism adjusts the installation height of the depth limiting wheel based on the first adjustment instruction, and thus adjusts the furrowing depth of the furrower.

[0154] In a possible implementation, the view range of the visual image further comprises two preset markers, and the two preset markers are respectively arranged at two lateral edges of the seeding machine. S7031: determining two preset markers in the visual image; wherein the preset markers are imaging points of the preset markers in the visual image; S7032: determining two auxiliary markers in the visual image by using pixel coordinates of the two preset markers and a preset geometric constraint function; wherein the two preset markers and the two auxiliary markers are used to form a trapezoid, a line connecting the two preset markers forms a first bottom side of the trapezoid, and a line connecting the two auxiliary markers forms a second bottom side of the trapezoid, and the first bottom side is a long bottom side and the second bottom side is a short bottom side. S7033: determining a target analysis region in the visual image; wherein the target analysis region is a trapezoidal region enclosed by the two preset markers and the two auxiliary markers. S7034: taking the two preset markers and the two auxiliary markers as grid vertices to perform grid processing on the target analysis region. In a possible implementation, the step S704 can comprise the following steps S7041-S7043. S7041: determining, by using the depth map, first ground elevation data corresponding to each grid cell in each column grid in the target analysis region. S7042: determining, for each column grid, a first average value and a first standard deviation of the respective first ground elevation data. S7043: for each column grid, traversing each grid cell corresponding to the column grid to determine N target grid cells; wherein the first ground elevation data of the target grid cells satisfies the following first formula: |x+t-μ|>σ; wherein x represents the first ground elevation data, t represents an adjustment parameter, t>0, μ represents the first average value, and σ represents the first standard deviation.

[0155] In a possible implementation, the step S7043 can further comprise the following steps S7044-S7046.

[0156] S7044: for each column grid, in the case of N=0, dividing all grid cells in the current column grid into a same cell group. S7045: for each column grid, in the case of N=1, if a sequence index P of the target grid cell is located in an interval [L-K, L], then the current column grid is divided into two cell groups, and a division point is located before the index position of the target grid cell; wherein L is a sequence length of the current column grid, and K is a preset positive integer. S7046: Otherwise, divide the current column grid into three cell groups, the first cell group includes the 1st to the P-M-1st grid cells, the second cell group includes the P-Mth to the P+Mth grid cells, and the third cell group includes the P+M+1st to the Lth grid cells, wherein M is a preset positive integer.

[0157] In a possible implementation, after step S7043, the method further includes steps S7047-S7049.

[0158] S7047: For each column grid, determine whether there is a continuous target grid cell in the case of N > 1; wherein the index difference of any adjacent target grid cells in the continuous target grid cell is not greater than a first preset threshold value. S7048: In the case where there is no continuous target grid cell, divide all the grid cells in the current column grid into the same cell group. S7049: In the case where there is a continuous target grid cell, divide each continuous target grid cell and the grid cells covered thereby into the same cell group, divide the continuous grid cells not covered by the continuous target grid cell into the same cell group, and divide the independent grid cells not covered by the continuous target grid cell and the adjacent target grid cells thereof into the same cell group.

[0159] In a possible implementation, after step S7049, the method further includes steps S70410-S70415.

[0160] S70410: Collect the real-time installation height of each depth limiting wheel 201 according to a first preset frequency. S70411: Collect the real-time running speed of the seeding machine according to a first preset frequency. S70412: For each column grid, calculate the average value of each first ground elevation data corresponding to at least one cell group to obtain at least one target ground elevation data. S70413: For each column grid, determine at least one installation height adjustment value based on the latest collected real-time installation height, at least one target ground elevation data, and / or a preset seeding depth. S70414: For each column grid, determine the instruction issuing time stamp corresponding to each cell group based on the latest collected real-time running speed, the current system time, and / or the starting index of the cell group. S70415: Generate at least one first adjustment instruction based on the at least one installation height adjustment value and the instruction issuing time stamp.

[0161] In a possible implementation, the first adjustment instruction further includes the instruction issuing time stamp.

[0162] Step S705 can include steps S7051-S7052.

[0163] S7051: receiving at least one first adjustment instruction; S7052: for each column of the grid, issuing the at least one first adjustment instruction to the corresponding depth-limiting adjustment mechanism 202 based on an instruction issue timestamp.

[0164] In a possible implementation, the seeding machine further includes a pressure sensor 300 installed on each depth-limiting wheel assembly 200; the method provided by the embodiment of the application can further include steps S7061-S7064.

[0165] S7061: collecting first pressure data of each pressure sensor 300 at a second preset frequency; S7062: for each pressure sensor 300, periodically aggregating a plurality of first pressure data at a third preset frequency to form a data set; wherein the third preset frequency is less than the second preset frequency; S7063: for each pressure sensor 300, calculating a second standard deviation of the corresponding data set; S7064: generating an alarm information in a case where the second standard deviation is greater than a second preset threshold.

[0166] As can be seen from the above, the embodiment of the application provides a hill and mountain area seeding machine depth control system and method, which enables the seeding machine to maintain stable operation in the hill and mountain area with large slope changes and frequent terrain undulations through real-time sensing and dynamic adjustment mechanism, effectively expanding the application range of the seeding machine under complex terrain conditions, and providing reliable technical support for hill and mountain area agricultural production. Further, the system and method realize precise and stable control of the seeding depth, create an excellent germination environment for crop growth, and ensure that the seeds are always at the optimal soil depth, significantly improving the uniformity of emergence and the robustness of seedlings.

[0167] Figure 8 A structural schematic diagram of an electronic device provided by the embodiment of the application is shown.

[0168] As shown in Figure 8 The embodiment of the application provides an electronic device, which includes a processor 10 and a memory 20, the memory 20 stores a computer program, and the processor 10 is configured to run the computer program to execute the steps in any of the hill and mountain area seeding machine depth control method embodiments.

[0169] In a specific implementation, the present invention also provides a computer storage medium, wherein the computer storage medium may store a program, which, when executed, may include some or all of the steps of the various embodiments of the hilly and mountainous area seeding depth control method provided by the present invention. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0170] It is readily understood that, based on the several embodiments provided in this application, those skilled in the art can combine, split, or reorganize the embodiments of this application to obtain other embodiments, none of which exceed the protection scope of this application.

[0171] The above detailed embodiments further illustrate the purpose, technical solution, and beneficial effects of the embodiments of this application. It should be understood that the above are merely specific embodiments of the embodiments of this application and are not intended to limit the protection scope of the embodiments of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solutions of the embodiments of this application should be included within the protection scope of the embodiments of this application.

Claims

1. A hilly terrain planter depth control system, characterized by, The application is applied to a seeding machine, the seeding machine comprises a plurality of furrow openers arranged side by side and a depth limiting wheel assembly corresponding to each of the furrow openers, the depth limiting wheel assembly comprises at least a depth limiting wheel and a depth limiting adjusting mechanism, the depth limiting wheel is used for rolling along the ground surface to cooperate with the furrow opener to limit the furrowing depth, and the depth limiting adjusting mechanism is used for adjusting the installation height of the depth limiting wheel to adjust the furrowing depth of the furrow opener. The system comprises: A visual module is configured to collect a visual image of a region to be seeded at a first preset frequency; wherein the field of view of the visual image covers at least the furrowing range of the furrow opener; A data processing module is configured to convert the visual image into a depth map using an image processing model; wherein the depth map comprises ground elevation data; The visual image is subjected to grid processing; wherein the size of the grid is determined according to the number of furrow openers, and the lateral size of the grid is equal to the number of furrow openers; For each column of the grid, at least one target ground elevation data corresponding to the column is determined based on the depth map; and at least one first adjusting instruction including an installation height adjustment value is determined based on a preset seeding depth and the at least one target ground elevation data; A control instruction output module is configured to issue at least one first adjusting instruction corresponding to each column of the grid to the corresponding depth limiting adjusting mechanism, so that the depth limiting adjusting mechanism adjusts the installation height of the depth limiting wheel based on the first adjusting instruction, thereby adjusting the furrowing depth of the furrow opener.

2. The hilly country planter depth control system of claim 1, wherein, The field of view of the visual image also includes two preset calibration objects, and the two preset calibration objects are arranged on the lateral edges of the seeding machine; The data processing module is specifically configured to: Determine two preset calibration points in the visual image; wherein the preset calibration points are imaging points of the preset calibration objects in the visual image; Determine two auxiliary calibration points in the visual image using the pixel coordinates of the two preset calibration points and a preset geometric constraint function; wherein the two preset calibration points and the two auxiliary calibration points are used to form a trapezoid, the connecting line of the two preset calibration points forms the first base of the trapezoid, the connecting line of the two auxiliary calibration points forms the second base of the trapezoid, and the first base is the long base and the second base is the short base; Determine a target analysis region in the visual image; wherein the target analysis region is a trapezoidal region enclosed by the two preset calibration points and the two auxiliary calibration points; Grid the target analysis region using the two preset calibration points and the two auxiliary calibration points as grid vertices.

3. The hilly country planter depth control system of claim 2, wherein, The data processing module is further configured to: Determine first ground elevation data corresponding to each grid cell in each column of the grid in the target analysis region using the depth map; For each column of the grid, determine the first average value and the first standard deviation of each of the first ground elevation data corresponding to the column; For each column of the grid, each corresponding grid cell is traversed to determine N target grid cells; wherein the first ground elevation data of the target grid cells satisfy the following first formula: |x+t-μ|>σ; wherein x represents the first ground elevation data, t represents an adjustment parameter, t>0, μ represents the first average value, and σ represents the first standard deviation.

4. The hilly country planter depth control system of claim 3, wherein, The data processing module is further configured to: For each column of the grid, when N=0, all grid cells in the current column of the grid are divided into a same cell group; For each column of the grid, when N=1, if the sequence index P of the target grid cell is located within the interval [L-K, L], the current column of the grid is divided into two cell groups, and the division point is located before the index position of the target grid cell; wherein L is the sequence length of the current column of the grid, and K is a preset positive integer; Otherwise, the current column of the grid is divided into three cell groups, the first cell group includes the 1st to the P-M-1th grid cells, the second cell group includes the P-Mth to the P+Mth grid cells, and the third cell group includes the P+M+1th to the Lth grid cells, wherein M is a preset positive integer.

5. The hilly country planter depth control system of claim 4, wherein, The data processing module is further configured to: For each column of the grid, when N>1, it is determined whether there is a continuous target grid cell; wherein the index difference of any adjacent target grid cells in the continuous target grid cell is not greater than a first preset threshold value; In the case where there is no continuous target grid cell, all grid cells in the current column of the grid are divided into a same cell group; In the case where there is a continuous target grid cell, each continuous target grid cell and the grid cells covered thereby are divided into a same cell group, the continuous grid cells not covered by the continuous target grid cells are divided into a same cell group, and the independent grid cells not covered by the continuous target grid cells and the target grid cells adjacent thereto are divided into a same cell group.

6. The hilly country planter depth control system of claim 4 or 5, wherein, The system further comprises: A height data acquisition module configured to acquire the real-time installation height of each depth limiting wheel according to the first preset frequency; A speed data acquisition module configured to acquire the real-time running speed of the seeding machine according to the first preset frequency; The data processing module is further configured to: For each column of the grid, the average value of each first ground elevation data corresponding to at least one cell group is calculated to obtain at least one target ground elevation data; For each column of the grid, at least one installation height adjustment value is determined based on the latest acquired real-time installation height, at least one target ground elevation data, and / or the preset seeding depth; For each column of the grid, the instruction issuance time stamp corresponding to each cell group is determined based on the latest acquired real-time running speed, the current system time, and / or the starting index of the cell group; At least one first adjustment instruction is generated based on at least one installation height adjustment value and the instruction issuance time stamp.

7. The hilly country planter depth control system of claim 6, wherein, The first adjustment instruction further comprises an instruction issuing timestamp; and the control instruction output module is specifically configured to: receive at least one first adjustment instruction; issue, for each column of the grid, at least one first adjustment instruction to the corresponding depth limiting adjustment mechanism based on the instruction issuing timestamp.

8. The hilly country planter depth control system of claim 1, wherein, The seeding machine further comprises a pressure sensor installed on each depth limiting wheel assembly; and the system further comprises: a pressure data acquisition module configured to acquire first pressure data of each pressure sensor at a second preset frequency; the data processing module is further configured to periodically aggregate a plurality of first pressure data of each pressure sensor at a third preset frequency to form a data set, wherein the third preset frequency is less than the second preset frequency; calculate, for each pressure sensor, a second standard deviation of the corresponding data set; generate an alarm information in the case where the second standard deviation is greater than a second preset threshold.

9. A method of controlling the depth of a hill and mountain land planter, characterized by, The method is applied to a seeding machine, which comprises a plurality of furrow openers arranged side by side and a depth limiting wheel assembly corresponding to each furrow opener, the depth limiting wheel assembly comprising at least a depth limiting wheel and a depth limiting adjustment mechanism, the depth limiting wheel being used to roll along the ground surface to limit the furrowing depth in cooperation with the furrow opener, and the depth limiting adjustment mechanism being used to adjust the installation height of the depth limiting wheel and thus adjust the furrowing depth of the furrow opener; the method comprises: acquiring a visual image of a to-be-seeded area at a first preset frequency; wherein the field of view of the visual image covers at least the furrowing range of the furrow opener; converting the visual image into a depth map using an image processing model; wherein the depth map comprises ground elevation data; performing grid processing on the visual image; wherein the size of the grid is determined according to the number of furrow openers, and the lateral size of the grid is equal to the number of furrow openers; for each column of the grid, determining at least one target ground elevation data corresponding thereto based on the depth map, and determining at least one first adjustment instruction based on a preset seeding depth and the at least one target ground elevation data, the first adjustment instruction comprising an installation height adjustment value; issuing the at least one first adjustment instruction corresponding to each column of the grid to the corresponding depth limiting adjustment mechanism, so that the depth limiting adjustment mechanism adjusts the installation height of the depth limiting wheel based on the first adjustment instruction, and thus adjusts the furrowing depth of the furrow opener.

10. An electronic device, comprising: comprise: one or more processors; and a memory configured to store one or more programs; wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the hilly and mountainous area seeding machine depth control method according to claim 9.

Images