Grassland precision mixed repair sowing method based on ground vegetation detection

Abstract

The application provides a grassland precision mixed reseeding method based on ground vegetation detection, grass seed information to be reseeded selected in advance is input, classified and numbered and stored according to properties and storage positions; soil information and dominant species of a current land plot are acquired, and properties of required reseeded grass are output; bare ground of the current land plot is identified in real time, bare area is calculated and operation coordinates are converted, and whether the area reaches a standard requiring reseeding is judged; when the standard is not reached, a next operation land plot is entered; when the standard is reached, reseeded grass required by the current land plot is matched, seeding monomer height is adjusted according to seeding depth, and when the operation position is reached, the matched grass is quantitatively discharged according to seeding amount, and precision reseeding is completed. The method can realize real-time regulation and control of mixed reseeding operation, improve applicability of a reseeding machine to patchy and small-area degraded grassland and selectivity of reseeded species, and reduce damage of ditching and seeding to original grassland vegetation and further degradation of the grassland caused by improper selection of reseeded species.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent agricultural equipment, in particular to a grassland precision mixed reseeding method based on surface vegetation detection. BACKGROUND

[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 grassland, but also seriously damages its ecological function. Grassland mixed reseeding is considered to be one of the effective means to quickly restore degraded grassland. Mixed reseeding can increase the species diversity of grassland, restore the productivity of grassland, improve the quality of forage grass, and repair the ecology of grassland, which is of great significance to the economic and ecological benefits of grassland resources.

[0003] At present, the method of machine seeding is mostly used for mixed reseeding of grassland. However, the existing grassland reseeding machine is mostly modified from a farmland seeding machine, which is not suitable for reseeding of patchy and small-area degraded grassland, and does not have selectivity for mixed reseeding species. In addition, blind ditching and seeding may also damage the original grassland vegetation, further affecting the grassland ecological environment and exacerbating grassland degradation. With the progress of intelligent agricultural technology, there is an urgent need for a grassland precision mixed reseeding method that can autonomously identify bare surfaces and determine mixed reseeding species to complete selective reseeding. SUMMARY

[0004] In view of the deficiencies in the prior art, the present application builds on the prior art and constructs a grassland precision mixed reseeding method based on surface vegetation detection. The method identifies the reseeding area and selects the reseeding species for the grassland that needs to be mixed and reseeded, and controls the mixed reseeding process of the grassland in real time based on bare area, original dominant population, soil conditions and the like.

[0005] In order to achieve the purpose, the technical scheme adopted by the present application is as follows:

[0006] A grassland precision mixed reseeding method based on surface vegetation detection, comprising the following steps:

[0007] S1, input all the information of the grass seeds to be reseeded, and classify and number them according to the characteristics and storage location of the species;

[0008] S2, demarcate the operation area in the Beidou navigation electronic map, and divide the operation plots;

[0009] S3, obtain the soil condition information in the current plot, and determine the characteristics of the grass seeds to be reseeded;

[0010] S4, take pictures and videos of the species growing in the current plot, and extract the dominant species;

[0011] S5, identify the bare surface of the current plot;

[0012] S6, calculate the current bare land surface area, extract the boundary points, and convert the reseeding machine operation coordinates;

[0013] S7, automatically match the required reseeding grass seeds for the current land plot and calculate the seeding amount;

[0014] S8, adjust the height position of the seeding unit, open the corresponding seed box of the matched grass seed, and perform quantitative seed sorting.

[0015] The S1 specifically comprises:

[0016] S1-1, fill each seed box of the reseeding machine with the selected reseeding grass seeds in advance;

[0017] S1-2, record the grass seed information Z in each seed box of the reseeding machine, perform classification numbering, and the numbering content includes the belonging department F, the stress resistance K, the seeding depth D, the seeding amount M, and the belonging seed box number. The reseeding grass seed information Z is represented by a data set and data entry and storage are performed. 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 a department set, Z K is a stress resistance set, Z Q is a full attribute set, and n z is the total number of seed boxes carried by the reseeding machine;

[0018] Further, the classification numbering method in the step S1-1 is:

[0019] The belonging department is divided into three categories according to the legume family, the grass family, and the chrysanthemum family, and is represented by the numbers L, G, and C, i.e. F1, F2, …, F n ∈{L、G、C};The stress resistance selects three characteristics of drought resistance, cold resistance, and salt and alkali resistance, and is represented by the numbers d, c, and s, i.e. K1, K2, …, K n ∈{d、c、s};The seeding depth is numbered by the seeding depth value required by the grass seed agronomy, with the unit of cm; The seeding amount is numbered by the unit area seeding amount required by the grass seed agronomy, with the unit of kg / hm 2 ; The belonging seed box number corresponds to the seed box number.

[0020] The S2 specifically comprises:

[0021] S2-1, manually demarcate the to-be-operated region in the Beidou navigation electronic map, demarcate the operation range as a quadrilateral, extract the longitude and latitude coordinate information of the four vertices as [x a1 , y a1 ], [x a2 , y a1 ], [x a1 , y a2 ], [x a2 , y a2 ] as the operation boundary conditions and store them;

[0022] S2-2, perform secondary division on the demarcated operation region, taking the operation width B of the reseeding machine as the width of the plot, dividing each row into m plots, and 1 degree of longitude is approximately equal to 111 km, so

[0023]

[0024] wherein m is the number of horizontal plots; x a1 and x a2 are the vertex longitudes, and the unit is °; B is the operation width of the reseeding machine, and the unit is km;

[0025] each column is divided into n plots, and the vertex longitude and latitude coordinates of each small plot are stored as [x h i , y h j ] (i=1, 2, …, m+1; j=1, 2, …, n+1).

[0026] The S3 is specifically:

[0027] S3-1, obtain the real-time longitude and latitude position of the reseeding machine through the Beidou navigation module, compare it with the vertex position coordinates of the small plot, judge the plot where the reseeding machine is currently located, and mark this plot as “in operation”;

[0028] S3-2, set the soil relative humidity threshold RH, the acid-base degree threshold PH, and the 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 of the operation region, use the soil pH value sensor to measure the soil acid-base degree PH true of the operation region, compare the obtained T true , RH true , PH true with the temperature threshold T, the relative humidity threshold RH, and the acid-base threshold PH set in advance, and output the required reseeding grass type K d according to the comparison result.

[0029] Further, the required reseeding grass type Kd The judgment method is:

[0030] When T true RH true , PH true Only T true When the set threshold is exceeded, the required grass seed characteristics K are output d c, only RH true When the set threshold is exceeded, the required grass seed characteristics K are output d d, only PH true When the set threshold is exceeded, the required grass seed characteristics K are output d When two or more items exceed the threshold, PH true Priority, RH true Next, T true Finally, when all three items exceed the threshold, the required grass seed characteristics K are output d For s.

[0031] The S4 is specifically:

[0032] S4-1. Use a CMOS machine vision camera to shoot a video of species images in the current plot, obtain species images from the video, and rotate, crop, and splice the filtered images to obtain an image of the overall planting situation of the current plot;

[0033] S4-2. Perform semantic segmentation on the overall planting image, label the species types contained in the image, count the total number of individuals of all species t and the number of individuals of each species p, calculate the dominance index SIM of each species, compare the species dominance rates, and the species with the highest dominance index is the dominant species in the current plot;

[0034] Furthermore, the calculation method of the single species dominance index in step S4-2 is:

[0035] Berger-Parker dominance index calculation formula

[0036]

[0037] Where t is the total number of individuals of all species; p is the number of individuals of a single species.

[0038] S4-3, match the dominant species to be reseeded with grass species, and output the family F of the grass species to be reseeded d .

[0039] The S5 is specifically:

[0040] S5-1. Draw the resistance-pressure calibration curve based on the calibration data set of the flexible pressure-sensitive sensor and fit the linear formula within the calibration resistance range.

[0041]

[0042] Wherein, P is pressure, unit is N; R is sensor resistance value, unit is Ω;

[0043] S5-2, the reseeding machine moves forward, and the real-time voltage values U1, U2, …, U of the flexible piezoresistive sensor array are acquired by using a single-point acquisition circuit p , the resistance values R1, R2, …, R of the sensors are calculated according to Ohm's law p , the actual pressures P1, P2, …, P are calculated by fitting a linear formula p , p is the number of seeding units, and one group of data is recorded and stored every 0.5 seconds;

[0044] S5-3, in the forward movement process, a pressure mutation point is monitored, when the real-time pressure value P measured by a row of sensors is mutated and the value is approximately 0, it is indicated that there is no plant growth in front, a blank spot appears in the row, and the point is marked as a bare point.

[0045] Further, the arrangement mode of the flexible piezoresistive sensor in the step S5-2 is:

[0046] The flexible piezoresistive sensor at the front end of the machine is arranged every L, L is the row spacing of the reseeding machine, unit is m, and the arrangement position is close to the ground and is not more than 2 cm away from the ground, the flexible piezoresistive sensor is one-to-one corresponding to the position of the rear-end seeding unit, is kept on the same axis, the arrangement number of the sensor is the same as the number of the seeding units;

[0047] Further, the determination method of the mutation in the step S5-3 is:

[0048] When the sensor returns a new group of data P u , the previous five groups 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 groups of actual pressures is taken , u is the number of data groups, when

[0049]

[0050] It is determined that the mutation is generated in the row of sensors, wherein, is the average value of the actual pressure, P u is the latest measured actual pressure value, is the minimum pressure value generated when passing through the plant stem in the horizontal direction, unit is N.

[0051] The S6 is specifically:

[0052] S6-1, using OpenMV to shoot the ground image with the exposed point identified in S5-3 as the center point, taking the smallest image that can be shot of the exposed point as the starting image, gradually expanding the coverage of the collected image, expanding one circle of pixel points each time and shooting an image, while extracting the features of the outermost circle of pixel points of the image, and stopping the expansion and shooting when all the outermost circle of pixel points of the image are points covered with vegetation;

[0053] S6-2, using open source image processing functions to perform edge detection and minimum rectangular contour fitting on the last shot image, extracting the center point coordinates (x o , y o ) and four vertex coordinates (x o1 , y o1 ), (x o2 , y o1 ), (x o1 , y o2 ), (x o2 , y o2 ) of the minimum rectangle, and calculating the area S of the exposed ground

[0054]

[0055] wherein S is the exposed ground area, unit: m 2 ; x o1 , x o2 , y o1 , y o2 are the vertex coordinates, unit: m;

[0056] S6-3, judging whether the exposed ground area S meets the condition of needing to be reseeded, if the exposed area reaches the standard of needing to be reseeded, extracting the horizontal boundary of the minimum rectangle fitted in step S6-2, taking a point every row distance L horizontally, and converting it into the rear-end working coordinates of the reseeding machine, otherwise directly entering the next working plot.

[0057] The S7 is specifically:

[0058] S7-1, returning the required reseeding grass seed characteristics K d obtained in step S3-2, the family F d of the reseeding grass seed obtained in step S4-3 to the upper computer, and matching them with all the reseeding grass seed information entered in step S2, according to Boolean operation:

[0059]

[0060]

[0061] wherein Z K , Z FFor anti-reversibility set intersection, when the intersection of two items is 1, it is proved that the matching of the to-be-supplemental-grass seed is successful and the unique to-be-supplemental-grass seed can be obtained from the full attribute set Z Q The corresponding grass seed number is extracted from the full attribute set Z

[0062] S7-2, according to the grass seed sowing data in the full attribute set, the exposed land surface area of the current land plot is converted into the sowing quality G required by the to-be-supplemental-grass seed in the current land plot.

[0063] Further, the conversion method of the to-be-supplemental-grass seed sowing quality in step S7-2 is as follows:

[0064]

[0065] Wherein, G is the sowing quality required by the to-be-supplemental-grass seed in the current land plot, with the unit of g; S is the exposed land surface area of the current land plot, with the unit of m 2 ; M is the unit area sowing quality required by the grass seed, with the unit of kg / hm 2 .

[0066] The S8 is specifically as follows:

[0067] S8-1, according to the matching grass seed sowing depth extracted in step S7-1, the hydraulic cylinder is controlled to be elongated or shortened, and the ultrasonic sensor installed below the four-bar linkage mechanism is used to monitor the moving height h of the sowing monomer in real time, when the height of the monomer descending or ascending meets the requirement of the grass seed sowing depth, the hydraulic cylinder stops elongating or shortening;

[0068] S8-2, when it is detected that the working coordinate of the back end of the supplemental sowing machine enters the boundary range of the exposed area of the current land plot, according to the storage position of the matching grass seed extracted in step S7-1, the corresponding rudder is controlled to rotate, the rotating partition plate below the matching grass seed seed box is driven to rotate, the seed falls into the central seed box, and then is discharged by the sowing monomer, the supplemental sowing is completed, and the weighing sensor below the seed box is enabled to monitor the mass change g in the sowing process in real time;

[0069] S8-3, when the mass change g is equal to the sowing quality G required by the to-be-supplemental-grass seed in the current land plot, the rudder is homed, the rotating partition plate is rotated back to the original position, and the matching grass seed seed box stops sowing outward.

[0070] The grassland precise mixed supplemental sowing method involves the following key components:

[0071] The soil temperature and humidity sensor is used to obtain the soil water content and soil temperature;

[0072] The soil pH value sensor is used to obtain the soil acidity and alkalinity;

[0073] The CMOS machine vision camera is used to shoot the species image video;

[0074] Flexible pressure-sensitive sensor for judging the bare patch in the row and locating the exposed point;

[0075] OpenMV vision module for taking images of the exposed ground surface and calculating the exposed area;

[0076] Weighing sensor installed below the seed box for detecting the seed rate;

[0077] Rotary partition plate for separating the seed box and the central seed box, installed below the seed box;

[0078] Servo motor for opening and closing the seed box, controlling the size of the seed rate, and connected to the rotary partition plate;

[0079] Central seed box for storing the grass seeds to be supplemented and distributing them to each seeding unit;

[0080] Ultrasonic sensor for monitoring the lifting height of the seeding unit;

[0081] Hydraulic cylinder installed on the parallel four-bar mechanism of the seeding unit for adjusting the lifting and lowering height.

[0082] The technical scheme of the present application has the following beneficial effects: The grassland precision supplemental seeding technology based on ground vegetation detection is adopted in the present application, and the supplemental seeding operation is adjusted in real time according to the soil conditions, dominant species, and real-time changing exposed ground area during grassland improvement, which improves the applicability of the grassland supplemental seeding machine to patchy and small-area degraded grasslands and the selectivity of the supplemental seeding species, reduces the damage to the original grassland vegetation caused by blind ditching and seeding, and prevents further degradation of the grassland caused by improper selection of supplemental seeding species.

[0083] The present application utilizes machine vision and sensor information fusion technology to realize autonomous identification of the exposed ground surface and accurate calculation of the exposed area, converts the supplemental seeding machine operation coordinates, accurately locates the area that needs to be supplemented, selects the grass seeds to be supplemented according to the soil conditions, dominant species, and exposed area of the operation plot, calls the related agronomic information of the grass seeds, adjusts the seeding depth and quantity, and completes the selective supplemental seeding of the grassland precision supplemental seeding. BRIEF DESCRIPTION OF DRAWINGS

[0084] Figure 1 The flowchart of the embodiment of the present application;

[0085] Figure 2 The structural diagram of the supplemental seeding control system in the present application. DETAILED DESCRIPTION

[0086] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts should fall into the scope of the present application.

[0087] The embodiment of the present application provides a grassland precision mixed reseeding method based on surface vegetation detection, which can accurately perform mixed reseeding operation according to soil conditions, dominant species and real-time changing bare surface area, and avoids damage to the original grassland caused by blind ditching and further degradation of the grassland caused by improper reseeding species.

[0088] As shown in Figure 1 , the specific steps of the present application are as follows:

[0089] S1-1, fill the selected to-be-reseeded grass seeds in each seed box of the reseeding machine;

[0090] S1-2, record the grass seed information Z in each seed box of the reseeding machine, perform classification numbering, and the numbering content includes the belonging department F, the stress resistance K, the seeding depth D, the seeding amount M, the belonging seed box number, the to-be-reseeded grass seed information Z is expressed by a data set and data entry and storage are performed;

[0091] S2-1, manually demarcate the to-be-operated area in the Beidou navigation electronic map, demarcate the operation range as a quadrilateral, extract the longitude and latitude coordinate information of the four vertices, and store [x a1 , y a1 ], [x a2 , y a1 ], [x a1 , y a2 ], [x a2 , y a2 ] as the operation boundary conditions;

[0092] S2-2, perform secondary division on the demarcated operation area, take the operation width B of the reseeding machine as the width of the plot, divide each row into m plots and each column into n plots, and store the longitude and latitude coordinates of the vertices of each small plot as [x h i , y h j ];

[0093] S3-1, obtain the real-time longitude and latitude position of the reseeding machine through the Beidou navigation module, compare it with the position coordinates of the vertices of the small plot, judge the plot where the reseeding machine is currently located, and mark this plot as "in operation";

[0094] S3-2, according to the drought, saline-alkali land, high temperature classification standard, set the soil relative humidity threshold RH, pH value threshold PH, temperature threshold T, using soil temperature and humidity sensor to determine the soil temperature T of the working area true and soil relative humidity RH true , using soil pH sensor to determine the soil pH value PH of the working area true , compare the acquired T true , RH true , PH true with the temperature threshold T, relative humidity threshold RH, pH value threshold PH set in advance, and output the required grass seed characteristics K according to the comparison result d ;

[0095] S4-1, using CMOS machine vision camera to shoot the species image video in the current plot, obtaining the species image from the video, and rotating, cropping and splicing the screened image to obtain the overall planting situation image of the current plot;

[0096] S4-2, semantic segmentation is performed on the overall planting situation image, the species type contained in the image is marked, and the total number of all species individuals t and the number of single species individuals p are counted, the dominance index SIM of single species is calculated, and the dominant species is compared with the dominant rate, and the highest dominance index is the dominant species of the current plot;

[0097] S4-3, the dominant species is matched with the grass seed, and the required grass seed belongs to the family F d ;

[0098] S5-1, draw the resistance-pressure calibration curve according to the calibration data set of flexible pressure sensor, and fit the linear formula in the calibration resistance interval;

[0099] S5-2, the overseeding machine moves forward, and the real-time voltage values U1, U2, …, U p of the flexible piezoresistive sensor array are obtained by using single-point acquisition circuit p , the resistance values R1, R2, …, R p of the sensor are calculated according to Ohm's law, and the actual pressures P1, P2, …, P p are calculated by fitting the linear formula, p is the number of seeding units, and a group of data is recorded and stored every 0.5 seconds;

[0100] S5-3, during the forward movement, the pressure mutation point is monitored, 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 in front, and there is a blank spot in this row, which is marked as a bare point;

[0101] S6-1, use OpenMV to shoot the ground image centered on the exposed point identified in S5-3, use the smallest image that can be shot of the exposed point as the starting image, gradually expand the coverage of the collected image, expand one pixel circle at a time and shoot an image, and extract the features of the outermost pixel points of the image at the same time, when all the outermost pixel points of the image are points covered with vegetation, stop expanding and shooting;

[0102] S6-2, use open source image processing functions to perform edge detection and minimum rectangular contour fitting on the last shot image, extract the center point coordinates (x o , y o ) and 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;

[0103] S6-3, determine whether the exposed ground area S meets the condition of needing to be replanted, if the exposed area reaches the standard of needing to be replanted, extract the horizontal boundary of the minimum rectangle fitted in step S6-2, take a point every row distance L horizontally, and convert it into the rear-end working coordinates of the reseeding machine, otherwise go directly to the next working plot.

[0104] S7-1, return the required reseeding grass characteristics K d obtained in step S3-2, the family F d of the grass to be reseeded obtained in step S4-3 to the host computer, and match them with all the reseeding grass information entered in step S2, and if the matching is successful, extract the corresponding grass number from the full attribute set Z Q to obtain the planting depth, planting quantity data and storage location;

[0105] S7-2, according to the grass planting quantity data called in the full attribute set, convert the exposed ground area of the current plot into the required planting quality G of the reseeding grass in the current plot.

[0106] S8-1, according to the matching grass planting depth extracted in step S7-1, control the extension or shortening of the hydraulic cylinder, and at the same time, the ultrasonic sensor installed below the four-bar linkage mechanism monitors the moving height h of the planting unit in real time, when the height of the unit descending or ascending meets the grass planting depth requirement, the hydraulic cylinder stops extending or shortening;

[0107] S8-2, when the back-end operation coordinates of the reseeding machine are detected to enter the boundary range of the current bare area of the land plot, according to the matching grass seed storage position extracted in step S7-1, the corresponding rudder is controlled to rotate, the rotating partition plate below the matching grass seed box is driven to rotate, the seed falls into the central seed box, and then is discharged by the seeding unit, the reseeding is completed, and the weighing sensor below the seed box is enabled to monitor the mass change g in the seed discharge process in real time;

[0108] S8-3, when the mass change g is equal to the required seeding mass G of the grass seed to be reseeded in the current land plot, the rudder is returned to the original position, the rotating partition plate is turned back to the original position, and the matching grass seed box stops discharging seeds outward.

[0109] As shown in Figure 2 , the grassland precision reseeding method involves the following key components:

[0110] A soil temperature and humidity sensor is used to obtain the soil water content and soil temperature.

[0111] A soil pH sensor is used to obtain the soil acidity and alkalinity.

[0112] A CMOS machine vision camera is used to take images and videos of the species.

[0113] A flexible pressure-sensitive sensor is used to determine the empty spots in the row and locate the bare spots.

[0114] An OpenMV vision module is used to take images of the bare ground surface and calculate the bare area.

[0115] A weighing sensor is installed below the seed box to detect the seed discharge amount.

[0116] A rotating partition plate is used to separate the seed box and the central seed box and is installed below the seed box.

[0117] A rudder is used to open and close the seed box and control the size of the seed discharge amount, and is connected to the rotating partition plate.

[0118] A central seed box is used to store the grass seed to be reseeded and distribute it to each seeding unit.

[0119] An ultrasonic sensor is used to monitor the lifting height of the seeding unit.

[0120] A hydraulic cylinder is installed on the parallel four-bar mechanism of the seeding unit to adjust the lifting and lowering height.

[0121] The specific implementation process of the present application can be clearly understood by those skilled in the art through the description of the above embodiments. It should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application 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 make equivalent replacement for part 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 embodiments of the present application.

Claims

1. A grassland precision mixed seeding method based on surface vegetation detection, characterized in that: The method comprises the following steps: S1. Enter the information of all grass seeds to be reseeded and classify and number them according to species characteristics and storage locations; S2. Define the operation area and divide the operation plots on the Beidou navigation electronic map; S3, obtaining soil condition information in the current plot and determining the characteristics of the grass seeds to be reseeded; said step S3 also includes: S3-1. Obtain the real-time latitude and longitude of the reseeding machine through the Beidou navigation module, compare it with the coordinates of the vertex position of the small plot, determine the plot where the reseeding machine is currently located, and mark this plot as "in operation"; S3-2. Set the soil relative humidity threshold RH, pH threshold PH, and temperature threshold T according to the drought, saline-alkali, and high temperature classification standards, and use the soil temperature and humidity sensor to measure the soil temperature T in the operation area. true and soil relative humidity RH true , use soil pH sensor to measure the pH of the soil in the working area true , the obtained T true RH true , PH true Compare with the pre-set temperature threshold T, relative humidity threshold RH, and pH threshold PH, and output the required grass seed characteristics K based on the comparison results d ; S4, shooting images and videos of species growing in the current plot to extract dominant species; said step S4 also includes: S4-1. Use a CMOS machine vision camera to shoot a video of species images in the current plot, obtain species images from the video, and rotate, crop, and splice the filtered images to obtain an image of the overall planting situation of the current plot; S4-2. Perform semantic segmentation on the overall planting image, label the species types contained in the image, count the total number of individuals of all species t and the number of individuals of each species p, calculate the dominance index SIM of each 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 dominant species to be reseeded with grass species, and output the family F of the grass species to be reseeded d ; S5. identifying the exposed surface of the current plot; S6. Calculate the exposed surface area of ​​the current plot, extract the boundary points, and convert the operation coordinates of the reseeder; S7, automatically matching the grass seeds required for the current plot and calculating the sowing rate; said step S7 also includes: S7-1, the required reseeding grass seed characteristics K obtained in step S3-2 d , the family F to which the grass species to be reseeded obtained in step S4-3 belongs d Return to the host computer, match with all the grass seed information to be re-seeded entered in step S1, extract the corresponding grass seed number, and obtain the sowing depth, sowing amount data and storage location; S7-2. Based on the grass seed sowing rate data retrieved from the full attribute collection, the exposed surface area of ​​the current plot is converted into the sowing mass G required for the grass seed to be re-sown in the current plot; S8. Adjust the height position of the sowing unit and open the corresponding seed box that matches the grass seed for quantitative sowing.

2. The grassland precision mixed seeding method according to claim 1, characterized in that: The step S1 further includes: S1-1, filling each seed box of the reseeding machine with grass seeds selected in advance to be reseeded; S1-2. Record the grass seed information Z in each seed box and classify and number them. The numbering content includes the family F, stress resistance K, sowing depth D, sowing amount M, and seed box number. The grass seed information Z to be re-sown is represented by a data set and the data is entered and stored.

3. The grassland precision mixed seeding method according to claim 1, characterized in that: The step S2 further includes: S2-1. Manually demarcate the operation area in the Beidou navigation electronic map. The operation range is demarcated as a quadrilateral. The latitude and longitude coordinates of the four vertices are extracted and the operation boundary conditions are stored. S2-2. Divide the designated operation area into two parts. Use the operating width B of the reseeder 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.

4. The grassland precision mixed seeding method according to claim 1, characterized in that: The step S5 further includes: S5-1. Draw a resistance-pressure calibration curve based on the calibration data set of the flexible pressure-sensitive sensor; S5-2, the reseeding 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 one set of data every 0.5 seconds; S5-3. During the forward movement, monitor the pressure mutation point. When the real-time pressure value P measured by a row of sensors suddenly changes and the value is close to 0, it means that there is no plant growth ahead and a blank spot appears in this row. Mark this point as a bare spot.

5. The grassland precision mixed seeding method according to claim 4, characterized in that: The step S6 further includes: S6-1. Use OpenMV to capture a surface image centered on the exposed point identified in step S5-3. Starting with the smallest image that can be captured at the exposed point, gradually expand the coverage of the captured image, capturing an image each time by expanding a circle of pixels. Extract features from the outermost pixel of the image. Stop expanding and capturing when all the pixels in the outermost circle 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 coordinates of the center point and four vertices of the minimum rectangle, and calculate the area of ​​the exposed surface. S6-3. Determine whether the exposed surface area S meets the conditions for reseeding. If the exposed area meets the standards for reseeding, extract the horizontal boundaries of the minimum rectangle fitted in step S6-2, take a point every row spacing L, and convert it into the back-end operation coordinates of the reseeder. Otherwise, go directly to the next operation plot.

6. The grassland precision mixed seeding method according to claim 1, characterized in that: The step S8 further includes: S8-1. Based on the matching grass seed sowing depth obtained in step S7-1, the hydraulic cylinder is controlled to extend or retract. Simultaneously, an ultrasonic sensor installed below the four-bar linkage monitors the moving height h of the sowing unit in real time. When the height of the unit descends or ascends to meet the grass seed sowing depth requirement, the hydraulic cylinder stops extending or retracting. S8-2. When the rear-end operating coordinates of the reseeding machine are detected to have entered the boundary of the exposed area of ​​the current plot, the corresponding numbered servo is controlled to rotate according to the matching grass seed storage location extracted in step S7-1, driving the rotating partition below the matching grass seed box to rotate, causing the seeds to fall into the central seed box and then be discharged by the sowing unit, completing the reseeding. At the same time, a weighing sensor below the seed box is used to monitor the mass change g during the seeding process in real time; S8-3. When the mass change g is equal to the sowing mass G required for the grass seeds to be re-sown in the current plot, the servo returns to its original position, the rotary partition returns to its original position, and the matching grass seed box stops discharging seeds outward.

Images