Grassland precise mixed reseeding method based on surface vegetation detection

Abstract

The invention provides a grassland precise mixed reseeding method based on surface vegetation detection, which comprises the following steps: inputting information of grass seeds to be reseeded selected in advance, and carrying out classified numbering and storage according to properties and storage positions; acquiring soil information and dominant species of the current land parcel, and outputting properties of grass seeds required to be reseeded; identifying the bare earth surface of the current land parcel in real time, calculating the bare area, converting the operation coordinates, and judging whether the area reaches the standard of reseeding or not; when the standard is not reached, entering the next operation land parcel; when the standard is reached, grass seeds needing to be resown in the current land parcel are matched, the height of the sowing single body is adjusted according to the sowing depth, the matched grass seeds are quantitatively discharged according to the sowing amount when the operation position is reached, and accurate resowing is completed. According to the method, the mixed reseeding operation can be regulated and controlled in real time, the applicability of a reseeding machine to plaque and small-area degraded grassland and the selectivity of reseeding species are improved, and the damage to original grassland vegetation caused by ditching and seeding and further grassland degradation caused by improper selection of the reseeding species are reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent agricultural machinery equipment, and particularly to a precise grassland mixed seeding method based on surface vegetation detection. Background Art

[0002] Grassland degradation is a global problem, mainly caused by overgrazing, climate change, soil erosion, and improper management. Grassland degradation not only affects the economic value of grasslands but also seriously damages their ecological functions. Grassland mixed seeding is considered one of the effective means to quickly restore degraded grasslands. Mixed seeding can increase grassland species diversity, restore grassland productivity, improve forage quality, and repair grassland ecology, which is of great significance for bringing into play the economic and ecological benefits of grassland resources.

[0003] Currently, machine seeding is mostly used for grassland mixed seeding. However, most existing grassland seeding machines are modified from farmland seeders, which are not suitable for seeding in patchy and small-area degraded grasslands and do not have selectivity for mixed seeding species. In addition, blindly opening ditches for seeding may damage the original grassland vegetation, further affecting the grassland ecological environment and exacerbating grassland degradation. With the progress of agricultural machinery intelligent technology, there is an urgent need for a precise grassland mixed seeding method that can autonomously identify bare ground surfaces, determine mixed seeding species, and complete selective seeding. Summary of the Invention

[0004] In view of the deficiencies in the prior art, on the basis of the prior art, the present invention constructs a precise grassland mixed seeding method based on surface vegetation detection, identifies seeding areas and selects seeding species for grasslands that need to be mixed-seeded, and real-time regulates the grassland mixed seeding process based on bare areas, native dominant populations, soil conditions, etc.

[0005] To achieve the objective, the technical solutions adopted by the present invention are as follows: A precise grassland mixed seeding method based on surface vegetation detection, comprising the following steps: S1. Input all information of grass seeds to be sown, and classify and number them according to species characteristics and storage locations; S2. Demarcate the operation area in the Beidou navigation electronic map and divide the operation plots; S3. Obtain the soil condition information within the current plot and judge the characteristics of the grass seeds to be sown; S4. Take images and videos of the species growing within the current plot and extract the dominant species; S5. Identify the bare ground surface of the current plot; S6. Calculate the area of the bare ground surface of the current plot, extract the boundary points, and convert them into the operation coordinates of the seeding machine; S7. Automatically match the grass seeds to be sown in the current plot and calculate the seeding rate; S8. Adjust the height position of the seeding unit, and open the corresponding seed box for the matching grass seeds for quantitative seeding.

[0006] The specific content of S1 is as follows: S1-1. Fill the selected grass seeds to be reseeded into each seed box of the reseeding machine respectively; S1-2. Record the grass seed information Z in each seed box of the reseeding machine, classify and number it. The numbering content includes the family F, stress resistance K, seeding depth D, seeding rate M, and the number of the seed box to which it belongs. The grass seed information Z to be reseeded is represented by a data set and the data is input and stored. Specifically: Z F ={F1,F2,…,F n z}, Z K ={K1,K2,…,K n z}, Z Q ={(F1, K1,D1,M1,1), (F2,K2,D2,M2, 2),…, (F n z , K n z ,D n z ,M nz , n z )}, Z F is the family set, Z K is the stress resistance set, Z Q is the full attribute set, n z is the total number of seed boxes carried by the reseeding machine; Further, the classification and numbering method in step S1-1 is as follows: The family to which it belongs is divided into three categories: legume, gramineae, and compositae, and is represented by the numbers L, G, and C respectively, that is, F1, F2,…, F n ∈{L, G, C}; The stress resistance selects three characteristics: drought resistance, cold resistance, and salt tolerance, and is represented by the numbers d, c, and s respectively, that is, K1, K2,…, K n ∈{d, c, s}; The seeding depth is numbered with the seeding depth value required by the agronomy of the grass seeds, and the unit is cm; The seeding rate is numbered with the seeding rate per unit area required by the agronomy of the grass seeds, and the unit is kg / hm 2 ; The number of the seed box to which it belongs corresponds to the seed box number.

[0007] The specific content of S2 is as follows: S2-1. Manually delimit the area to be operated in the Beidou navigation electronic map. The operation range is delimited as a quadrilateral, and the longitude and latitude coordinate information of the four vertices is extracted, in the form of [x a1 , y a1 , [x a2 , y a1 , [x a1 , y a2 , [x a2 , y a2Store it as the operation boundary condition. S2-2. Perform secondary division on the demarcated operation area. Take the working width B of the overseeder as the width of the plot, and divide each row into m plots. Since 1 degree of longitude and latitude is approximately equal to 111 km, then where m is the number of horizontal plots; x a1 and x a2 are the vertex longitudes, with the unit of °; B is the working width of the overseeder, with the unit of km; Divide each column into n plots, and store the vertex longitude and latitude coordinates of each small plot with [x h i , y h j (i = 1, 2,..., m + 1; j = 1, 2,..., n + 1).

[0008] Specifically, S3 is as follows: S3-1. Obtain the real-time longitude and latitude position of the overseeder through the Beidou navigation module, compare it with the vertex position coordinates of the small plot, determine the plot where the overseeder is currently located, and mark this plot as "under operation"; S3-2. Set the soil relative humidity threshold RH, pH threshold PH, and temperature threshold T according to the drought, saline-alkali land, and high-temperature grading standards. Use the soil temperature and humidity sensor to measure the soil temperature T true and the soil relative humidity RH true in the operation area, and use the soil pH value sensor to measure the soil acidity and alkalinity PH true in the operation area. Compare the obtained T true , RH true , PH true with the pre-set temperature threshold T, relative humidity threshold RH, and pH threshold PH, and output the required characteristics K of the overseeded grass seeds according to the comparison results d .

[0009] Furthermore, the discrimination method for the required type K d of the overseeded grass seeds in step S3-2 is as follows: When only T true , RH true , PH true exceeds the set threshold, output that the required characteristics K true of the overseeded grass seeds is c; when only RH d exceeds the set threshold, output that the required characteristics K true of the overseeded grass seeds is d; when only PH d exceeds the set threshold, output that the required characteristics K true of the overseeded grass seeds is s. When two or more exceed the threshold, PH d takes priority, RH true ​true Second, T true Finally, when all three exceed the threshold, output the characteristics K of the grass seeds to be reseeded d as s.

[0010] Specifically, S4 is as follows: S4-1. Use a CMOS machine vision camera to capture the species image video in the current plot, obtain the species image from the video, and rotate, crop, and splice the filtered image to obtain the overall planting situation image of the current plot; S4-2. Perform semantic segmentation on the overall planting situation image, mark the species types contained in the image, count the total number t of all species individuals and the number p of individuals of a single species, calculate the dominance index SIM of a single species, compare the species dominance rates, and the species with the highest dominance index is the dominant species in the current plot; Furthermore, the calculation method of the dominance index of a single species in step S4-2 is: Berger-Parker dominance index calculation formula where t is the total number of all species individuals; p is the number of individuals of a single species.

[0011] S4-3. Match the reseeding grass seeds for the dominant species and output the family F to which the grass seeds to be reseeded belong d .

[0012] Specifically, S5 is as follows: S5-1. Draw a resistance-pressure calibration curve based on the calibration data set of the flexible piezoresistive sensor and fit the linear formula within the calibrated resistance range where P is the pressure, in N; R is the resistance value of the sensor, in Ω; S5-2. The seeder moves forward, and uses a single-point acquisition circuit to obtain the real-time voltage values U1, U2,..., U of the flexible piezoresistive sensor array p , calculate the resistance values R1, R2,..., R of the sensor according to Ohm's law p , calculate the actual pressures P1, P2,..., P through the fitted linear formula p , p is the number of seeding monomers, and 1 set of data is recorded and stored every 0.5 seconds; S5-3. During the forward movement, monitor the pressure mutation points. When the real-time pressure value P measured by a certain row of sensors mutates and the value is approximately 0, it means that there is no plant growth ahead and there is an empty spot in this row, and mark this point as a bare spot.

[0013] Furthermore, the arrangement method of the flexible piezoresistive sensor in step S5-2 is: Flexible piezoresistive sensors are arranged at intervals of L at the very front of the machine. L is the row spacing of the reseeding machine, with the unit of m. The arrangement position is close to the ground, not exceeding 2 cm from the ground. The flexible piezoresistive sensors correspond one by one to the positions of the sowing units at the back and are on the same axis. The number of arranged sensors is the same as the number of sowing units. Furthermore, the determination method for mutation in the step S5-3 is as follows: When the sensor returns a new set of data P u , the previous five sets of returned data P u-5 , P u-4 , P u-3 , P u-2 , P u-1 are called and the average value of the five sets of actual pressures is taken, where u is the number of data sets. When it is determined that a mutation has occurred in the sensor of this row. Among them, is the average value of the actual pressure, P u is the latest measured actual pressure value, and is the minimum pressure value generated when passing through the plant stem in the horizontal direction, with the unit of N.

[0014] The specific content of S6 is as follows: S6-1: Use OpenMV to take a surface image with the exposed point identified in S5-3 as the center point. Starting from the smallest image that can be taken of the exposed point, gradually expand the coverage range of the collected image. Each time, expand one circle of pixel points and take an image, and at the same time extract the features of the outermost pixel points of the image. Stop expanding and taking pictures when all the outermost pixel points of the image are points covered with vegetation. S6-2: Use open-source image processing functions to perform edge detection and minimum rectangle contour fitting on the last taken image, extract the center point coordinates (x o , y o ) and the four vertex coordinates (x o1 , y o1 ), (x o2 , y o1 ), (x o1 , y o2 ), (x o2 , y o2 ) of the minimum rectangle, and calculate the area of the exposed ground where S is the area of the exposed ground, with the unit of m 2 ; x o1 , x o2 , y o1 , y o2is the vertex position coordinate, with the unit of m; S6-3. Determine whether the bare ground area S meets the conditions for reseeding. If the bare area reaches the standard for reseeding, extract the horizontal boundaries of the smallest rectangle fitted in step S6-2 respectively, take a point every row spacing L horizontally, and convert it into the operation coordinates at the rear end of the seeder. Otherwise, directly enter the next operation plot.

[0015] The specific content of S7 is as follows: S7-1. Send the characteristics K of the grass seeds to be reseeded obtained in step S3-2 d and the family F to which the grass seeds to be reseeded obtained in step S4-3 belong d back to the host computer, match them with all the information of the grass seeds to be reseeded entered in step S2, and according to the Boolean operation: where Z K and Z F are the stress resistance set and the family set. When the intersection of the two is all 1, it proves that the grass seeds to be reseeded are successfully matched and a unique grass seed to be reseeded can be obtained. Extract the corresponding grass seed number from the full attribute set Z Q to obtain the seeding depth, seeding rate data and storage location; S7-2. According to the seeding rate data of the grass seeds called from the full attribute set, convert the bare ground area of the current plot into the seeding mass G required for the grass seeds to be reseeded in the current plot.

[0016] Furthermore, the conversion method of the seeding rate of the grass seeds to be reseeded in step S7-2 is as follows: where G is the seeding mass required for the grass seeds to be reseeded in the current plot, with the unit of g; S is the bare ground area of the current plot, with the unit of m 2 ; M is the seeding rate per unit area required by the agronomy of the grass seeds, with the unit of kg / hm 2 .

[0017] The specific content of S8 is as follows: S8-1. According to the seeding depth of the matched grass seeds extracted in step S7-1, control the hydraulic cylinder to extend or shorten. At the same time, the ultrasonic sensor installed under the four-bar mechanism monitors the moving height h of the seeding unit in real time. When the height of the unit descending or rising meets the seeding depth requirement of the grass seeds, the hydraulic cylinder stops extending or shortening; S8-2. When it is detected that the operation coordinates at the rear end of the reseeding machine enter the boundary range of the bare area of the current plot, according to the storage position of the matching grass seeds extracted in step S7-1, control the rotation of the servo with the corresponding number, drive the rotation of the rotating partition below the matching grass seed box, the seeds fall into the central seed box, and then are discharged by the seeding unit, completing the reseeding. At the same time, enable the weighing sensor below the seed box to monitor the quality change g during the seeding process in real time; S8-3. When the quality change g is equal to the seeding quality G required for the grass seeds to be reseeded in the current plot, the servo returns to its original position, the rotating partition turns back to its original position, and the matching grass seed box stops discharging seeds outward.

[0018] The precise grassland mixed reseeding method involves the following key components: Soil temperature and humidity sensor, used to obtain soil moisture content and soil temperature; Soil pH value sensor, used to obtain soil acidity and alkalinity; CMOS machine vision camera, used to capture species image videos; Flexible pressure-sensitive sensor, used to judge the empty spots in the row and locate the bare points; OpenMV vision module, used to capture images of the bare ground surface and calculate the bare area; Weighing sensor, installed below the seed box, used to detect the seeding amount; Rotating partition, used to separate the seed box and the central seed box, installed below the seed box; Servo, used to open and close the seed box, control the size of the seeding amount, and is connected to the rotating partition; Central seed box, used to store the grass seeds to be reseeded and distribute them to each seeding unit; Ultrasonic sensor, used to monitor the lifting height of the seeding unit; Hydraulic cylinder, installed on the parallel four-bar mechanism of the seeding unit, used to adjust the lifting and lowering height.

[0019] The technical solution of the present invention has the following beneficial effects: The present invention adopts the precise grassland reseeding technology based on surface vegetation detection, and during the grassland improvement, it adjusts the reseeding operation in real time according to the soil conditions, dominant species, and the real-time changing bare ground surface area, improving the applicability of the grassland reseeding machine for reseeding patchy and small-area degraded grasslands and the selectivity of reseeding species, reducing the damage to the original grassland vegetation caused by blindly opening ditches for seeding and the further degradation of the grassland caused by improper selection of reseeding species; The present invention utilizes machine vision and sensor information fusion technology to achieve autonomous recognition of bare ground and accurate calculation of the bare area, convert the operation coordinates of the overseeder, accurately locate the areas to be overseeded, select the grass seeds to be overseeded according to the soil conditions, dominant species conditions, and bare area of the operation plot, call the relevant agronomic information of the grass seeds, and adjust the seeding depth and seeding rate, thus completing the precise overseeding of grassland for selective overseeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic flowchart of an embodiment of the present invention; Figure 2 It is a schematic structural diagram of the overseeding control system in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] The embodiments of the present invention provide a precise mixed overseeding method for grassland based on surface vegetation detection, which can accurately perform mixed overseeding operations according to soil conditions, dominant species, and the area of bare ground that changes in real time, avoiding the damage to the original grassland caused by blind ditching and the further degradation of the grassland caused by inappropriate overseeding species.

[0023] As Figure 1 shown, the specific steps of the present invention are as follows: S1-1. Fill the grass seeds to be overseeded selected in advance into each seed box of the overseeder; S1-2. Record the grass seed information Z in each seed box of the overseeder, classify and number it. The numbered content includes the family F, stress resistance K, seeding depth D, seeding rate M, and the number of the seed box to which it belongs. The grass seed information Z to be overseeded is represented by a data set and the data is entered and stored; S2-1. Manually delimit the area to be operated in the Beidou navigation electronic map. The operation range is delimited as a quadrilateral, and the longitude and latitude coordinate information of the four vertices is extracted, using [x a1 , y a1 , [x a2 , y a1 , [x a1 , y a2 , [x a2 , y a2 as the operation boundary conditions and store them; S2-2. Re-divide the defined operation area. Take the working width B of the seeding machine as the width of the plot. Divide each row into m plots and each column into n plots. Store the longitude and latitude coordinates of the vertices of each small plot as [x h i , y h j . S3-1. Obtain the real-time longitude and latitude position of the seeding machine through the Beidou navigation module, compare it with the vertex position coordinates of the small plot, determine the plot where the seeding machine is currently located, and mark this plot as "under operation". S3-2. Set the soil relative humidity threshold RH, pH threshold PH, and temperature threshold T according to the drought, saline-alkali land, and high-temperature grading standards. Use the soil temperature and humidity sensor to measure the soil temperature T true and the soil relative humidity RH true in the operation area, and use the soil pH value sensor to measure the soil pH value PH true in the operation area. Compare the obtained T true , RH true , and PH true with the pre-set temperature threshold T, relative humidity threshold RH, and pH threshold PH, and output the characteristics K of the grass seeds to be re-seeded according to the comparison results d . S4-1. Use a CMOS machine vision camera to take the species image video in the current plot, obtain the species image from the video, and perform rotation, cropping, and splicing on the filtered image to obtain the overall planting situation image of the current plot. S4-2. Perform semantic segmentation on the overall planting situation image, mark the species types contained in the image, count the total number t of all species individuals and the number p of individuals of a single species, calculate the dominance index SIM of a single species, compare the species dominance rates, and the species with the highest dominance index is the dominant species in the current plot. S4-3. Match the grass seeds to be re-seeded for the dominant species and output the family F d to which the grass seeds to be re-seeded belong. S5-1. Draw the resistance-pressure calibration curve according to the calibration data set of the flexible piezoresistive sensor, and fit the linear formula within the calibrated resistance range. S5-2. The seeding machine moves forward. Use the single-point acquisition circuit to obtain the real-time voltage values U1, U2,..., U p of the flexible piezoresistive sensor array. Calculate the resistance values R1, R2,..., R p of the sensor according to Ohm's law, and calculate the actual pressures P1, P2,..., P p through the fitted linear formula, where p is the number of seeding monomers, and record and store 1 set of data every 0.5 seconds. S5-3. During the forward movement, monitor the pressure mutation points. When the real-time pressure value P measured by a certain row of sensors mutates and the value is approximately 0, it indicates that there is no plant growth ahead in this row and there is an empty spot. Mark this point as the exposed point. S6-1. Use OpenMV to take a ground image with the exposed point identified in S5-3 as the center point. Take the smallest image that can be captured at the exposed point as the starting image, and gradually expand the coverage range of the captured image. Expand one pixel circle each time and take an image, and at the same time extract the features of the outermost pixel points of the image. Stop expanding and taking pictures when all the outermost pixel points of the image are covered with vegetation. S6-2. Use open-source image processing functions to perform edge detection and minimum rectangle contour fitting on the last captured image, and extract the center point coordinates (x o , y o ) and the four vertex coordinates (x o1 , y o1 ), (x o2 , y o1 ), (x o1 , y o2 ), (x o2 , y o2 ) of the minimum rectangle, and calculate the area of the exposed ground surface. S6-3. Judge whether the area S of the exposed ground surface meets the conditions for reseeding. If the exposed area reaches the standard for reseeding, extract the horizontal boundaries of the minimum rectangle fitted in step S6-2 respectively, take a point every row spacing L horizontally, and convert it into the operation coordinates at the rear end of the seeding machine. Otherwise, directly enter the next operation plot.

[0024] S7-1. Return the characteristics K d of the grass seeds to be reseeded obtained in step S3-2 and the family F d to which the grass seeds to be reseeded belong obtained in step S4-3 to the host computer, and match them with all the information of the grass seeds to be reseeded entered in step S2. After successful matching, extract the corresponding grass seed numbers from the full attribute set Z Q , and obtain the seeding depth, seeding rate data and storage location. S7-2. According to the seeding rate data of the grass seeds called from the full attribute set, convert the area of the exposed ground surface of the current plot into the seeding quality G required for the grass seeds to be reseeded in the current plot.

[0025] S8-1. According to the seeding depth of the matching grass seeds extracted in step S7-1, control the hydraulic cylinder to extend or shorten. At the same time, the ultrasonic sensor installed under the four-bar mechanism monitors the moving height h of the seeding unit in real time. When the height of the unit descending or ascending meets the seeding depth requirement of the grass seeds, the hydraulic cylinder stops extending or shortening. S8-2. When it is detected that the coordinates of the rear end of the reseeding machine enter the boundary range of the bare area of the current plot, according to the storage position of the matching grass seeds extracted in step S7-1, control the rotation of the servo with the corresponding number, drive the rotation of the rotating partition below the matching grass seed box, the seeds fall into the central seed box, and then are discharged by the seeding unit, completing the reseeding. At the same time, enable the weighing sensor below the seed box to monitor the quality change g during the seeding process in real time; S8-3. When the quality change g is equal to the seeding quality G required for the grass seeds to be reseeded in the current plot, the servo returns to its original position, the rotating partition rotates back to its original position, and the matching grass seed box stops discharging seeds outward.

[0026] As Figure 2 shown, the precise grass mixture reseeding method for grasslands involves the following key components: Soil temperature and humidity sensor, used to obtain soil moisture content and soil temperature; Soil pH value sensor, used to obtain soil acidity and alkalinity; CMOS machine vision camera, used to capture species image videos; Flexible pressure-sensitive sensor, used to judge the empty spots in the row and locate the bare points; OpenMV vision module, used to capture images of the bare ground surface and calculate the bare area; Weighing sensor, installed below the seed box, used to detect the seeding amount; Rotating partition, used to separate the seed box and the central seed box, installed below the seed box; Servo, used to open and close the seed box, control the size of the seeding amount, and is connected to the rotating partition; Central seed box, used to store the grass seeds to be reseeded and distribute them to each seeding unit; Ultrasonic sensor, used to monitor the lifting height of the seeding unit; Hydraulic cylinder, installed on the parallelogram mechanism of the seeding unit, used to adjust the lifting and lowering height.

[0027] Through the description of the above embodiments, those skilled in the art can clearly understand the specific implementation process of the present invention. It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A precise grassland mixed seeding method based on surface vegetation detection, characterized in that The method includes the following steps: S1. Enter all the information of the grass seeds to be reseeded, and classify and number them according to species characteristics and storage locations; S2. Define the operation area in the Beidou navigation electronic map and divide the operation plots; S3. Obtain the soil condition information in the current plot and judge the characteristics of the grass seeds to be reseeded; S4. Take images and videos of the species growing in the current plot and extract the dominant species; S5. Identify the bare ground surface of the current plot; S6. Calculate the area of the bare ground surface of the current plot, extract the boundary points, and convert the operation coordinates of the seeding machine; S7. Automatically match the grass seeds to be reseeded in the current plot and calculate the seeding rate; S8. Adjust the height position of the seeding unit, and open the corresponding seed box of the matched grass seeds for quantitative seed metering.

2. The precise mixed seeding method for grassland according to claim 1, characterized in that The step S1 also includes: S1-1. Fill the grass seeds to be reseeded, which are selected in advance, into each seed box of the seeding machine respectively; S1-2. Record the grass seed information Z in each seed box, classify and number it. The numbering content includes the family F, stress resistance K, seeding depth D, seeding rate M, and the number of the seed box to which it belongs. The grass seed information Z to be reseeded is represented by a data set and the data is entered and stored.

3. The precise mixed seeding method for grassland according to claim 1, characterized in that, The step S2 also includes: S2-1. Manually delimit the area to be operated in the Beidou navigation electronic map. The operation range is delimited as a quadrilateral, and the longitude and latitude coordinate information of the four vertices is extracted and the operation boundary conditions are stored; S2-2. Re-divide the delimited operation area. Take the working width B of the seeding machine as the width of the plot. Each row is divided into m plots, and each column is divided into n plots, and the longitude and latitude coordinates of the vertices of each small plot are stored.

4. The precise mixed seeding method for grassland according to claim 3, characterized in that, The step S3 also includes: S3-1. Obtain the real-time longitude and latitude position of the seeding machine through the Beidou navigation module, compare it with the vertex position coordinates of the small plot, judge the plot where the seeding machine is currently located, and mark this plot as "under operation"; S3-2. Set the soil relative humidity threshold RH, the pH threshold PH, and the temperature threshold T according to the drought, saline-alkali land, and high-temperature grading standards, and use a soil temperature and humidity sensor to measure the soil temperature T in the operation area true and the soil relative humidity RH true , use a soil pH sensor to measure the soil pH value PH in the operation area true , and compare the obtained T true , RH true , PH true with the pre-set temperature threshold T, relative humidity threshold RH, and pH threshold PH, and output the characteristics K of the grass seeds to be reseeded according to the comparison results d .

5. The method for precise mixed seeding of grassland according to claim 1, characterized in that The step S4 also includes: S4-1. Use a CMOS machine vision camera to take images and videos of the species in the current plot, obtain the species images from the videos, and rotate, crop and splice the selected images to obtain an image of the overall planting situation of the current plot; S4-2. Perform semantic segmentation on the overall planting situation image, mark the species types contained in the image, count the total number t of all species individuals and the number p of individuals of a single species, calculate the dominance index SIM of a single species, compare the species dominance rates, and the species with the highest dominance index is the dominant species in the current plot.

6. The method for precise mixed seeding and supplementary seeding of grassland according to claim 1, wherein, The step S5 also includes: S5-1. Draw a resistance-pressure calibration curve according to the calibration data set of the flexible piezoresistive sensor; S5-2. The seeding machine moves forward, uses a single-point acquisition circuit to obtain the real-time voltage value of the flexible piezoresistive sensor array, calculates the actual pressure, and records and stores 1 set of data every 0.5 seconds; S5-3. During the forward movement, monitor the pressure mutation points. When the real-time pressure value P measured by a certain row of sensors mutates and the value is approximately 0, it indicates that there is no plant growth in the front and there is an empty spot in this row, and mark this point as a bare point.

7. The precise mixed seeding method for grassland according to claim 6, characterized in that, The step S6 also includes: S6-1. Use OpenMV to capture surface images with the exposed point identified in step S5-3 as the center point. Starting from the smallest image that can capture the exposed point, gradually expand the coverage of the captured images. Expand the image by one pixel ring each time and capture one image. At the same time, extract the features of the outermost pixel ring of the image. Stop expanding and capturing when all the outermost pixel points of the image are covered by vegetation. S6-2. Use open-source image processing functions to perform edge detection and minimum rectangle contour fitting on the last captured image, extract the center point coordinates and the four vertex coordinates of the minimum rectangle, and calculate the area of the exposed ground surface. S6-3. Determine whether the area S of the exposed ground surface meets the conditions for reseeding. If the exposed area reaches the standard for reseeding, extract the horizontal boundaries of the minimum rectangle fitted in step S6-2 respectively. Take a point every row spacing L horizontally and convert it into the operation coordinates at the rear end of the reseeding machine. Otherwise, directly enter the next operation plot.

8. The precise mixed seeding method for grassland according to claim 4, characterized in that, The following is also included in step S7: S7-1. Match the characteristics K of the required reseeding grass seeds obtained in step S3-2 d with the family F to which the reseeding grass seeds obtained in step S4-3 belong d Return to the host computer, match with all the information of the reseeding grass seeds entered in step S2, extract the corresponding grass seed numbers, and obtain the seeding depth, seeding rate data and storage locations; S7-2. According to the grass seeding rate data called from the full attribute set, convert the exposed ground surface area of the current plot into the seeding mass G required for the grass seeds to be reseeded in the current plot.

9. The precise mixed seeding method for grassland according to claim 8, characterized in that, The following is also included in step S8: S8-1. According to the matching grass seeding depth extracted in step S7-1, control the hydraulic cylinder to extend or shorten. At the same time, the ultrasonic sensor installed under the four-bar linkage mechanism monitors the moving height h of the seeding unit in real time. When the height of the unit's descent or ascent meets the grass seeding depth requirement, the hydraulic cylinder stops extending or shortening. S8-2. When it is detected that the operation coordinates at the rear end of the reseeding machine enter the boundary range of the exposed area of the current plot, according to the matching grass seed storage position extracted in step S7-1, control the corresponding numbered servo motor to rotate, drive the rotating partition under the matching grass seed box to rotate, the seeds fall into the central seed box, and then are discharged by the seeding unit to complete the reseeding. At the same time, use the weighing sensor under the seed box to monitor the mass change g during the seed discharging process in real time. S8-3. When the mass change g is equal to the seeding mass G required for the grass seeds to be reseeded in the current plot, the servo motor returns to its original position, the rotating partition turns back to its original position, and the matching grass seed box stops discharging seeds outward.

Images