An automatic monitoring method and system suitable for economic crops
Through automated monitoring and prediction of plant height changes, combined with plant, environment and soil data, precise control of the height of economic crops is achieved, solving the problems of low monitoring and regulation efficiency and large errors in the existing technology, and improving crop yield and quality.
Patent Information
- Application Number
- CN202410963550.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-07-18
AI Technical Summary
The existing technology cannot effectively monitor and regulate the height of cash crop plants, resulting in loss of yield and quality. The traditional methods are inefficient and have large errors, making it difficult to meet the precise and efficient management needs of modern agriculture.
Using automated monitoring methods, by obtaining plant, environment and soil status data, using a pre-trained regression network model to predict plant height changes, and screen abnormal areas based on plant growth amount, determine the dosage ratio of plant growth regulators, and realize the plant's Highly accurate regulation.
Accurate control of plant height has been achieved, the ventilation and light transmittance in the field has been improved, lodged, and the yield and quality of cash crops have been ensured.
Smart Images

Figure CN118504795B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of agricultural technology, and more specifically, to an automatic monitoring method and system suitable for cash crops. Background Art
[0002] Cotton and other crops are important economic crops worldwide, and their yield and quality have a significant impact on the textile industry and agricultural economy. During the planting process, the measurement and control of plant height is one of the key factors affecting yield and quality. Plants that are too high are prone to lodging, especially in windy and rainy weather conditions, which can lead to serious yield losses. In addition, too high plant height will also affect field ventilation and light conditions, increase the probability of disease and insect pests, and thus affect yield and quality. Traditional plant height measurement methods mainly rely on manual measurement, which has many disadvantages, such as low measurement efficiency, large errors, high labor intensity, and is not suitable for large-scale planting environments. In addition, the method of manually controlling plant height also has problems of inaccuracy and untimeliness, which makes it difficult to meet the needs of modern agriculture for precise and efficient management.
[0003] At present, there is a lack of plant height prediction methods. Although there are some relevant technical documents, for example, the patent with announcement number CN113962474A discloses a method, processor, storage medium and computer program product for predicting plant height. The method predicts the plant height according to the humidity and accumulated temperature of the farmland. However, research and practical application of the above method and the prior art have found that the above method and the prior art have at least the following defects:
[0004] (1) The monitoring method is single and the monitoring data is not comprehensive enough, which makes it difficult to obtain complete data on internal and external influences on plant height, making it difficult to determine the growth amount of plant height in different growth stages and growth environments, and thus difficult to provide important data support for regulating plant height;
[0005] (2) It is impossible to screen out abnormal planting sub-areas based on the growth of plant height. Furthermore, it is impossible to determine the dosage ratio of plant growth regulators required for different abnormal planting sub-areas, and thus it is impossible to regulate the plant height. It is difficult to avoid shading inside the canopy caused by excessively tall plants, and it is impossible to improve the ventilation and light transmittance of the planting sub-areas, which is not conducive to ensuring the yield and quality of economic crops. Summary of the invention
[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an automated monitoring method and system applicable to cash crops.
[0007] To achieve the above object, the present invention provides the following technical solutions:
[0008] An automated monitoring method applicable to cash crops, the method comprising:
[0009] S101: within a predetermined time period [TN, T], obtaining plant height influence characteristic data of plants in the planting sub-region; the plant height influence characteristic data includes plant state data, environmental state data and soil state data in the planting sub-region at each moment, where T and N are integers greater than zero;
[0010] S102: Retrieving the initial average plant height of the plants in the planting area, inputting the plant height influencing feature data into a pre-trained plant height prediction model, and predicting the average plant height change of the plants at the future time T+N;
[0011] S103: Calculate the average growth amount of the plants in the planting area based on the initial average plant height and the changed average plant height of the plants.
[0012] Furthermore, the plant status data includes the average plant height, plant variety and plant density coefficient of the plants in the planting area; the environmental status data includes environmental temperature, environmental humidity and rainfall; and the soil status data includes soil temperature, soil humidity and soil status coefficient.
[0013] Furthermore, the logic for obtaining the density coefficient of the plant is as follows:
[0014] When the drone is controlled to arrive above the airspace of the planting sub-area, a bird's-eye view of the plants in the planting sub-area is obtained;
[0015] Convert the top view from the standard RGB color space to the HSV color space, use the color threshold method to perform color segmentation on the converted top view, separate the plant area in the top view from the background according to the target color of the plant's growth stage, and define the mask G of the target color:
[0016] ;
[0017] Where: For top view at position The pixel value at , 1 represents plant pixels, and 0 represents background pixels;
[0018] According to the mask G of the target color, the coverage area A of the plant area is calculated, and the coverage area A is normalized to obtain the coverage rate C of the plant area:
[0019] ;
[0020] ;
[0021] Where: is the mask value of the target color, M and N represent the width and height of the top view respectively;
[0022] Get the actual coverage area and image resolution of the top view, and calculate the density coefficient of the plant based on the actual coverage area and image resolution:
[0023] ;
[0024] Where: D is the density coefficient of the plant, S is the actual coverage area, and is the resolution of the top view.
[0025] Furthermore, the logic for obtaining the soil state coefficient is as follows:
[0026] Obtain the actual soil porosity, actual soil nitrogen content and actual soil pH of the planting sub-area;
[0027] The process of obtaining the actual porosity of the soil is as follows:
[0028] Obtain soil samples from the planting sub-area, and insert four electrodes into the soil samples at equal intervals to a predetermined depth;
[0029] Use the two outer electrodes to conduct a known current R into the soil and measure the voltage V through the two inner electrodes;
[0030] Substitute the current R and voltage V into the pre-built porosity calculation formula to obtain the actual porosity of the soil;
[0031] Wherein, the porosity calculation formula is as follows:
[0032] ;
[0033] Where: ∅ is the actual porosity of the soil, K is the geometric factor, α is the empirical constant, F is the formation factor, and β is the empirical constant;
[0034] Substituting the actual soil porosity, actual soil nitrogen content and actual soil pH into the pre-built mathematical model to obtain the soil state coefficient;
[0035] The mathematical model is expressed as follows:
[0036] ;
[0037] Where: SC is the soil state coefficient, μ is the actual nitrogen content of the soil, is the soil reference nitrogen content, ∅ is the actual soil porosity, is the soil reference porosity, pH is the actual soil acidity and alkalinity, is the reference pH of the soil, and is a correction factor greater than zero, > .
[0038] Furthermore, the training logic of the plant height prediction model is as follows:
[0039] Acquire historical plant height training data, and divide the historical plant height training data into a plant height prediction training set and a plant height prediction test set; the historical plant height training data includes plant height influencing feature data and its corresponding changed average plant height;
[0040] Constructing a first regression network, taking the plant height influencing feature data in the plant height prediction training set as input data of the first regression network, and taking the changing average plant height in the plant height prediction training set as output data of the first regression network, training the first regression network, and obtaining an initial plant height prediction network;
[0041] The plant height prediction test set is used to verify the model of the initial plant height prediction network, and the initial plant height prediction network with a preset first test error is output as the plant height prediction model.
[0042] Furthermore, the method further comprises:
[0043] S201: Obtain the average growth amount of the plants in all planting sub-areas, and the growth stage of the plants in each planting sub-area at the future time T+N, where the number of all planting sub-areas is Q, which is an integer greater than zero;
[0044] S202: Screening out a plurality of abnormally implanted sub-regions based on the average growth amount and the growth stage, and obtaining dose impact characteristic data of each abnormally implanted sub-region;
[0045] S203: inputting the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator;
[0046] S204: Control the spraying equipment according to the dosage ratio to spray the plants in the abnormal planting area with plant growth regulator.
[0047] Furthermore, the screening out of a plurality of abnormally planted sub-regions includes:
[0048] a1: Extract the average growth of the r-th planting sub-region, and extract the growth stage of the r-th planting sub-region at the future time T+N;
[0049] a2: Based on the preset relationship between the growth stage and the growth amount interval threshold, the corresponding growth amount threshold is obtained, and the average growth amount of the r-th planting sub-area is compared with the corresponding growth amount threshold; if the average growth amount is greater than or equal to the corresponding growth amount threshold, the r-th planting sub-area is determined to be a non-abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1; if the average growth amount is less than the corresponding growth amount threshold, the r-th planting sub-area is determined to be an abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1;
[0050] a3: Repeat the above steps a1 to a2 until r=Q, then end the loop and obtain multiple abnormal planting sub-areas.
[0051] Furthermore, the dose impact characteristic data include the density coefficient, average growth amount, growth rate and growth stage of the plants in the abnormal seedling area;
[0052] The training process of the dose regression model is as follows:
[0053] Obtaining historical dose spraying training data, and dividing the historical dose spraying training data into a dose regression training set and a dose regression test set; the historical dose spraying training data includes dose impact characteristic data and its corresponding plant growth regulator dosage ratio;
[0054] Constructing a second regression network, taking the dose effect feature data in the dose regression training set as input data of the second regression network, and taking the dosage ratio in the dose regression training set as output data of the second regression network, training the second regression network, and obtaining an initial dose regression network;
[0055] The initial dose regression network is model verified using the dose regression test set, and the initial dose regression network with a value less than or equal to the preset second test error is output as the dose regression model.
[0056] An automatic monitoring system suitable for cash crops, comprising:
[0057] The data acquisition module is used to obtain the plant height influence characteristic data of the plants in the planting sub-area within a predetermined time period [TN, T]; the plant height influence characteristic data includes the plant state data, environment state data and soil state data in the planting sub-area at each moment, and T and N are integers greater than zero;
[0058] The plant height prediction module is used to retrieve the initial average plant height of the plants in the planting area, input the plant height influencing feature data into the pre-trained plant height prediction model, and predict the average plant height change of the plants at the future time T+N;
[0059] The increment calculation module is used to calculate the average growth amount of plants in the planting area based on the initial average plant height and the changed average plant height of the plants.
[0060] Furthermore, the system further comprises:
[0061] The data retrieval module is used to obtain the average growth of plants in all planting sub-areas and the growth stage of plants in each planting sub-area at the future time T+N. The number of all planting sub-areas is Q, which is an integer greater than zero.
[0062] An abnormal screening module is used to screen out multiple abnormal planting sub-areas based on average growth amount and growth stage, and obtain dose impact characteristic data of each abnormal planting sub-area;
[0063] A ratio regression module is used to input the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator;
[0064] The plant height control module is used to control the spraying equipment to spray plant growth regulators on plants in the abnormal planting area according to the dosage ratio.
[0065] Compared with the prior art, the present invention has the following beneficial effects:
[0066] The present application discloses an automated monitoring method and system suitable for economic crops, comprising: obtaining plant height influence characteristic data of plants in a planting sub-area within a given time period [TN, T]; retrieving the initial average plant height of plants in the planting sub-area, inputting the plant height influence characteristic data into a pre-trained plant height prediction model, and predicting the changed average plant height of plants at the future time T+N; calculating the average growth amount of plants in the planting sub-area based on the initial average plant height and the changed average plant height; based on the above process, the present invention can determine the growth amount of plant height in different generation stages and generation environments; further, it is beneficial to provide important data support for regulating the height of crop plants; in addition, the present invention screens out abnormal planting sub-areas based on the average growth amount, and determines the dosage ratio of plant growth regulators according to the dosage influence characteristic data of the abnormal planting sub-areas, thereby facilitating the regulation of plant height, and further, it is beneficial to avoid lodging caused by excessive plant height, improve the ventilation and light transmittance of abnormal planting sub-areas, and thus ensure the yield and quality of economic crops. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] Figure 1 A flow chart of an automated monitoring method applicable to cash crops provided by the present invention;
[0068] Figure 2 A module structure diagram of an automatic monitoring system suitable for economic crops provided by the present invention. DETAILED DESCRIPTION
[0069] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Example
[0070] See also Figure 1 As shown, this embodiment discloses an automated monitoring method applicable to cash crops, the method comprising:
[0071] S101: within a predetermined time period [TN, T], obtaining plant height influence characteristic data of plants in the planting sub-region; the plant height influence characteristic data includes plant state data, environmental state data and soil state data in the planting sub-region at each moment, where T and N are integers greater than zero;
[0072] It should be understood that: the predetermined time period is a fixed time period, which can be in units of hours, such as 1 hour, 2 hours or 3 hours, or in units of days, such as 1 day, 2 days or 3 days; there are Q planting sub-areas, which are obtained by equally dividing the planting area of the plant (such as cotton); each planting sub-area contains a number of plant (such as cotton) plants;
[0073] The plant status data include the average plant height, plant variety and plant density coefficient of the plants in the planting sub-area;
[0074] It should be noted that: the average height of the plants in the planting sub-area at each moment is achieved by manual measurement or existing intelligent measurement methods, such as an image-based plant height measurement method, an ultrasonic radar-based plant height measurement method, or a point cloud data-based plant height measurement method; and the plant varieties of the plants in each planting sub-area are the same, and the plant varieties of the plants in each planting sub-area are pre-entered into the system database by the technicians;
[0075] Wherein, the environmental status data includes environmental temperature, environmental humidity and rainfall; the environmental status data is collected by temperature sensors, humidity sensors and rainfall sensors exposed to the environment;
[0076] The soil state data includes soil temperature, soil moisture and soil state coefficient; the soil temperature and soil moisture are collected by temperature sensors and humidity sensors buried in the soil;
[0077] In implementation, the logic for obtaining the density coefficient of the plant is as follows:
[0078] When the drone is controlled to arrive above the airspace of the planting sub-area, a bird's-eye view of the plants in the planting sub-area is obtained;
[0079] It should be noted that: controlling the drone to arrive above the airspace of the planting sub-area can be achieved through a preset flight route or can be achieved manually by a technician; the drone is equipped with an airborne camera for collecting a bird's-eye view of the plants in the planting sub-area;
[0080] The top view is converted from the standard RGB color space to the HSV color space, and the color threshold method is used to perform color segmentation on the converted top view. According to the target color of the plant's growth stage, the plant area in the top view is separated from the background, and the mask G of the target color is defined:
[0081] ;
[0082] Where: For top view at position The pixel value at , 1 represents plant pixels, and 0 represents background pixels;
[0083] It should be noted that: different plants have different color characteristics at different growth stages, so it is necessary to dynamically select the target color according to the actual situation to ensure the accuracy of the obtained plant area;
[0084] For example, the color characteristic of cotton in the seedling stage is green, so a green channel is needed to distinguish cotton plants and leaves; in the bud stage, the flowers may appear white or yellow, so a combination of multiple color channels is needed to distinguish plants and leaves; therefore, it is understandable that when facing different plants and different growth stages, it should be handled according to the actual situation;
[0085] According to the mask G of the target color, the coverage area A of the plant area is calculated, and the coverage area A is normalized to obtain the coverage rate C of the plant area:
[0086] ;
[0087] ;
[0088] Where: is the mask value of the target color, M and N represent the width and height of the top view respectively;
[0089] Get the actual coverage area and image resolution of the top view, and calculate the density coefficient of the plant based on the actual coverage area and image resolution:
[0090] ;
[0091] Where: D is the density coefficient of the plant, S is the actual coverage area, and is the resolution of the top view;
[0092] It should be noted that the actual coverage area refers to the real physical area covered by the top view, which can be obtained by actual measurement by technicians using surveying and mapping equipment; or determined based on the actual records of the planting area in the system database; and the resolution and It refers to the actual physical size of each pixel in the top view, which can be determined by the camera parameters and shooting height;
[0093] In implementation, the logic for obtaining the soil state coefficient is as follows:
[0094] Obtain the actual soil porosity, actual soil nitrogen content and actual soil pH of the planting sub-area;
[0095] In a specific embodiment, the process of obtaining the actual porosity of the soil is as follows:
[0096] Obtain soil samples from the planting sub-area, and insert four electrodes into the soil samples at equal intervals to a predetermined depth;
[0097] It should be noted that the predetermined depth is set by technicians based on experimental data or experience;
[0098] Use the two outer electrodes to conduct a known current R into the soil and measure the voltage V through the two inner electrodes;
[0099] Substitute the current R and voltage V into the pre-built porosity calculation formula to obtain the actual porosity of the soil;
[0100] Wherein, the porosity calculation formula is as follows:
[0101] ;
[0102] Where: ∅ is the actual porosity of the soil, K is the geometric factor, α is the empirical constant, F is the formation factor, and β is the empirical constant;
[0103] It should be noted that: the actual porosity of the soil is calculated by the resistivity method, and the geometric factor is determined according to the electrode arrangement. For example, the time formula of the geometric factor of the electrodes arranged at equal intervals is: K=2πa, where a is the interval; and the formation factor and formation factor are obtained through specific experiments; among them, the formation factor reflects the complexity of the pore structure in the soil, and the empirical constant reflects the relationship between the soil pore structure and resistivity;
[0104] In another specific embodiment, the actual porosity of the soil can be achieved by a volume ratio method or a gas exhaust method;
[0105] Substituting the actual soil porosity, actual soil nitrogen content and actual soil pH into the pre-built mathematical model to obtain the soil state coefficient;
[0106] The mathematical model is expressed as follows:
[0107] ;
[0108] Where: SC is the soil state coefficient, μ is the actual nitrogen content of the soil, is the soil reference nitrogen content, ∅ is the actual soil porosity, is the soil reference porosity, pH is the actual soil acidity and alkalinity, is the reference pH of the soil, and is a correction factor greater than zero, > ;
[0109] It should be noted that the actual nitrogen content of the soil and the actual pH of the soil are collected by a nitrogen content sensor and a pH sensor, respectively.
[0110] S102: Retrieving the initial average plant height of the plants in the planting area, inputting the plant height influencing feature data into a pre-trained plant height prediction model, and predicting the average plant height change of the plants at the future time T+N;
[0111] In implementation, the training logic of the plant height prediction model is as follows:
[0112] Acquire historical plant height training data, and divide the historical plant height training data into a plant height prediction training set and a plant height prediction test set; the historical plant height training data includes plant height influencing feature data and its corresponding changed average plant height;
[0113] It should be noted that: the plant height influence characteristic data in the historical plant height training data is obtained by collection and calculation, and the details are described above, which will not be repeated here; the change average plant height in the historical plant height training data is obtained by actual measurement by technicians;
[0114] Constructing a first regression network, taking the plant height influencing feature data in the plant height prediction training set as input data of the first regression network, and taking the changing average plant height in the plant height prediction training set as output data of the first regression network, training the first regression network, and obtaining an initial plant height prediction network;
[0115] The initial plant height prediction network is verified using the plant height prediction test set, and the initial plant height prediction network with a value less than or equal to a preset first test error is output as the plant height prediction model;
[0116] It should be noted that: the first regression network is specifically one of the regression algorithm models such as random forest, support vector machine, polynomial regression, linear regression, decision tree or neural network.
[0117] S103: Calculating an average growth amount of plants in the planting area based on an initial average plant height and a changed average plant height of the plants;
[0118] In practice, the calculation formula for the average growth of plants in the planting sub-area is as follows:
[0119] ;
[0120] Where: ZH is the average growth rate, is the average plant height change, is the initial average plant height.
[0121] In a specific embodiment, the method further comprises:
[0122] S201: Obtain the average growth amount of the plants in all planting sub-areas, and the growth stage of the plants in each planting sub-area at the future time T+N, where the number of all planting sub-areas is Q, which is an integer greater than zero;
[0123] It should be noted that different types of plants have different growth stages. For example, cotton includes the seedling stage, bud stage, boll stage and mature stage. The growth rate of cotton plants is different in different growth stages, and the conditions for plant height control are also different. Therefore, it is necessary to determine the growth stage of the plants in the planting area, which can generally be calculated through experience and time.
[0124] S202: Screening out a plurality of abnormally implanted sub-regions based on the average growth amount and the growth stage, and obtaining dose impact characteristic data of each abnormally implanted sub-region;
[0125] In implementation, the screening out of multiple abnormally planted sub-areas includes:
[0126] a1: Extract the average growth of the r-th planting sub-region, and extract the growth stage of the r-th planting sub-region at the future time T+N;
[0127] a2: Based on the preset relationship between the growth stage and the growth amount interval threshold, the corresponding growth amount threshold is obtained, and the average growth amount of the r-th planting sub-area is compared with the corresponding growth amount threshold; if the average growth amount is greater than or equal to the corresponding growth amount threshold, the r-th planting sub-area is determined to be a non-abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1; if the average growth amount is less than the corresponding growth amount threshold, the r-th planting sub-area is determined to be an abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1;
[0128] It should be noted that: a plurality of preset relationships between growth stages and growth interval thresholds are pre-stored in the system database, and in each preset relationship there is a growth interval threshold associated with the growth stage; the preset relationship is set by a technician based on specific experiments;
[0129] It should also be noted that: when the rth planting sub-area is determined to be an abnormal planting sub-area, it means that the corresponding planting sub-area has excessive plant growth, and plant height suppression must be carried out in a timely manner to avoid a series of hazards caused by excessive plant height (such as large-scale plant lodging, etc.);
[0130] a3: Repeat the above steps a1 to a2 until r=Q, then end the loop and obtain multiple abnormal planting sub-regions;
[0131] The dose impact characteristic data include the density coefficient, average growth amount, growth rate and growth stage of the plants in the abnormal planting area;
[0132] The calculation formula for the growth rate of plants in the abnormal planting area is as follows:
[0133] ;
[0134] Where: ZS is the growth rate, It is the duration from TN to T+N.
[0135] S203: inputting the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator;
[0136] In implementation, the training process of the dose regression model is as follows:
[0137] Obtaining historical dose spraying training data, and dividing the historical dose spraying training data into a dose regression training set and a dose regression test set; the historical dose spraying training data includes dose impact characteristic data and its corresponding plant growth regulator dosage ratio;
[0138] It should be noted that the dose effect characteristic data in the historical dose spraying training data is obtained through collection and calculation, and the detailed parameters are described above and will not be repeated here; and the dosage ratio of the plant growth regulator is determined by the technicians in advance according to the experimental data;
[0139] It should be understood that plant growth regulators are a class of chemical substances, including chemical control agents and water, which are used to adjust the hormone balance in plants to affect the growth and development of plants. The chemical control agent is mepiquat. The rational use of plant growth regulators can effectively control the growth height of crops, enhance stress resistance, and improve yield and quality.
[0140] Constructing a second regression network, taking the dose effect feature data in the dose regression training set as input data of the second regression network, and taking the dosage ratio in the dose regression training set as output data of the second regression network, training the second regression network, and obtaining an initial dose regression network;
[0141] The initial dose regression network is model verified using the dose regression test set, and the initial dose regression network with a second test error less than or equal to a preset value is output as the dose regression model;
[0142] It should be noted that: like the above-mentioned first regression network, the second regression network is specifically one of the regression algorithm models such as random forest, support vector machine, polynomial regression, linear regression, decision tree or neural network.
[0143] S204: Controlling the spraying equipment to spray the plants in the abnormal planting area with plant growth regulator according to the dosage ratio;
[0144] It should be noted that: the spraying equipment is arranged around the field of the planting sub-area, and when the dosage ratio is obtained, the plant growth regulator is sprayed evenly into the corresponding planting sub-area;
[0145] It can be understood that: by obtaining and calculating the plant height influencing characteristic data, and predicting the plant height at a future moment based on the plant height influencing characteristic data, and combining the initial average plant height of the plant, the growth amount of the plant height at different generation stages and generation environments can be determined; further, it is helpful to provide important data support for regulating the plant height;
[0146] Furthermore, on the basis of determining the average growth of the plants, abnormal planting sub-areas are screened out, and the dosage ratio of plant growth regulators is determined according to the dose influence characteristic data of the abnormal planting sub-areas, which is conducive to spraying different dosage ratios of plant growth regulators on different abnormal planting sub-areas, thereby facilitating the regulation of plant height. Further, it is conducive to avoiding lodging caused by plants being too tall, improving the ventilation and light transmittance of abnormal planting sub-areas, reducing the probability of occurrence of plant diseases and insect pests, and thus ensuring the yield and quality of economic crops.
[0147] Example 2
[0148] See also Figure 2 As shown, based on the same inventive concept, this embodiment discloses an automatic monitoring system suitable for cash crops. For details not provided in this embodiment, please refer to the relevant parts of the above embodiment 1. The system includes:
[0149] The data acquisition module 310 is used to obtain plant height influence characteristic data of plants in the planting sub-area within a predetermined time period [TN, T]; the plant height influence characteristic data includes plant state data, environmental state data and soil state data in the planting sub-area at each moment, and T and N are integers greater than zero;
[0150] The plant height prediction module 320 is used to retrieve the initial average plant height of the plants in the planting area, input the plant height influencing feature data into the pre-trained plant height prediction model, and predict the average plant height change of the plants at the future time T+N;
[0151] An increment calculation module 330 is used to calculate the average growth amount of the plants in the planting area based on the initial average plant height and the changed average plant height of the plants;
[0152] In a specific embodiment, the system further comprises:
[0153] The data retrieval module 410 is used to obtain the average growth of the plants in all the planting sub-areas and the growth stage of the plants in each planting sub-area at the future time T+N, where the number of all the planting sub-areas is Q, which is an integer greater than zero;
[0154] The abnormal screening module 420 is used to screen out a plurality of abnormal implantation sub-regions based on the average growth amount and the growth stage, and obtain the dose impact characteristic data of each abnormal implantation sub-region;
[0155] The ratio regression module 430 is used to input the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator;
[0156] The plant height control module 440 is used to control the spraying equipment to spray the plant growth regulator on the plants in the abnormal planting area according to the dosage ratio.
[0157] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters, weights and thresholds in the formula are set by technicians in this field according to actual conditions.
[0158] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired network or a wireless network. The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD) or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.
[0159] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.
[0160] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0161] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0162] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0163] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0164] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
[0165] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. An automated monitoring method for cash crops, characterized in that: The method comprises: S101: within a predetermined time period [TN, T], obtaining plant height influence characteristic data of plants in the planting sub-region; the plant height influence characteristic data includes plant state data, environment state data and soil state data in the planting sub-region at each moment, T and N are integers greater than zero, and T is greater than N; The plant status data include the average plant height, plant variety and plant density coefficient of the plants in the planting sub-area; The logic for obtaining the density coefficient of the plant is as follows: When the drone is controlled to arrive above the airspace of the planting sub-area, a bird's-eye view of the plants in the planting sub-area is obtained; Convert the top view from the standard RGB color space to the HSV color space, use the color threshold method to perform color segmentation on the converted top view, separate the plant area in the top view from the background according to the target color of the plant's growth stage, and define the mask G of the target color: Where: I ij is the pixel value at position (i, j) in the top view, 1 represents plant pixel and 0 represents background pixel; According to the mask G of the target color, the coverage area A of the plant area is calculated, and the coverage area A is normalized to obtain the coverage rate C of the plant area: Where: G ij is the mask value of the target color, M and N represent the width and height of the top view respectively; Get the actual coverage area and image resolution of the top view, and calculate the density coefficient of the plant based on the actual coverage area and image resolution: Where: D is the density coefficient of the plant, S is the actual coverage area, Δx and Δy are the resolutions of the top view; S102: Retrieving the initial average plant height of the plants in the planting area, inputting the plant height influencing feature data into a pre-trained plant height prediction model, and predicting the average plant height change of the plants at the future time T+N; S103: Calculate the average growth amount of the plants in the planting area based on the initial average plant height and the changed average plant height of the plants.
2. The automated monitoring method for cash crops according to claim 1, characterized in that: in, The environmental state data include environmental temperature, environmental humidity and rainfall; the soil state data include soil temperature, soil humidity and soil state coefficient.
3. The automated monitoring method for cash crops according to claim 2, characterized in that: The logic for obtaining the soil state coefficient is as follows: Obtain the actual soil porosity, actual soil nitrogen content and actual soil pH of the planting sub-area; The process of obtaining the actual porosity of the soil is as follows: Obtain soil samples from the planting sub-area, and insert four electrodes into the soil samples at equal intervals to a predetermined depth; Use the two outer electrodes to conduct a known current R into the soil and measure the voltage V through the two inner electrodes; Substitute the current R and voltage V into the pre-built porosity calculation formula to obtain the actual porosity of the soil; Wherein, the porosity calculation formula is as follows: Where: is the actual porosity of the soil, K is the geometric factor, α is the empirical constant, F is the formation factor, and β is the empirical constant; Substituting the actual soil porosity, actual soil nitrogen content and actual soil pH into the pre-built mathematical model to obtain the soil state coefficient; The mathematical model is expressed as follows: Where: SC is the soil state coefficient, μ is the actual nitrogen content of the soil, μ opt is the soil reference nitrogen content, is the actual porosity of the soil, is the soil reference porosity, pH is the actual soil acidity, pH opt is the reference pH of the soil, k1 and k2 are correction factors greater than zero, k1>k2.
4. The automated monitoring method for cash crops according to claim 3, characterized in that: The training logic of the plant height prediction model is as follows: Acquire historical plant height training data, and divide the historical plant height training data into a plant height prediction training set and a plant height prediction test set; the historical plant height training data includes plant height influencing feature data and its corresponding changed average plant height; Constructing a first regression network, taking the plant height influencing feature data in the plant height prediction training set as input data of the first regression network, and taking the changing average plant height in the plant height prediction training set as output data of the first regression network, training the first regression network, and obtaining an initial plant height prediction network; The plant height prediction test set is used to verify the model of the initial plant height prediction network, and the initial plant height prediction network with a preset first test error is output as the plant height prediction model.
5. The automated monitoring method for cash crops according to claim 4, characterized in that: The method further comprises: S201: Obtain the average growth amount of plants in all planting sub-areas, and the growth stage of plants in each planting sub-area at the future time T+N, where the number of all planting sub-areas is Q, which is an integer greater than zero; S202: Screening out a plurality of abnormally implanted sub-regions based on the average growth amount and the growth stage, and obtaining dose impact characteristic data of each abnormally implanted sub-region; S203: inputting the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator; S204: Control the spraying equipment according to the dosage ratio to spray the plants in the abnormal planting area with plant growth regulator.
6. The automated monitoring method for cash crops according to claim 5, characterized in that: The screening out of multiple abnormally implanted sub-areas includes: a1: Extract the average growth of the r-th planting sub-region, and extract the growth stage of the r-th planting sub-region at the future time T+N; a2: Based on the preset relationship between the growth stage and the growth amount interval threshold, the corresponding growth amount threshold is obtained, and the average growth amount of the r-th planting sub-area is compared with the corresponding growth amount threshold; if the average growth amount is greater than or equal to the corresponding growth amount threshold, the r-th planting sub-area is determined to be a non-abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1; if the average growth amount is less than the corresponding growth amount threshold, the r-th planting sub-area is determined to be an abnormal planting sub-area, and r=r+1 is set, and the process returns to step a1; a3: Repeat the above steps a1 to a2 until r=Q, then end the loop and obtain multiple abnormal planting sub-regions.
7. The automated monitoring method for cash crops according to claim 6, characterized in that: The dose impact characteristic data include the density coefficient, average growth amount, growth rate and growth stage of the plants in the abnormal planting area; The training process of the dose regression model is as follows: Acquire historical dose spraying training data, and divide the historical dose spraying training data into a dose regression training set and a dose regression test set; the historical dose spraying training data includes dose influence characteristic data and its corresponding dosage ratio of plant growth regulator; Constructing a second regression network, taking the dose effect feature data in the dose regression training set as input data of the second regression network, and taking the dosage ratio in the dose regression training set as output data of the second regression network, training the second regression network, and obtaining an initial dose regression network; The initial dose regression network is model verified using the dose regression test set, and the initial dose regression network with a value less than or equal to the preset second test error is output as the dose regression model.
8. An automated monitoring system for cash crops, characterized in that: include: The data acquisition module is used to obtain the plant height influence characteristic data of the plants in the planting sub-area within a predetermined time period [TN, T]; the plant height influence characteristic data includes the plant state data, environment state data and soil state data in the planting sub-area at each moment, and T and N are integers greater than zero; The plant status data include the average plant height, plant variety and plant density coefficient of the plants in the planting sub-area; The logic for obtaining the density coefficient of the plant is as follows: When the drone is controlled to arrive above the airspace of the planting sub-area, a bird's-eye view of the plants in the planting sub-area is obtained; Convert the top view from the standard RGB color space to the HSV color space, use the color threshold method to perform color segmentation on the converted top view, separate the plant area in the top view from the background according to the target color of the plant's growth stage, and define the mask G of the target color: Where: I ij is the pixel value at position (i, j) in the top view, 1 represents plant pixel and 0 represents background pixel; According to the mask G of the target color, the coverage area A of the plant area is calculated, and the coverage area A is normalized to obtain the coverage rate C of the plant area: Where: G ij is the mask value of the target color, M and N represent the width and height of the top view respectively; Get the actual coverage area and image resolution of the top view, and calculate the density coefficient of the plant based on the actual coverage area and image resolution: Where: D is the density coefficient of the plant, S is the actual coverage area, Δx and Δy are the resolutions of the top view; The plant height prediction module is used to retrieve the initial average plant height of the plants in the planting area, input the plant height influencing feature data into the pre-trained plant height prediction model, and predict the average plant height change of the plants at the future time T+N; The increment calculation module is used to calculate the average growth amount of plants in the planting area based on the initial average plant height and the changed average plant height of the plants.
9. The automatic monitoring system for cash crops according to claim 8, characterized in that: The system further comprises: The data retrieval module is used to obtain the average growth of plants in all planting sub-areas and the growth stage of plants in each planting sub-area at the future time T+N. The number of all planting sub-areas is Q, which is an integer greater than zero. An abnormal screening module is used to screen out multiple abnormal planting sub-areas based on average growth amount and growth stage, and obtain dose impact characteristic data of each abnormal planting sub-area; A ratio regression module is used to input the dosage effect characteristic data into a pre-trained dosage regression model to obtain the dosage ratio of the plant growth regulator; The plant height control module is used to control the spraying equipment to spray plant growth regulators on plants in the abnormal planting area according to the dosage ratio.
