An AI crop phenological period recognition system

Abstract

The present invention discloses an AI-powered crop phenological period recognition system, comprising a memory, a timing device, an image data acquisition device, an image data processing device, a crop phenological period analysis module, a crop yield prediction module, and an output module. The memory is used to store trained crop recognition models and crop yield prediction models. The crop recognition model inputs image data and outputs crop phenological periods. The crop yield prediction model inputs image data and outputs predicted yields. The image data processing device processes image data obtained by the image data acquisition device and extracts and outputs image data with significant differences between the two images before and after the output. The present invention addresses the technical issues of low phenological period recognition accuracy, insufficient precision in phenological period recognition, and the inability to accurately predict yields.

Description

Technical Field

[0001] The present invention relates to the technical field of crop phenological period recognition, and more specifically, to an AI crop phenological period recognition system. Background Art

[0002] The phenological phase of a crop refers to the morphological characteristics of its growth and development under certain external conditions. These specific standards are artificially established to scientifically understand the crop's growth process. Examples include: rice seedling emergence; incomplete sheath emergence; and leaf color turning green.

[0003] Identifying the phenological periods of crops can guide farmers to accurately plant crops, prevent pests, avoid risks, rationally allocate production resources, and accurately predict the time to market, thereby improving crop yield and quality.

[0004] Currently, existing technologies mostly extract features from images of crop phenological periods, generate a feature library, and establish a mathematical model based on the characteristics of each phenological period. This approach automatically identifies phenological periods by extracting features from crop images to be identified and analyzing them based on the mathematical model. However, crop growth is continuous and gradual, with a gradual transition between two consecutive phenological periods. Simply acquiring images of each phenological period and establishing a mathematical model cannot accurately divide phenological periods, resulting in low phenological period identification accuracy. Phenological period divisions are relatively crude, making them difficult to refine. Crop phenological periods are often used to predict yield, but current crop phenological period identification technology cannot accurately predict yield. Summary of the Invention

[0005] The purpose of the present invention is to provide an AI crop phenological period recognition system to solve the technical problems of low accuracy of phenological period recognition, insufficient precision of phenological period recognition, and inability to accurately predict yield.

[0006] To achieve the above-mentioned purpose, an AI crop phenological period recognition system is provided, which includes a memory, a timing device, an image data acquisition device, an image data processing device, a crop phenological period analysis module, a crop yield prediction module, and an output module;

[0007] The memory is used to store trained crop recognition models and crop yield prediction models; the crop recognition model inputs image data and outputs crop phenological periods; the crop yield prediction model inputs image data and outputs predicted yields;

[0008] The timing device is used to set time intervals and time;

[0009] The image data acquisition device is used to acquire image data of crops according to time intervals;

[0010] The image data processing device is used to process the image data obtained by the image data acquisition device, and extract and output image data with a large difference between the two image data before and after;

[0011] The crop phenological period analysis module is used to receive image data output by the image data processing device, call the crop recognition model through the memory, obtain the crop phenological period and annotate it to the image data; the crop phenological period analysis module is also used to train the crop recognition model;

[0012] The crop yield prediction module is used to receive the image data obtained by the crop phenological period analysis module, call the crop yield prediction model through the memory, obtain the predicted yield and mark it on the image data;

[0013] The output module is used to output the image data output by the phenological period prediction module.

[0014] Particularly, the time interval is 1 day.

[0015] Particularly, the image data acquisition device adopts a crop image high-definition acquisition instrument.

[0016] Particularly, the output module is a printer.

[0017] An AI crop phenological period recognition method includes the following steps:

[0018] Step S1. Establishing a crop recognition model; the crop recognition model inputs image data and outputs crop phenological periods;

[0019] Step S2. Establishing a crop yield prediction model; the crop yield prediction model inputs image data and outputs predicted yield;

[0020] Step S3. Set the time interval;

[0021] Step S4. collecting image data of the crop according to time intervals;

[0022] Step S5. Process the image data of step S4, extract and output image data with a large difference between the two images before and after;

[0023] Step S6. Based on the image data output in step S5 and the crop recognition model, obtain the crop phenological period and annotate the image data;

[0024] Step S7. Based on the image data obtained in step S6 and the crop yield prediction model, the predicted yield is obtained and annotated to the image data.

[0025] In particular, the specific method of establishing a crop recognition model is:

[0026] Collect a certain number of crop image data of a certain type and quantity and mark the crop phenological periods to establish a target field dataset;

[0027] The target domain dataset is divided into a crop phenological period training set, a crop phenological period validation set, and a crop phenological period test set according to a certain ratio;

[0028] The crop phenological period training set and the crop phenological period verification set are input into the neural network model for training; the network model can input image data and output crop phenological period;

[0029] The trained neural network model is then input into the crop phenological period test set for testing until a network model with a preset accuracy threshold is trained as a crop recognition model.

[0030] In particular, the specific method for establishing a crop yield prediction model is:

[0031] Collect a certain number of crop image data of a certain type and quantity, as well as marked crop phenological periods, and predict yields to establish a joint feature vector database of crop growth trends and yields;

[0032] The joint feature vector database is divided into a yield prediction training set, a yield prediction validation set and a yield prediction test set according to a certain ratio;

[0033] Input the yield prediction training set and the yield prediction verification set into the neural network model for training; the network model can input image data and output predicted yield;

[0034] The trained neural network model is then input into the yield prediction test set for testing until a network model with an accuracy rate reaching a preset threshold is trained as the crop yield prediction model.

[0035] In particular, the processing of the image data in step S4 includes the following steps:

[0036] Adjust the contrast, brightness, and size of image data and extract image features of crops; remove image features outside of crops.

[0037] In particular, in step S5, the method of outputting the image data with a large difference between the two images is as follows:

[0038] Step A1. reducing the image data to a specified size;

[0039] Step A2. grayscale processing is performed on the image data to obtain the grayscale of each pixel of the image data;

[0040] Step A3. Calculate the grayscale average of the image data;

[0041] Step A4. Arrange the pixels of the image data in order, compare the grayscale of each pixel with the average value, and record the grayscale greater than or equal to the average value as 1, and less than the average value as 0, to obtain an image sequence;

[0042] Step A5. Setting an image difference threshold;

[0043] Step A6. Obtain an image sequence from a designated image data according to steps A1 to A4; set integer value n = 1;

[0044] Step A7. Obtain an image sequence by performing the image data of the last n time intervals according to steps A1-A4;

[0045] Step A8. Compare the different bits of the two image sequences;

[0046] Step A9. If the number of bits exceeds the image difference threshold, the image data corresponding to the two image sequences are extracted and output; if the number of bits does not exceed the image difference threshold, set n = n + 1, and execute steps A7-A9 in sequence until the number of bits exceeds the image difference threshold, then the image data corresponding to the two image sequences are extracted and output.

[0047] The technical principles and beneficial effects of the present invention are as follows:

[0048] The present invention can remove similar image data by outputting image data with large differences between the two images before and after. The extracted image data can identify the crop phenological period through the crop phenological period analysis module and the crop yield prediction module, and predict the crop yield based on the crop phenological period and image data, thereby solving the technical problems of low accuracy in phenological period identification, insufficient precision in phenological period identification, and inability to accurately predict yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 2 is an overall structural diagram of a system according to an embodiment of the present invention.

[0051] Figure 2 Flowchart of a method according to an embodiment of the present invention. Implementation Method

[0052] The preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more precise definition of the protection scope of the present invention.

[0053] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not require further definition or explanation in subsequent drawings.

[0054] It should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer" and the like indicate positions or locations based on the positions shown in the accompanying drawings, or the positions or locations in which the inventive product is typically placed when in use. These terms are intended solely to facilitate and simplify the description of the present invention and are not intended to indicate or imply that the devices or components referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first," "second," and "third," etc., are used solely to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0055] Furthermore, terms such as "horizontal," "vertical," and "overhanging" do not necessarily imply that a component must be absolutely horizontal or overhanging, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but rather that it can be slightly tilted.

[0056] like Figure 1 , an AI crop phenological period recognition system according to an embodiment of the present invention includes a memory, a timing device, an image data acquisition device, an image data processing device, a crop phenological period analysis module, a crop yield prediction module, and an output module;

[0057] The memory is used to store the trained crop recognition model and the crop yield prediction model; the crop recognition model inputs image data and outputs the crop phenological period; the crop yield prediction model inputs image data and outputs the predicted yield;

[0058] The timing device is used to set time intervals and measure time;

[0059] The image data acquisition device is used to acquire image data of crops according to time intervals;

[0060] The image data processing device is used to process the image data obtained by the image data acquisition device, and extract the image data with a large difference between the two image data before and after output;

[0061] The crop phenological period analysis module is used to receive image data output by the image data processing device, call the crop recognition model through the memory, obtain the crop phenological period and annotate it to the image data; the crop phenological period analysis module is also used to train the crop recognition model;

[0062] The crop yield prediction module is used to receive the image data obtained by the crop phenological period analysis module, call the crop yield prediction model through the memory, obtain the predicted yield and mark it on the image data;

[0063] The output module is used to output the image data output by the phenological period prediction module.

[0064] The time interval is 1 day.

[0065] The image data acquisition device adopts a crop image high-definition acquisition instrument.

[0066] The output module uses a printer.

[0067] like Figure 2 As shown, an AI crop phenological period recognition method includes the following steps:

[0068] Step S1. Establishing a crop recognition model; the crop recognition model inputs image data and outputs crop phenological periods;

[0069] Step S2. Establishing a crop yield prediction model; the crop yield prediction model inputs image data and outputs predicted yield;

[0070] Step S3. Set the time interval;

[0071] Step S4. collecting image data of the crop according to time intervals;

[0072] Step S5. Process the image data of step S4, extract and output image data with a large difference between the two images before and after;

[0073] Step S6. Based on the image data output in step S5 and the crop recognition model, obtain the crop phenological period and annotate the image data;

[0074] Step S7. Based on the image data obtained in step S6 and the crop yield prediction model, the predicted yield is obtained and annotated to the image data.

[0075] The specific method of establishing a crop recognition model is as follows:

[0076] Collect a certain number of crop image data of a certain type and quantity and mark the crop phenological periods to establish a target field dataset;

[0077] The target domain dataset is divided into a crop phenological period training set, a crop phenological period validation set, and a crop phenological period test set according to a certain ratio;

[0078] The crop phenological period training set and the crop phenological period verification set are input into the neural network model for training; the network model can input image data and output crop phenological period;

[0079] The trained neural network model is then input into the crop phenological period test set for testing until a network model with a preset accuracy threshold is trained as a crop recognition model.

[0080] The specific method of establishing a crop yield prediction model is as follows:

[0081] Collect a certain number of crop image data of a certain type and quantity, as well as marked crop phenological periods, and predict yields to establish a joint feature vector database of crop growth trends and yields;

[0082] The joint feature vector database is divided into a yield prediction training set, a yield prediction validation set and a yield prediction test set according to a certain ratio;

[0083] Input the yield prediction training set and the yield prediction verification set into the neural network model for training; the network model can input image data and output predicted yield;

[0084] The trained neural network model is then input into the yield prediction test set for testing until a network model with an accuracy rate reaching a preset threshold is trained as the crop yield prediction model.

[0085] Processing the image data of step S4 includes the following steps:

[0086] Adjust the contrast, brightness, and size of image data and extract image features of crops; remove image features outside of crops.

[0087] In step S5, the method for outputting the image data with a large difference between the two images is as follows:

[0088] Step A1. reducing the image data to a specified size;

[0089] Step A2. grayscale processing is performed on the image data to obtain the grayscale of each pixel of the image data;

[0090] Step A3. Calculate the grayscale average of the image data;

[0091] Step A4. Arrange the pixels of the image data in order, compare the grayscale of each pixel with the average value, and record the grayscale greater than or equal to the average value as 1, and less than the average value as 0, to obtain an image sequence;

[0092] Step A5. Setting an image difference threshold;

[0093] Step A6. Obtain an image sequence from a designated image data according to steps A1 to A4; set integer value n = 1;

[0094] Step A7. Obtain an image sequence by performing the image data of the last n time intervals according to steps A1-A4;

[0095] Step A8. Compare the different bits of the two image sequences;

[0096] Step A9. If the number of bits exceeds the image difference threshold, the image data corresponding to the two image sequences are extracted and output; if the number of bits does not exceed the image difference threshold, set n = n + 1, and execute steps A7-A9 in sequence until the number of bits exceeds the image difference threshold, then the image data corresponding to the two image sequences are extracted and output.

[0097] Although the embodiments of the present invention are described in conjunction with the accompanying drawings, the patent owner may make various changes or modifications within the scope of the appended claims. As long as they do not exceed the scope of protection described in the claims of the present invention, they should be within the scope of protection of the present invention.

Claims

1. An AI crop phenological period recognition system, characterized by: It includes a memory, a timing device, an image data acquisition device, an image data processing device, a crop phenological period analysis module, a crop yield prediction module, and an output module; The memory is used to store trained crop recognition models and crop yield prediction models; the crop recognition model inputs image data and outputs crop phenological periods; the crop yield prediction model inputs image data and outputs predicted yields; The timing device is used to set time intervals and time; The image data acquisition device is used to acquire image data of crops according to time intervals; The image data processing device is used to process the image data obtained by the image data acquisition device, and extract and output image data with a large difference between the two image data before and after; The method for outputting image data having a large difference between the two images is specifically as follows: Step A1. reducing the image data to a specified size; Step A2. grayscale processing is performed on the image data to obtain the grayscale of each pixel of the image data; Step A3. Calculate the grayscale average of the image data; Step A4. Arrange the pixels of the image data in order, compare the grayscale of each pixel with the average value, and record the grayscale greater than or equal to the average value as 1, and less than the average value as 0, to obtain an image sequence; Step A5. Setting an image difference threshold; Step A6. Obtain an image sequence from a designated image data according to steps A1 to A4; set integer value n = 1; Step A7. Obtain an image sequence by performing the image data of the last n time intervals according to steps A1-A4; Step A8. Compare the different bits of the two image sequences; Step A9. If the number of bits exceeds the image difference threshold, extract and output the image data corresponding to the two image sequences. If the number of bits does not exceed the image difference threshold, set n = n + 1, and execute steps A7 through A9 sequentially until the number of bits exceeds the image difference threshold, then extract and output the image data corresponding to the two image sequences. The crop phenological period analysis module is used to receive image data output by the image data processing device, call the crop recognition model through the memory, obtain the crop phenological period and annotate it to the image data; the crop phenological period analysis module is also used to train the crop recognition model; The crop yield prediction module is used to receive the image data obtained by the crop phenological period analysis module, call the crop yield prediction model through the memory, obtain the predicted yield and mark it on the image data; The output module is used to output the image data output by the crop yield prediction module.

2. The AI ​​crop phenological period recognition system according to claim 1, characterized in that: The time interval is 1 day.

3. The AI ​​crop phenological period recognition system according to claim 1, characterized in that: The image data acquisition device adopts a crop image high-definition acquisition instrument.

4. The AI ​​crop phenological period recognition system according to claim 1, characterized in that: The output module adopts a printer.

5. An AI-based crop phenological phase identification method comprising the following steps: Step S1. Establishing a crop recognition model; the crop recognition model inputs image data and outputs crop phenological periods; Step S2. Establishing a crop yield prediction model; the crop yield prediction model inputs image data and outputs predicted yield; Step S3. Set the time interval; Step S4. collecting image data of the crop according to time intervals; Step S5. Process the image data of step S4, extract and output image data with a large difference between the two images before and after; Step S6. Based on the image data output in step S5 and the crop recognition model, obtain the crop phenological period and annotate the image data; Step S7. Based on the image data obtained in step S6 and the crop yield prediction model, the predicted yield is obtained and annotated to the image data.

6. The AI ​​crop phenological period recognition method according to claim 5, characterized in that: The specific method of establishing a crop recognition model is as follows: Collect a certain number of crop image data of a certain type and quantity and mark the crop phenological periods to establish a target field dataset; The target domain dataset is divided into a crop phenological period training set, a crop phenological period validation set, and a crop phenological period test set according to a certain ratio; The crop phenological period training set and the crop phenological period verification set are input into the neural network model for training; the network model can input image data and output crop phenological period; The trained neural network model is then input into the crop phenological period test set for testing until a network model with a preset accuracy threshold is trained as a crop recognition model.

7. The AI ​​crop phenological period recognition method according to claim 5, characterized in that: The specific method of establishing a crop yield prediction model is as follows: Collect a certain number of crop image data of a certain type and quantity, as well as marked crop phenological periods, and predict yields to establish a joint feature vector database of crop growth trends and yields; The joint feature vector database is divided into a yield prediction training set, a yield prediction validation set and a yield prediction test set according to a certain ratio; Input the yield prediction training set and the yield prediction verification set into the neural network model for training; the network model can input image data and output predicted yield; The trained neural network model is then input into the yield prediction test set for testing until a network model with an accuracy rate reaching a preset threshold is trained as the crop yield prediction model.

8. The AI ​​crop phenological period recognition method according to claim 5, characterized in that: The image data processing step S4 includes the following steps: Adjust the contrast, brightness, and size of image data and extract image features of crops; remove image features outside of crops.

9. The AI ​​crop phenological phase recognition method according to claim 5, wherein in step S5, the method of outputting image data with a large difference between the two images is specifically: Step A1. reducing the image data to a specified size; Step A2. grayscale processing is performed on the image data to obtain the grayscale of each pixel of the image data; Step A3. Calculate the grayscale average of the image data; Step A4. Arrange the pixels of the image data in order, compare the grayscale of each pixel with the average value, and record the grayscale greater than or equal to the average value as 1, and less than the average value as 0, to obtain an image sequence; Step A5. Setting an image difference threshold; Step A6. Obtain an image sequence from a designated image data according to steps A1 to A4; set integer value n = 1; Step A7. Obtain an image sequence by performing the image data of the last n time intervals according to steps A1-A4; Step A8. Compare the different bits of the two image sequences; Step A9. If the number of bits exceeds the image difference threshold, the image data corresponding to the two image sequences are extracted and output; if the number of bits does not exceed the image difference threshold, set n = n + 1, and execute steps A7-A9 in sequence until the number of bits exceeds the image difference threshold, then the image data corresponding to the two image sequences are extracted and output.

Images