Method and system for identifying track of agricultural machine

Abstract

The invention provides an agricultural machine track identification method and system, and the method comprises the steps: obtaining the feature information of an agricultural machine track; wherein the feature information at least comprises an instantaneous speed feature of a target track point in the agricultural machine track, a time difference feature of the target track point and a previous track point of the target track point, a direction feature of the agricultural machine track, a direction variance feature of a plurality of track points in the agricultural machine track, and a speed variance feature of a plurality of track points in the agricultural machine track; speed mean value features of a plurality of track points in the agricultural machine track; and inputting the feature information into a pre-trained recognition model to output an agricultural machine track type. According to the method and the device, the recognition precision of the recognition model is improved through the multi-dimensional feature information, and particularly after the speed variance, the direction variance and the speed variance of a plurality of track points in the agricultural machine track are added to serve as the input features of the recognition model, the burst point recognition problem on the agricultural machine track can be optimized.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent agricultural machinery technology, and in particular to a method and system for identifying agricultural machinery trajectories. Background Technology

[0002] With the development of technology, agricultural mechanization is becoming increasingly intelligent. Among these, the identification of agricultural machinery operation trajectories plays a crucial role in statistical analysis of operation area, research on crop distribution, and planning operation routes.

[0003] In existing technologies, it is impossible to accurately identify whether agricultural machinery tracks are on roads or in farmland, which is a technical problem that urgently needs to be solved. Summary of the Invention

[0004] The technical problem to be solved by this disclosure is to overcome the shortcomings of existing technologies in accurately identifying agricultural machinery trajectory analogies, and to provide a method and system for identifying agricultural machinery trajectories.

[0005] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0006] This disclosure provides a method for identifying the trajectory of agricultural machinery, the method comprising:

[0007] Obtain feature information of agricultural machinery trajectory; wherein, the feature information includes at least: instantaneous velocity features of target trajectory points in the agricultural machinery trajectory, time difference features between the target trajectory point and the previous trajectory point of the target trajectory point, direction features of the agricultural machinery trajectory, direction variance features of several trajectory points in the agricultural machinery trajectory, velocity variance features of several trajectory points in the agricultural machinery trajectory, and velocity mean features of several trajectory points in the agricultural machinery trajectory;

[0008] The feature information is input into a pre-trained recognition model to output the agricultural machinery trajectory type; wherein, the agricultural machinery trajectory type includes road trajectory type and farmland trajectory type.

[0009] Preferably, the pre-trained recognition model includes a feature co-optimization module; the processing steps of the pre-trained recognition model for the input feature information include:

[0010] The feature information is processed by the feature collaborative optimization module, which includes at least one of the channel attention module, spatial attention module, and multi-scale fusion module.

[0011] Preferably, the method for identifying the agricultural machinery trajectory further includes:

[0012] A three-channel trajectory map is determined based on the aforementioned feature information; wherein, the three-channel trajectory map includes a first three-channel trajectory map and a second three-channel trajectory map; the first three-channel trajectory map is determined based on the instantaneous velocity characteristics of the target trajectory point in the agricultural machinery trajectory, the time difference characteristics between the target trajectory point and the previous trajectory point, and the directional characteristics of the agricultural machinery trajectory; the second three-channel trajectory map is determined based on the directional variance characteristics of several trajectory points in the agricultural machinery trajectory, the velocity variance characteristics of several trajectory points in the agricultural machinery trajectory, and the velocity mean characteristics of several trajectory points in the agricultural machinery trajectory.

[0013] The agricultural machinery trajectory type is visualized based on the three-channel trajectory diagram.

[0014] Preferably, the recognition model includes a post-processing layer, wherein the post-processing layer is used to distinguish the background of the three-channel trajectory map from the agricultural machinery trajectory type.

[0015] Preferably, the three-channel trajectory map is a satellite image at level 18.

[0016] Preferably, the step of determining the three-channel trajectory map based on the feature information includes:

[0017] The original image is determined based on the aforementioned feature information;

[0018] By using an overlay technique, 128 pixels of edge information are added to the image edges of the original image to determine the final three-channel trajectory map.

[0019] This disclosure provides a system for identifying the trajectory of agricultural machinery, the system comprising:

[0020] An acquisition module is used to acquire feature information of agricultural machinery trajectory; wherein, the feature information includes at least: instantaneous velocity features of target trajectory points in the agricultural machinery trajectory, time difference features between the target trajectory point and the previous trajectory point of the target trajectory point, direction features of the agricultural machinery trajectory, direction variance features of several trajectory points in the agricultural machinery trajectory, velocity variance features of several trajectory points in the agricultural machinery trajectory, and velocity mean features of several trajectory points in the agricultural machinery trajectory;

[0021] The recognition module is used to input the feature information into a pre-trained recognition model to output the agricultural machinery trajectory type; wherein, the agricultural machinery trajectory type includes road trajectory type and farmland trajectory type.

[0022] This disclosure provides an electronic device, including a memory, a processor, and a computer program stored in the memory and used to run on the processor, wherein the processor executes the computer program to implement the above-described method for identifying agricultural machinery trajectories.

[0023] This disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for identifying agricultural machinery trajectories.

[0024] This disclosure provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for recognizing agricultural machinery trajectories.

[0025] Based on common knowledge in the field, the above-mentioned preferred conditions can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0026] The positive and progressive effects of this disclosure are as follows:

[0027] This disclosure improves the recognition accuracy of the recognition model by using multi-dimensional feature information. In particular, by adding the variances of velocity, direction, and speed of several trajectory points in the agricultural machinery trajectory as input features of the recognition model, the problem of identifying explosion points on the agricultural machinery trajectory can be optimized. Attached Figure Description

[0028] Figure 1 A trajectory diagram is provided as an exemplary embodiment of this disclosure;

[0029] Figure 2 A schematic diagram of a trajectory type provided for an exemplary embodiment of this disclosure;

[0030] Figure 3 A flowchart illustrating a method for identifying agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0031] Figure 4 A first trajectory point distribution map of an example of a method for identifying agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0032] Figure 5 A second trajectory point distribution map illustrating an example of a method for identifying agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0033] Figure 6 A schematic diagram of a first trajectory line, illustrating an example of a method for recognizing agricultural machinery trajectories provided in an exemplary embodiment of this disclosure;

[0034] Figure 7 A schematic diagram of a second trajectory line, illustrating an example of a method for identifying agricultural machinery trajectories provided in an exemplary embodiment of this disclosure;

[0035] Figure 8 A first three-channel trajectory diagram illustrating an example of a method for recognizing agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0036] Figure 9A second three-channel trajectory diagram illustrating an example of a method for recognizing agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0037] Figure 10 A schematic diagram of the framework structure of an example recognition model for a method of recognizing agricultural machinery trajectories provided in an exemplary embodiment of this disclosure;

[0038] Figure 11 An example application effect diagram of a method for recognizing agricultural machinery trajectories provided as an exemplary embodiment of this disclosure;

[0039] Figure 12 Training convergence graph of the recognition model of an example of a method for recognizing agricultural machinery trajectories provided in an exemplary embodiment of this disclosure;

[0040] Figure 13 A schematic diagram of the structure of an agricultural machinery trajectory recognition system provided as an exemplary embodiment of this disclosure;

[0041] Figure 14 This is a schematic diagram of the structure of an electronic device provided as an exemplary embodiment of the present disclosure. Detailed Implementation

[0042] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0043] The prefixes such as "first" and "second" used in this disclosure are merely for distinguishing different descriptive objects and do not limit the position, order, priority, quantity, or content of the described objects. The use of ordinal numbers and other prefixes used to distinguish descriptive objects in this disclosure does not constitute a limitation on the described objects. The description of the described objects is given in the claims or the context of the embodiments, and should not be construed as an unnecessary limitation. Furthermore, in the description of this embodiment, unless otherwise stated, "multiple" means two or more.

[0044] In this embodiment of the disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information comply with relevant laws and regulations and do not violate public order and good morals.

[0045] Example 1

[0046] Farmers drive their agricultural machinery from home to the fields every day to work and return home afterwards. The GPS (Global Positioning System) on the machinery uploads a track point (e.g., ...) at regular intervals. Figure 1As shown in the image, these points contain information such as location, speed, time, and distance. We need to identify the collected points to determine whether each point is on the road or in the field. Finally, we connect the trajectory points of the day in chronological order to create a trajectory map (e.g., ...). Figure 2 As shown in the image, mark which parts are roads (red) and which parts are farmland (green) using different colors to distinguish them.

[0047] This task is of great significance in smart agriculture, as it can be used to count the area of ​​operation, study crop distribution, and plan operation routes. Therefore, this embodiment has developed a set of agricultural machinery operation trajectory recognition algorithms based on artificial intelligence image recognition. As long as the trajectory of the agricultural machinery for a day is input, it can accurately determine whether each point is a field trajectory or a road trajectory, and draw an agricultural machinery operation trajectory map.

[0048] The innovations of this embodiment include: 1. The use of six relatively independent and precise feature information increases the effectiveness and richness of the input features of the recognition model. 2. Addressing the "explosion point" problem, which refers to a data anomaly that occurs when analyzing agricultural machinery operation trajectory data in the fields of agricultural remote sensing, precision agriculture, or agricultural machinery navigation. Explosion points can also be called "anomalies" or "drift points." Explosion point data is often clumped together and disordered, with fluctuating positions. The current trajectory point alone cannot adequately determine the state. Therefore, this embodiment adds range feature information, using the dynamic range of several trajectory points before and after the target trajectory point as feature information, namely the variance of velocity, the variance of direction, and the mean of velocity, as input features of the recognition model. This allows the recognition model to learn rich range features to optimize the handling of the explosion point problem. Third, the structure of the recognition model is optimized. Channel Attention, Spatial Attention, and multi-scale fusion are used in the decoding layer of the recognition model to extract key information and multi-scale contextual information more accurately, thereby improving the recognition model's discrimination ability.

[0049] Figure 3 A flowchart illustrating a method for recognizing agricultural machinery trajectories provided as an exemplary embodiment of this disclosure is shown below. Figure 3 As shown, the methods for identifying agricultural machinery trajectories include:

[0050] S1. Obtain the feature information of the agricultural machinery trajectory.

[0051] The feature information includes at least: the instantaneous velocity features of the target trajectory point in the agricultural machinery trajectory, the time difference features between the target trajectory point and the previous trajectory point, the directional features of the agricultural machinery trajectory, the directional variance features of several trajectory points in the agricultural machinery trajectory, the velocity variance features of several trajectory points in the agricultural machinery trajectory, and the velocity mean features of several trajectory points in the agricultural machinery trajectory.

[0052] It should be noted that the direction of the agricultural machinery trajectory in this embodiment is represented by the angle between the line connecting the target trajectory point and the previous trajectory point and the horizontal or vertical coordinate axis.

[0053] S2. Input the feature information into the pre-trained recognition model to output the agricultural machinery trajectory type.

[0054] Among them, agricultural machinery trajectory types include road trajectory types and farmland trajectory types.

[0055] This implementation improves the recognition accuracy of the recognition model by using multi-dimensional feature information. In particular, by adding the variances of velocity, direction, and speed of several trajectory points in the agricultural machinery trajectory as input features of the recognition model, the problem of identifying explosion points on the agricultural machinery trajectory can be optimized.

[0056] In an optional implementation, the pre-trained recognition model includes a feature co-optimization module. The pre-trained recognition model processes the input feature information through the following steps:

[0057] The feature collaborative optimization module processes the feature information. The feature collaborative optimization module includes at least one channel attention module, spatial attention module, and multi-scale fusion module.

[0058] The Channel Attention module focuses on the channel dimension of the feature map, with the goal of enabling the recognition model to automatically learn and weigh the importance of different feature channels.

[0059] Spatial Attention focuses on the spatial dimensions of the feature map (i.e., the height and width of the feature map). Its goal is to enable the recognition model to automatically focus on the region with the most information in the feature map and ignore irrelevant background or noisy regions.

[0060] The multi-scale fusion module focuses on how to effectively combine feature maps of different depths (different scales) in the recognition model. Its goal is to extract multi-scale contextual information from the feature maps and improve recognition accuracy.

[0061] In this embodiment, by using the feature co-optimization module to decode the feature map, it is possible to extract key information and multi-scale contextual information more accurately, thereby improving the recognition performance of the recognition model.

[0062] In an optional implementation, the method for identifying agricultural machinery trajectories further includes:

[0063] S3. Determine the three-channel trajectory map based on feature information.

[0064] The three-channel trajectory map includes a first three-channel trajectory map and a second three-channel trajectory map. The first three-channel trajectory map is determined based on the instantaneous velocity characteristics of the target trajectory point in the agricultural machinery trajectory, the time difference characteristics between the target trajectory point and the previous trajectory point, and the directional characteristics of the agricultural machinery trajectory. The second three-channel trajectory map is determined based on the directional variance characteristics, the velocity variance characteristics, and the velocity mean characteristics of several trajectory points in the agricultural machinery trajectory.

[0065] S4. Visualize agricultural machinery trajectory types based on three-channel trajectory graph.

[0066] A typical color image is composed of three channels—red, green, and blue (RGB). In an RGB image, the value corresponding to each channel represents the brightness value under red, green, and blue light. By combining these channels appropriately, a color image can be generated, where the intensity and saturation of each color can be adjusted by distorting the channel values. Semantic segmentation can assign each pixel in an RGB image to a specific category.

[0067] In this embodiment, the first three-channel trajectory map uses custom data to define the three channels of the trajectory map as the instantaneous velocity feature of the target trajectory point, the time difference feature between the target trajectory point and the previous trajectory point, and the directional feature of the agricultural machinery trajectory. That is, the first three-channel trajectory map simultaneously includes information on velocity, time difference, and direction, composed of three channels—velocity, direction, and time difference (SDT). It should be noted that the velocity information value in the background area is zero. It should also be noted that because the instantaneous velocity obtained through GPS is inaccurate, and many instances of zero velocity during agricultural machinery operation cause significant interference to the data, there is a correlation and redundancy between time and mileage. Therefore, the ratio of mileage to time difference between the target trajectory point and the previous trajectory point is used as the instantaneous velocity feature of the target trajectory point to improve the accuracy of the instantaneous velocity of the target trajectory point. The feature information corresponding to the first three-channel trajectory map increases the effectiveness and richness of the features.

[0068] Similarly, in this embodiment, the second and third channel trajectory map uses custom data to define the three channels of the trajectory map as the directional variance feature, the velocity variance feature, and the velocity mean feature of several trajectory points in the agricultural machinery trajectory. That is, the second and third channel trajectory map simultaneously includes information on directional variance, velocity variance, and velocity mean, and is composed of three channels—directional variance, velocity variance, and velocity mean (SSD RANGE). In this embodiment, typically three trajectory points before and after the target trajectory point, for a total of seven trajectory points, are used to calculate the directional variance, velocity variance, and velocity mean.

[0069] In this embodiment, the trajectory of the agricultural machinery can be clearly seen on the map through the first three-channel trajectory map and the second three-channel trajectory map. In particular, the variance color of the direction from the second three-channel trajectory map to the explosion point is reddish, which greatly improves the accuracy of explosion point identification.

[0070] In an alternative implementation, the recognition model includes a post-processing layer, which is used to distinguish the background and agricultural machinery trajectory type of the three-channel trajectory map.

[0071] The post-processing layer determines the background through matrix operations on the first vector and determines the type of different trajectory segments through matrix operations on the second vector. Both the first and second vectors are binary vectors of length 3. Specifically, the post-processing layer is a Softmax layer.

[0072] For example, normalization is performed using the Softmax function, where the model outputs three classification probability values ​​[p1, p2, p3] at each point after Softmax, representing the probabilities of [background, road trajectory, farmland trajectory]. Specifically, the first vector is [1, 0, 0]. Multiplying [1, 0, 0] by [p1, p2, p3] yields [p1, 0, 0], representing the classification result of the background region. In this case, since the probabilities of road trajectory and farmland trajectory are both set to zero, the classification result for these regions can only be background. The second vector is [0, 1, 1]. Multiplying [0, 1, 1] by [p1, p2, p3] yields [0, p2, p3]. Here, since the probability of background is also set to zero, the classification result for the trajectory can only be either road trajectory type or farmland trajectory type.

[0073] In one alternative implementation, the three-channel trajectory map is a satellite image at level 18.

[0074] In this embodiment, the first three-channel trajectory map determined based on multiple trajectory points can be selected with different sizes or map levels. For example, to avoid the loss of details due to a map level that is too low or the inability to obtain more global information due to a map level that is too high, a level 18 satellite image can be selected; to avoid the image being too large and reducing the running speed of the model, the image size can be selected as 1024×1024.

[0075] In an optional implementation, step S3 includes:

[0076] S31. Determine the original image based on feature information.

[0077] S32. By using an overlay technique, 128 pixels of edge information are added to the image edges of the original image to determine the final three-channel trajectory map.

[0078] In this embodiment, to ensure that the model can extract sufficiently rich edge information, or to meet the size requirements of the model input, an overlap technique is used to add edge information and expand the original image. For example, adding 128 pixels of edge information to the edges of an image with a size of 1024×1024 expands the image to 1280×1280.

[0079] The following is a specific example that explains in detail the method for recognizing agricultural machinery trajectories.

[0080] I. Data Acquisition

[0081] The GPS devices equipped on the agricultural machinery continuously collect data and upload it to a server via a wireless network. This data includes the latitude and longitude coordinates, instantaneous speed, and time of each trajectory point. After obtaining the data, we can plot... Figure 4 and Figure 5 The distribution map of trajectory points.

[0082] For each trajectory point distribution map, an appropriate map level needs to be selected (similar to satellite maps). If the level is too low, the trajectory may be too fine, resulting in the loss of many details. If the level is too high, the image will be trapped in local details, failing to capture more global information. After extensive experimentation, level 18 satellite imagery has been found to be the most suitable level.

[0083] Because the trajectories are long and the map size is limited, a single trajectory often requires multiple images, sometimes even thousands. Theoretically, larger input images can extract better global information. However, if the images are too large, the model's running speed will be very slow, and it will also contain more invalid information (pure black backgrounds without trajectories), resulting in a waste of computational resources. After experimentation, a size of 1024×1024 is found to be more suitable.

[0084] Then, by overlapping, 128 pixels of information are added to the edges to ensure that the model has sufficient edge information. The image is then expanded to 1280×1280 and input into the model for calculation.

[0085] Ultimately, as Figure 6 and Figure 7 As shown, a trajectory is divided into many 1280×1280 images, and the points are connected to form a trajectory line. These lines are then successively input into the recognition model to predict the category.

[0086] The trajectory points are plotted into a visual driving trajectory map according to the time sequence. For example: Figure 8 As shown, based on S (the instantaneous velocity characteristic of the target trajectory point, i.e., the ratio of mileage to time difference between the target trajectory point and the previous trajectory point), D (the time difference characteristic between the target trajectory point and the previous trajectory point), and T (the directional characteristic of the agricultural machinery trajectory, i.e., the angle between the line connecting the target trajectory point and the previous trajectory point and the horizontal or vertical coordinate axis), the RGB channel values ​​are determined, and the color is represented by the corresponding numerical value, distinguishing between road trajectory types and farmland trajectory types. For example: Figure 9 As shown, the RGB values ​​are composed of the variance of the direction, the variance of the velocity, and the mean of the velocity of the seven trajectory points, which are the three trajectory points before and after the target trajectory point. It can be clearly seen that the variance of the direction at the explosion point is particularly large, and the color is reddish.

[0087] The framework structure of the recognition model in this example is as follows: Figure 10 As shown, three relatively independent and precise features (i.e., the instantaneous velocity feature of the target trajectory point in the agricultural machinery trajectory, the time difference feature between the target trajectory point and the previous trajectory point, and the direction feature of the agricultural machinery trajectory) and three range features (i.e., the direction variance feature of several trajectory points in the agricultural machinery trajectory, the velocity variance feature of several trajectory points in the agricultural machinery trajectory, and the velocity mean feature of several trajectory points in the agricultural machinery trajectory) were added to the input features of the recognition model. This not only increased the effectiveness and richness of the input features, but also enabled the recognition model to learn rich range features to optimize the explosion point problem.

[0088] Actual application effect diagrams are as follows Figure 11 As shown, red represents road trajectory type, and green represents farmland trajectory type.

[0089] The following is a supplementary explanation of the training process of the recognition model:

[0090] The recognition model was trained using Ubuntu (an open-source operating system based on Linux), with the code completed in the Tensorflow (an open-source deep learning framework) environment. The model was trained on an RTX3090 GPU (a high-performance graphics processor) with 24GB of GPU memory.

[0091] The training dataset accurately labeled 1,032 trajectories based on the actual work trajectories of farmers in multiple regions. These trajectories included various work modes such as sowing, rotary tillage, and harvesting. On average, each trajectory contained about 40 images, for a total of about 42,000 images.

[0092] During training, the batch size (the number of samples input in one iteration) is set to 2, for a total of 50 batches. The learning rate is set to 5e. -3 (The step size for parameter updates during model training) first goes through a warm-up (a learning rate scheduling strategy) from low to high, and then gradually decreases after reaching the maximum learning rate.

[0093] like Figure 12 As shown, the trained recognition model converges rapidly in the first 20 epochs (training periods) and then tends to level off, with the best accuracy of the recognition model being 98.6%.

[0094] Example 2

[0095] Corresponding to the aforementioned embodiments of the agricultural machinery trajectory identification method, this disclosure also provides embodiments of the agricultural machinery trajectory identification system.

[0096] Figure 13 A schematic diagram of a module for recognizing agricultural machinery trajectories provided as an exemplary embodiment of this disclosure, as shown below. Figure 13 As shown, the agricultural machinery trajectory recognition system includes:

[0097] Module 1 is used to acquire feature information of the agricultural machinery trajectory. The feature information includes at least: the instantaneous velocity feature of the target trajectory point in the agricultural machinery trajectory, the time difference feature between the target trajectory point and the previous trajectory point, the direction feature of the agricultural machinery trajectory, the direction variance feature of several trajectory points in the agricultural machinery trajectory, the velocity variance feature of several trajectory points in the agricultural machinery trajectory, and the velocity mean feature of several trajectory points in the agricultural machinery trajectory.

[0098] Recognition module 2 is used to input feature information into a pre-trained recognition model to output the agricultural machinery trajectory type. The agricultural machinery trajectory type includes road trajectory type and farmland trajectory type.

[0099] In one alternative implementation, see Figure 13 The agricultural machinery trajectory recognition system also includes:

[0100] The feature collaborative optimization module 3 is used to process feature information through the feature collaborative optimization module. The feature collaborative optimization module includes at least one channel attention module, spatial attention module, and multi-scale fusion module.

[0101] In one alternative implementation, see Figure 13 The agricultural machinery trajectory recognition system also includes:

[0102] Module 4 is used to determine the three-channel trajectory map based on feature information. The three-channel trajectory map includes a first three-channel trajectory map and a second three-channel trajectory map. The first three-channel trajectory map is determined based on the instantaneous velocity characteristics of the target trajectory point, the time difference characteristics between the target trajectory point and the previous trajectory point, and the directional characteristics of the agricultural machinery trajectory. The second three-channel trajectory map is determined based on the directional variance characteristics, velocity variance characteristics, and mean velocity characteristics of several trajectory points in the agricultural machinery trajectory.

[0103] Visualization module 5 is used to visualize agricultural machinery trajectory types based on a three-channel trajectory graph.

[0104] In an alternative implementation, the recognition model includes a post-processing layer, which is used to distinguish the background and agricultural machinery trajectory type of the three-channel trajectory map.

[0105] In one alternative implementation, the three-channel trajectory map is a satellite image at level 18.

[0106] In an alternative implementation, the determining module 4 is also used to determine the original image based on feature information.

[0107] In one alternative implementation, see Figure 13 The agricultural machinery trajectory recognition system also includes:

[0108] Edge module 6 is used to add 128 pixels of edge information to the image edges of the original image using an overlay technique to determine the final three-channel trajectory map.

[0109] For the system embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs.

[0110] Example 3

[0111] Figure 14 This is a schematic diagram of the structure of an electronic device according to an example embodiment of the present disclosure. The electronic device includes a memory, a processor, and a computer program stored in the memory and used to run on the processor. When the processor executes the computer program, it implements the agricultural machinery trajectory recognition method of any of the above embodiments. Figure 14 The electronic device 90 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0112] like Figure 14 As shown, the electronic device 90 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different system components (including memory 92 and processor 91).

[0113] Bus 93 includes a data bus, an address bus, and a control bus.

[0114] The memory 92 may include volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0115] The memory 92 may also include a program tool 925 (or utility) having a set (at least one) program module 924, such program module 924 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0116] The processor 91 executes various functional applications and data processing by running computer programs stored in the memory 92, such as the agricultural machinery trajectory recognition method provided in any of the above embodiments.

[0117] Electronic device 90 can also communicate with one or more external devices 94 (e.g., keyboard, pointing device, etc.). This communication can be performed through input / output (I / O) interface 95. Furthermore, electronic device 90 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 96. As shown, network adapter 96 communicates with other modules of electronic device 90 via bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with electronic device 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0118] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0119] Example 4

[0120] This disclosure also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the agricultural machinery trajectory recognition method provided in any of the above embodiments.

[0121] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0122] Example 5

[0123] This disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the method for identifying agricultural machinery trajectories as described above.

[0124] The program code for executing the computer program product disclosed herein can be written in any combination of one or more programming languages. The program code can be executed entirely on a user device, partially on a user device, as a stand-alone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0125] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A method for identifying a track of an agricultural machine, characterized in that The agricultural machine track recognition method comprises: obtaining feature information of the agricultural machine track; wherein the feature information at least comprises: instantaneous speed feature of a target track point in the agricultural machine track, time difference feature of the target track point and a previous track point of the target track point, direction feature of the agricultural machine track, direction variance feature of several track points in the agricultural machine track, speed variance feature of the several track points in the agricultural machine track, and speed mean value feature of the several track points in the agricultural machine track; inputting the feature information into a pre-trained recognition model to output an agricultural machine track type; wherein the agricultural machine track type comprises a road track type and a farmland track type.

2. The agricultural machine track identification method of claim 1, wherein, The pre-trained recognition model comprises a feature cooperative optimization module; The processing steps of the pre-trained recognition model on the input feature information comprise: processing the feature information through the feature cooperative optimization module, wherein the feature cooperative optimization module comprises at least one of a channel attention module, a spatial attention module, and a multi-scale fusion module.

3. The agricultural machine track identification method of claim 1, wherein, The agricultural machine track recognition method further comprises: determining a three-channel track graph based on the feature information; wherein the three-channel track graph comprises a first three-channel track graph and a second three-channel track graph; the first three-channel track graph is determined according to the instantaneous speed feature of the target track point in the agricultural machine track, the time difference feature of the target track point and the previous track point of the target track point, and the direction feature of the agricultural machine track; the second three-channel track graph is determined according to the direction variance feature of the several track points in the agricultural machine track, the speed variance feature of the several track points in the agricultural machine track, and the speed mean value feature of the several track points in the agricultural machine track; visualizing the agricultural machine track type based on the three-channel track graph.

4. The method of claim 3, wherein the agricultural machine trajectory is identified based on the agricultural machine trajectory data. The recognition model comprises a post-processing layer, wherein the post-processing layer is used to distinguish the background of the three-channel track graph and the agricultural machine track type.

5. The method of claim 3, wherein the agricultural machine trajectory is identified based on the agricultural machine trajectory data. The map level of the three-channel track graph is an 18-level satellite image.

6. The method of claim 3, wherein the agricultural machine trajectory is identified based on the agricultural machine trajectory data. The step of determining the three-channel track graph based on the feature information comprises: determining an original image based on the feature information; adding 128-pixel edge information to the image edge of the original image through an overlapping technology to determine a final three-channel track graph.

7. An agricultural machine track identification system, comprising: The agricultural machine track recognition system comprises: an obtaining module configured to obtain feature information of an agricultural machine track; wherein the feature information at least comprises: instantaneous speed feature of a target track point in the agricultural machine track, time difference feature of the target track point and a previous track point of the target track point, direction feature of the agricultural machine track, direction variance feature of several track points in the agricultural machine track, speed variance feature of the several track points in the agricultural machine track, and speed mean value feature of the several track points in the agricultural machine track; an identification module configured to input the feature information into a pre-trained recognition model to output an agricultural machine track type; wherein the agricultural machine track type comprises a road track type and a farmland track type.

8. An electronic device comprising a memory, a processor, and a computer program stored on the memory for running on the processor, characterized in that, The computer program is executed by the processor to implement the agricultural machine track identification method in any one of claims 1 to 6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the agricultural machine track identification method in any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the agricultural machine track identification method in any one of claims 1 to 6.

Images