Crop identification method and system

Abstract

The invention provides a crop identification method and system, and the method comprises the steps: obtaining a time sequence multi-polarization SAR image which comprises a crop sample and a to-be-identified ground object, and determining a logic time weight according to a time sequence post-scattering coefficient of the time sequence multi-polarization SAR image, the logic time weight is used for distinguishing the difference between the ground features under different seasons and different phenological characteristics, determining the category center of the ground feature to be identified according to the spatial information of the time sequence multi-polarization SAR image, and determining the target distance between the ground feature to be identified and the category center of the ground feature to be identified; and identifying the ground feature to be identified according to the logic time weight and the target distance to obtain an identification result. The accuracy of crop identification is improved.

Description

Technical Field

[0001] This specification relates to the technical field of agricultural applications, and particularly to a crop recognition method and system. Background Art

[0002] The accurate recognition of crops is of great significance for agricultural production management, food security assessment, etc.

[0003] Synthetic Aperture Radar (SAR), as an active microwave sensor, performs high-resolution earth observation through the synthetic aperture principle and electromagnetic waves, and has the advantage of being unaffected by weather and lighting conditions. In addition, SAR is sensitive to the dielectric and structural characteristics of crops and can effectively capture changes in crop growth status. Therefore, it has become an important data source in the fields of crop recognition, growth monitoring, and yield prediction. In the related art, crop recognition is mainly based on single-temporal SAR images.

[0004] However, single-temporal images can only reflect the information at a specific time point, thus limiting the recognition accuracy.

[0005] It should be noted that the content of the above related art is only the information known to the inventor personally, and does not mean that the above information has entered the public domain before the filing date of this specification, nor does it mean that it can become the prior art of this specification. Summary of the Invention

[0006] This specification provides a crop recognition method and system to avoid at least one of the above technical problems.

[0007] In a first aspect, this specification provides a crop recognition method, including:

[0008] Obtaining a time series of multi-polarization SAR images, where the time series of multi-polarization SAR images includes crop samples and the ground objects to be recognized;

[0009] Determining a logical time weight according to the temporal backscattering coefficient of the time series of multi-polarization SAR images, where the logical time weight is used to distinguish the differences between crops under different seasons and different phenological characteristics;

[0010] Determining the category center to which the ground object to be recognized belongs according to the spatial information of the time series of multi-polarization SAR images, and determining the target distance between the ground object to be recognized and its category center; and

[0011] Recognizing the ground object to be recognized according to the logical time weight and the target distance to obtain a recognition result.

[0012] In a second aspect, this specification provides a crop recognition system, including:

[0013] At least one storage medium storing at least one instruction set for crop recognition;

[0014] At least one processor communicatively connected to the at least one storage medium, wherein when the at least one processor runs, it reads the at least one instruction set and executes the method described in the first aspect according to the instructions of the at least one instruction set.

[0015] In a third aspect, this specification provides a computer-readable non-transitory storage medium, wherein the computer-readable non-transitory storage medium stores at least one instruction set, and the at least one instruction set is executed by at least one processor to implement the method described in the first aspect.

[0016] As can be seen from the above technical solutions, the crop recognition method and system provided in this specification obtain the temporal characteristics (such as logical time weights) and spatial characteristics (such as target distances) of ground objects respectively, and combine the temporal and spatial characteristics for crop recognition. This is equivalent to considering both the growth characteristics of crops in terms of time, such as seasonality and phenological characteristics, and the distribution and geometric attributes of ground objects in terms of space during the crop recognition process. Therefore, the accuracy, reliability, and precision of crop recognition can be improved.

[0017] Other functions of the crop recognition method and system provided in this specification will be partially listed in the following description. The creative aspects of the crop recognition method and system provided in this specification can be fully explained through practice or the use of the methods, devices, and combinations described in the detailed examples below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] To more clearly illustrate the technical solutions in the embodiments of this specification, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of this specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 A schematic diagram of the effect of the crop recognition method provided in the embodiments of this specification;

[0020] Figure 2 A schematic diagram of the application scenario of the crop recognition method provided in the embodiments of this specification;

[0021] Figure 3 A schematic diagram of the structure of the crop recognition system provided in the embodiments of this specification;

[0022] Figure 4 A schematic diagram of the process of the crop recognition method provided in an embodiment of this specification;

[0023] Figure 5 Schematic flowchart of the crop recognition method provided in another embodiment of this specification;

[0024] Figure 6 Schematic diagram of the image data vectors of known and unknown pixels in the time series multi-polarization SAR image provided in an embodiment of this specification. Detailed implementation manners

[0025] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with this specification. On the contrary, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.

[0026] It should be understood that in the embodiments of this specification, the terms "include" and "have" and any variations thereof are intended to cover but not be exclusive of inclusion. For example, a product or device including a series of components does not necessarily have to be limited to those components clearly listed, but may include other components not clearly listed or inherent to these products or devices.

[0027] In the embodiments of this specification, the term "and / or" describes the association relationship of associated objects and indicates that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after.

[0028] In the embodiments of this specification, the term "plurality" refers to two or more, and other quantifiers are similar thereto.

[0029] The terms "first", "second", "initial", "target", etc. in this specification are used to distinguish similar or homogeneous objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise indicated (Unless otherwise indicated). It should be understood that such terms can be interchanged under appropriate circumstances, for example, it is possible to implement in an order other than those given in the illustration or description of the embodiments of this specification.

[0030] The term "unit / module" used in this specification refers to any known or later developed hardware, software, firmware, artificial intelligence, fuzzy logic, or a combination of hardware or / and software code that can perform functions related to the element.

[0031] Based on the relevant descriptions in the above technical background, it can be known that there are technical problems with low recognition accuracy in the crop recognition methods in the related technologies. To avoid this technical problem, this specification presents a technical concept through creative labor: obtaining time series multi-polarized SAR images, such as obtaining SAR image data at different time points (time series) and using at least two polarization methods (multi-polarization). Analyze the characteristics of sample crops and the objects to be recognized on the time dimension and space dimension respectively based on the time series multi-polarized SAR images. And obtain the recognition result of the object to be recognized based on the characteristics in both time and space dimensions.

[0032] Relatively speaking, by using the method based on time series multi-polarized SAR images for crop recognition, it not only enhances the monitoring ability of surface dynamic changes, but also greatly improves the understanding and analysis ability of complex surface features due to its multi-dimensional information collection method. In addition, by combining the characteristics in both time and space dimensions for crop recognition, it not only utilizes the rich spatio-temporal characteristics, but also utilizes the rich spatial characteristics, that is, it fuses the spatio-temporal domain characteristics, and can effectively improve the accuracy of crop recognition.

[0033] For the convenience of readers' understanding of this specification, the application scenarios of this specification are introduced below.

[0034] The technical solution provided in this specification is applicable to scenarios where objects (such as objects to be recognized) need to be recognized.

[0035] Exemplarily, based on the technical solution provided in this specification, the objects in a certain area can be recognized to obtain corresponding recognition results, such as the types of objects. Among them, the objects can be crops, and specifically can be agricultural crops.

[0036] For example, as Figure 1 shown, through the technical solution provided in this specification, the agricultural crops in area A can be recognized. For the convenience of distinction, the blank positions in area A are areas without agricultural crops, and the positions corresponding to the black spots in area A are areas with agricultural crops.

[0037] In some embodiments, through the technical solution provided in this specification, the types of agricultural crops can be determined. For example, continuing to combine Figure 1 , through the technical solution provided in this specification, the types of agricultural crops in area A can be recognized, such as wheat or others.

[0038] In other embodiments, through the technical solution provided in this specification, the types of multiple different agricultural crops can be determined. For example, if there are multiple agricultural crops in a certain area, through the technical solution provided in this specification, the types of different agricultural crops can be obtained respectively.

[0039] In addition, in the scenario of determining the types of multiple different crops, through the technical solution provided in this specification, an identification result map can be obtained. And in the identification result map, different types of crops correspond to different color identifiers. For example, if different types of crops include wheat and corn, then in the identification result map, wheat can be marked with red, and corn can be marked with yellow.

[0040] Figure 2 FIG. is a schematic diagram of an application scenario of the crop identification method according to an embodiment of this specification, where the crop identification method of this specification can be applied to, for example Figure 2 the scenario 200 shown. As Figure 2 shown, the scenario 200 may include a target user 201, a client 202, a server 203, and a network 204.

[0041] The target user 201 may be a user who triggers the identification of crops. For example, the target user 201 may perform a target operation on the client 202 to trigger the identification of crops. In some embodiments, the target user 201 may upload time-series multi-polarization SAR images on the client 202 to trigger the identification of crops.

[0042] The client 202 may be an electronic device that provides an interaction function to the target user 201. For example, the client 202 may provide an interaction interface to the target user 201, and the target user 201 may perform interaction operations on the interaction page. In some embodiments, in response to detecting an operation of the target user 201 triggering the identification of crops, the client 202 executes the crop identification method described in this specification. At this time, the client 202 may store data or instructions for executing the crop identification method described in this specification and may execute or be used to execute the data or instructions. In some embodiments, the client 202 may include a hardware device with data information processing functions and necessary programs for driving the hardware device to work to execute the crop identification method described in this specification.

[0043] In some embodiments, the client 202 may include a mobile device, a tablet computer, a laptop computer, a built-in device of a motor vehicle, or the like, or any combination thereof. In some embodiments, the mobile device may include a smart home device, a smart mobile device, a virtual reality device, an augmented reality device, or the like, or any combination thereof. In some embodiments, the smart home device may include a smart TV, a desktop computer, etc., or any combination. In some embodiments, the smart mobile device may include a smart phone, a personal digital assistant, a game device, a navigation device, etc., or any combination thereof. In some embodiments, the built-in device in a motor vehicle may include an in-vehicle computer, an in-vehicle TV, etc.

[0044] In some embodiments, the client 202 may be installed with one or more applications (APPs). The APPs can provide the target user 201 with the ability to interact with the outside world through the network 204 and an interface. The APPs include, but are not limited to: web browser APPs, search APPs, chat APPs, shopping APPs, video APPs, financial management APPs, instant messaging tools, email clients, social platform software, and so on. In some embodiments, a target APP may be installed on the client 202. The target APP can collect time series multi-polarization SAR images for the client 202.

[0045] As Figure 2 shown, the client 202 can be communicatively connected to the server 203. Among them, the server 203 can be communicatively connected to one client 202 or multiple clients 202. In some embodiments, the client 202 can interact with the server 203 through the network 204 to receive or send messages, etc.

[0046] The server 203 can be a server that provides various services. For example, the server 203 can be a cloud server or a local server. The server 203 can be communicatively connected to one client 202 and receive the data sent by the client 202, or can be communicatively connected to multiple clients 202 and receive the data sent by each client 202 respectively.

[0047] In some embodiments, the crop recognition method described in this specification can be executed on the server 203. At this time, the server 203 can store the data or instructions for executing the crop recognition method described in this specification and can execute or be used to execute the data or instructions. The server 203 can include a hardware device with data information processing functions and the necessary programs for driving the hardware device to work.

[0048] The network 204 is a medium for providing a communication connection between the client 202 and the server 203. The network 204 can facilitate the exchange of information or data. As Figure 2 shown, the client 202 and the server 203 can be respectively connected to the network 204 and transmit information or data to each other through the network 204.

[0049] In some embodiments, network 204 can be any type of wired or wireless network, or a combination thereof. For example, network 204 can include a cable network, a wired network, an optical fiber network, a telecommunication network, an intranet, the Internet, a Local Area Network (LAN), a Wide Area Network (WAN), a Wireless Local Area Network (WLAN), a Metropolitan Area Network (MAN), a Public Switched Telephone Network (PSTN), a Bluetooth networkTM, a short-range wireless network (ZigBeeTM), a Near Field Communication (NFC) network, or a similar network.

[0050] In some embodiments, network 204 can include one or more network access points. For example, network 204 can include a wired or wireless network access point, such as a base station or an Internet exchange point, through which one or more components of client 202 and server 203 can be connected to network 204 to exchange data or information.

[0051] It should be noted that Figure 2 the numbers of client 202, server 203, and network 204 in

[0052] are merely illustrative. According to the implementation requirements, there can be any number of client 202, server 203, and network 204. And the crop recognition method provided in this specification can be executed entirely on client 202, entirely on server 203, or partially on client 202 and partially on server 203. Figure 2 That is to say, Figure 2 and the above description for

[0053] Figure 3Shows a hardware structure diagram of a crop recognition system 300 provided according to an embodiment of this specification. The crop recognition system 300 can execute the crop recognition method described in this specification. The crop recognition method is introduced in other parts of this specification. When the crop recognition method is executed on the client 202, the crop recognition system 300 can be the client 202. When the crop recognition method is executed on the server 203, the crop recognition system 300 can be the server 203. When the crop recognition method is partially executed on the client 202 and partially executed on the server 203, the crop recognition system 300 can be a system including the client 202 and the server 203.

[0054] As Figure 3 shown, the crop recognition system 300 may include at least one storage medium 303 and at least one processor 302. In some embodiments, the crop recognition system 300 may further include a communication port 304 and an internal communication bus 301. The crop recognition system 300 may further include I / O components 305.

[0055] The internal communication bus 301 can connect different system components. For example, the internal communication bus 301 can connect the storage medium 303, the processor 302, the communication port 304, and the I / O components 305.

[0056] The I / O components 305 support input / output between the crop recognition system 300 and other components.

[0057] The communication port 304 is used for data communication between the crop recognition system 300 and the outside world. For example, the communication port 304 can be used for data communication between the crop recognition system 300 and the network 204. The communication port 304 can be a wired communication port or a wireless communication port.

[0058] The storage medium 303 may include a data storage device. The data storage device can be a non-transitory storage medium or a transitory storage medium. For example, the data storage device can include one or more of a magnetic disk 3031, a read-only storage medium (ROM) 3032, or a random access storage medium (RAM) 3033. The storage medium 303 further includes at least one instruction set stored in the data storage device. The instruction set includes computer program code, and the computer program code can include programs, routines, objects, components, data structures, processes, modules, etc. for executing the crop recognition method provided in this specification.

[0059] At least one processor 302 may be communicatively connected to at least one storage medium 303. The at least one processor 302 is configured to execute the above-mentioned at least one instruction set. When the crop recognition system 300 runs, the at least one processor 302 reads the at least one instruction set and, according to the indication of the at least one instruction set, executes the crop recognition method provided in this specification. The processor 302 may execute all steps included in the crop recognition method. The processor 302 may be in the form of one or more processors. In some embodiments, the processor 302 may include one or more hardware processors, such as a microcontroller, a microprocessor, a reduced instruction set computer (RISC), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a central processing unit (CPU), a graphics processing unit (GPU), a physics processing unit (PPU), a microcontroller unit, a digital signal processor (DSP), a field-programmable gate array (FPGA), an advanced RISC machine (ARM), a programmable logic device (PLD), any circuit or processor capable of executing one or more functions, etc., or any combination thereof.

[0060] For illustrative purposes only, only one processor 302 is shown in the crop recognition system 300 in the drawings. However, it should be noted that the crop recognition system 300 in this specification may also include multiple processors. Therefore, the operations and / or method steps disclosed in this specification may be executed by one processor or jointly executed by multiple processors. For example, if it is described in this specification that the processor 302 of the crop recognition system 300 executes step A and step B, it should be understood that step A and step B may also be jointly or separately executed by two different processors 302 (e.g., the first processor executes step A, the second processor executes step B, or the first and second processors jointly execute steps A and B).

[0061] Please refer to Figure 4 , Figure 4 which is a schematic flowchart of the crop recognition method provided by an embodiment of this specification. Among them, Figure 4 the execution subject of the shown crop recognition method may be the crop recognition system. For the description of the crop recognition system, reference may be made to the above examples and will not be elaborated here.

[0062] As Figure 4 shown, the method includes the following S401 to S404:

[0063] S401: Obtain time-series multi-polarization SAR images, where the time-series multi-polarization SAR images include crop samples and features to be recognized.

[0064] The time-series multi-polarization SAR images can be understood as SAR image data collected using at least two polarization methods at different time points.

[0065] That is to say, the SAR image data for crop recognition in the specification includes multi-polarized SAR image data at different times in the time series. For example, for a certain area, the crop recognition system can obtain multi-polarized SAR image data at a certain moment every Monday within several months.

[0066] Therefore, relatively speaking, these images obtained by the crop recognition system can capture the changes of surface coverings (i.e., ground objects, such as crop samples and ground objects to be recognized) over time and seasons.

[0067] Among them, the multi-polarization can be dual-polarization, such as Horizontal transmit-Horizontal receive (HH), Vertical transmit-Vertical receive (VV). The multi-polarization can also be full-polarization, such as on the basis of the aforementioned HH and VV, adding Horizontal transmit-Vertical transmit (HV) and Vertical transmit-Horizontal transmit (VH).

[0068] The crop sample can be understood as the crop used for reference when performing crop recognition on the ground object to be recognized.

[0069] In this embodiment, the quantity, type, etc. of the crop samples are not limited. Taking the quantity as an example, the quantity of the crop samples can be one or multiple. When the quantity of the crop samples is one, it is equivalent to using a certain crop as a reference to recognize the ground object to be recognized. When the quantity of the crop samples is multiple, it is equivalent to using multiple crops as references to recognize the ground object to be recognized.

[0070] In addition, based on the above analysis, it can be known that when the quantity of the crop samples is multiple, the multiple crop samples can be crop samples of the same category, such as all being wheat. Or, the multiple crop samples can be crop samples of different categories, such as wheat and corn, etc.

[0071] Correspondingly, the ground object to be recognized can be understood as the ground object for which the type is obtained by performing recognition based on the reference crop (i.e., the crop sample). Similarly, the quantity of the ground object to be recognized can be one or multiple, which is not limited in this embodiment.

[0072] It is worth noting that this specification mainly takes the ground objects in time - series multi - polarimetric SAR images including crop samples and objects to be recognized as an example for elaboration. In some other embodiments, the ground objects in time - series multi - polarimetric SAR images may also only include crop samples or objects to be recognized.

[0073] That is to say, the crops used as a reference (i.e., crop samples) and the objects to be recognized (i.e., objects to be recognized) can be ground objects in the same image or in different images.

[0074] Regarding obtaining time - series multi - polarimetric SAR images, the following examples can be adopted to achieve it:

[0075] In one example, the crop recognition system can be connected to a collection device and receive the time - series multi - polarimetric SAR images collected and sent by the collection device.

[0076] In another example, the crop recognition system can provide a tool for loading data, and users can transmit the time - series multi - polarimetric SAR images to the crop recognition system through this data - loading tool.

[0077] Among them, the data - loading tool can be an interface for connecting to external devices, such as an interface for connecting to other storage devices, and obtain the time - series multi - polarimetric SAR images transmitted by the external device through this interface; the data - loading tool can also be a display device. For example, the crop recognition system can output an interface with a data - loading function on the display device, and users can import the time - series multi - polarimetric SAR images into the crop recognition system through this interface.

[0078] S402: Determine the logical time weight according to the temporal backscatter coefficient of the time - series multi - polarimetric SAR image, where the logical time weight is used to distinguish the differences between ground objects in different seasons and different phenological characteristics.

[0079] The backscatter coefficient is a key parameter in SAR images, which is used to describe the reflection ability of the earth's surface to radar waves. It quantifies the energy intensity of the radar waves returning from the earth's surface to the sensor.

[0080] Correspondingly, the temporal backscatter coefficient can be understood as the change of the reflection ability of ground objects (such as crop samples) to radar waves over time. Different ground objects (such as crop samples) have different reflection characteristics to radar waves at different growth stages.

[0081] The logical time weight can be understood as that the crop recognition system assigns different importance or weights to the SAR images collected at different times according to the characteristics of ground objects (such as crop samples) in different seasons and phenological periods, so as to reflect the dependence relationship between the time series of crop samples and seasonal and phenological characteristics, thereby increasing the separability between different ground objects, that is, more accurately distinguishing different crops.

[0082] S403: Determine the category center of the ground object to be recognized according to the spatial information of the time series multi-polarized SAR images, and determine the target distance between the ground object to be recognized and its category center.

[0083] Spatial information can be understood as the information containing the spatial distribution characteristics and geometric attributes of ground objects. For example, spatial information can include: geographical location, texture characteristics, shape and size, spatial relationship, backscattering coefficient, polarization information, etc.

[0084] Among them, the texture characteristic can be understood as the change rule of the gray level between adjacent pixels, which can reflect physical attributes such as the roughness and flatness of the ground object surface. The spatial relationship can be understood as the relative position relationship between different ground objects, including proximity, inclusion, etc.

[0085] The category center can be understood as the central position of a certain type of crop sample in the multi-dimensional feature space. This center can be determined by analyzing the spatial information of all samples in this category. Correspondingly, the category center of the ground object to be recognized can be understood as the central position of the ground object to be recognized in the multi-dimensional feature space.

[0086] In this step, determining the category center of the ground object to be recognized based on spatial information can be understood as determining the category center not only based on the characteristics of the ground object to be recognized itself, but also in combination with the characteristics of the ground objects adjacent to it in the physical space of the ground object to be recognized, so as to further determine the target distance based on the characteristics of the adjacent ground objects.

[0087] It should be noted that "adjacent" can be the adjacency of one pixel or the adjacency of multiple pixels, which is not limited in this embodiment.

[0088] In some embodiments, "determining the target distance between the ground object to be recognized and its category center" in S403 may include the following step 11 and step 12:

[0089] Step 11: Obtain the time series feature curves corresponding to the ground object to be recognized and its category center respectively.

[0090] The time series feature curve can be simply referred to as the feature curve.

[0091] The time-series feature curve of the object to be recognized can be understood as the curve representing the features of the object to be recognized in the time series. The time-series feature curve of the category center can be understood as the curve representing the features of the category to which the object to be recognized belongs in the time series.

[0092] Step 12: Determine the target distance when the respective time-series feature curves are warped and aligned.

[0093] Warped alignment can be abbreviated as alignment, which can be understood as deforming the time-series feature curve of the object to be recognized and / or the time-series feature curve of the category center so that the highest or lowest points of the two appear at the same time point as much as possible.

[0094] Correspondingly, when the time-series feature curves of the object to be recognized and the category center are warped and aligned, the crop recognition system calculates the distance between the two. This distance is the minimum Euclidean distance between the two, and this minimum distance is determined as the target distance.

[0095] Exemplarily, the spatial clustering method based on Iterative Self-Organizing Data Analysis (ISODATA) in the crop recognition system can calculate the minimum Euclidean distance from the object to be recognized in the time-series multi-polarized SAR image to the center of its category by dynamically adjusting the number of clusters and the center, so as to obtain the target distance.

[0096] For example, taking the number of objects to be recognized as multiple as an example:

[0097] If the set of objects to be recognized in the time-series multi-polarized SAR image is X = {x (x,y)}, where is a vector containing d-dimensional features such as time-series backscattering coefficients and texture features.

[0098] The ISODATA algorithm realizes the following process through iterative optimization:

[0099] In the initialization stage, the crop recognition system randomly selects K initial cluster centers j is the th cluster center. And set the merging threshold T m and the splitting threshold T s and the maximum number of iterations T max . In each iteration, the crop recognition system can calculate the Euclidean distance i from the object to be recognized x to the current cluster center Formula 1:

[0100]

[0101] Thus, the crop recognition system assigns each object to be recognized to the cluster with the minimum distance That is, it satisfies Subsequently, the crop recognition system obtains the updated cluster centers according to Equation 2 Equation 2:

[0102]

[0103] wherein, wherein is the number of samples within the current class. If the distance between the j-th and k-th cluster centers then they are merged into the same class; if the sample standard within a certain class then it splits into two classes along the direction of the maximum variance. The algorithm terminates when the maximum number of iterations is reached or the change in the center converges. The finally output distance is the minimum Euclidean distance from the object x i to its final belonging cluster center as shown in Equation 3: as shown in Equation 3:

[0104]

[0105] Combining the analysis of the above steps 11 and 12, it can be seen that in this embodiment, the crop recognition system first performs warping alignment on the time series feature curve, and then determines the target distance under the condition of warping alignment, which can improve the effectiveness and reliability of the target distance.

[0106] For example, taking the same kind of crop as an example, some crops may be planted first and some later. The growth trends of the two parts of the crops may not differ much, but due to time issues, when the time series feature curves of the two are not warping aligned, they may show great differences. After warping alignment, the two parts of the crops can show better commonalities in growth trends. Thus, the determined target distance can be relatively more accurate.

[0107] S404: Identify the object to be recognized according to the logical time weight and the target distance, and obtain the recognition result.

[0108] Combining the above analysis for S401 to S404, it can be seen that in this embodiment, the crop recognition system can respectively obtain the time characteristics (such as logical time weight) and spatial characteristics (such as target distance) of the object, so as to combine the time characteristics and spatial characteristics for crop recognition. It is equivalent to considering both the growth characteristics of the crop in terms of time, such as seasonality and phenological characteristics, and the distribution and geometric attributes of the object in space during the crop recognition process. Therefore, the accuracy, reliability, and precision of crop recognition can be improved.

[0109] In some embodiments, in order to further improve the recognition accuracy, the crop recognition system can combine the farmland mask corresponding to the recognition area to implement the recognition of the object to be recognized.

[0110] Exemplarily, refer to Figure 5 , Figure 5 which is a schematic flow chart of the crop recognition method provided in another embodiment of this specification. As Figure 5 shown, the method includes the following S501 to S505:

[0111] S501: Obtain time series multi-polarization SAR images, where the time series multi-polarization SAR images include crop samples and ground objects to be recognized.

[0112] It can be understood that, in order to avoid cumbersome statements, for the same or similar technical features in this embodiment and the above embodiments, reference can be made to the above examples, and this embodiment will not be elaborated here.

[0113] For example, for the implementation principle of S501, reference can be made to the description of S401 in the above example, and details will not be repeated here.

[0114] S502: Determine the logical time weight according to the temporal backscattering coefficient of the time series multi-polarization SAR images, where the logical time weight is used to distinguish the differences between ground objects under different seasons and different phenological characteristics.

[0115] Similarly, for the implementation principle of S502, reference can be made to the description of S402 in the above example, and details will not be repeated here.

[0116] S503: Determine the class center to which the crop samples belong according to the spatial information of the time series multi-polarization SAR images, and determine the target distance between the crop samples and their class center.

[0117] Similarly, for the implementation principle of S503, reference can be made to the description of S403 in the above example, and details will not be repeated here.

[0118] S504: Determine the farmland mask corresponding to at least one image in the time series multi-polarization SAR images, where the farmland mask is used to characterize the distribution information of farmland and non-farmland in at least one image.

[0119] The farmland mask can be understood as a binary or classification map used to identify and extract the location and scope of the farmland area. The farmland mask is a geospatial data product, usually represented in a raster format, where each pixel is labeled as "farmland" or "non-farmland".

[0120] Exemplarily, combining the above analysis, it can be known that the time series multi-polarization SAR images include multi-polarization SAR images at different time points. Therefore, the time series multi-polarization SAR images include multiple images.

[0121] The crop recognition system can obtain an image from multiple images and determine the farmland mask of this image. That is, based on this image, it determines the distribution information of farmland and non-farmland in the corresponding area. For example, which part of the corresponding area is the area corresponding to farmland and which part is the area corresponding to non-farmland.

[0122] In addition, the crop recognition system can also determine the farmland mask based on multiple images. For example, the crop recognition system first determines the initial farmland masks based on each image respectively, and then determines the final farmland mask by averaging.

[0123] S505: Identify the object to be recognized according to the logical time weight, target distance, and farmland mask to obtain the recognition result.

[0124] Correspondingly, after determining the distribution information of farmland and non-farmland in the corresponding area, the crop recognition system can combine this distribution information based on the logical time weight and target distance to determine the recognition result for identifying the object to be recognized.

[0125] Based on the above analysis of S501 to S505, it can be seen that in this embodiment, when the crop recognition system determines the recognition result, it not only considers the time characteristics and spatial characteristics of the crops, but also further considers the distribution information of farmland and non-farmland in the area where the crops are located. And the distribution information of farmland and non-farmland can indicate whether there are really crops in the corresponding area. Therefore, the accuracy, reliability, and precision of crop recognition can be further improved.

[0126] In some embodiments, in order to improve the quality of time-series multi-polarized SAR images, reduce noise effects, enhance useful information, and prepare more suitable data for subsequent analysis. The crop recognition system can preprocess the time-series multi-polarized SAR images to perform operations such as Figure 4 , Figure 5 based on the preprocessed time-series multi-polarized SAR images.

[0127] This embodiment does not limit the preprocessing method. For example, the preprocessing can include radiometric calibration, multi-looking, filtering, terrain correction, decibel conversion, image registration, etc.

[0128] In some embodiments, the filtering preprocessing can include two dimensions: performing the first filtering process on the images in the time-series multi-polarized SAR images, and performing the second filtering process on the image data of the same pixel point in the time-series multi-polarized SAR images at different times.

[0129] That is to say, the filtering preprocessing can include the first filtering process from the dimension of the image itself, and can also include the second filtering process of the image data from the time dimension.

[0130] Correspondingly, the farmland mask can be determined based on the result of the first filtering process, and the logical time weight and the target distance are determined based on the result of the second filtering process.

[0131] For example, before determining the farmland mask, the crop recognition system can first perform a first filtering process on the time series multi-polarization SAR images to filter the images themselves. Then, the farmland mask is determined from at least one of the filtered images. Alternatively, after obtaining at least one image, the crop recognition system can filter the at least one image and then determine the farmland mask based on the filtering, so as to make the determined farmland mask more accurate.

[0132] Again, for example, before determining the logical time weight and the target distance, the crop recognition system can first perform a second filtering process on the time series multi-polarization SAR images to filter a series of image data in the time dimension. And based on this, the logical time weight and the target distance are determined, so that the determined logical time weight and target distance are affected by as little interference as possible, that is, the accuracy and reliability of the logical time weight and the target distance can be improved.

[0133] In some embodiments, S505 may include the following steps 21 and 22:

[0134] Step 21: Determine the similarity between the object to be recognized and the crop sample according to the logical time weight and the target distance.

[0135] Exemplarily, combining the above analysis, the logical time weight can characterize the dependence relationship between the crop in terms of season and phenological characteristics, so as to increase the separability between different objects. The target distance can characterize the difference between the object to be recognized and the crop sample determined based on the spatial information of the object.

[0136] Therefore, the similarity determined by the crop recognition system based on the logical time weight and the target distance is equivalent to determining the correlation between the object to be recognized and the crop sample from multiple dimensions such as season, phenological characteristics, and space. Therefore, the accuracy and reliability of the determined similarity can be improved.

[0137] In some embodiments, the crop recognition system can use the warping distance between the object to be recognized and the crop sample as the crop similarity.

[0138] Exemplarily, the crop recognition system can determine the warping distance between the object to be recognized and the crop sample according to the logical time weight and the target distance, and determine the warping distance as the similarity between the object to be recognized and the crop sample.

[0139] Among them, the warping distance is used to characterize the pixel offset for registering the object to be recognized and the crop sample.

[0140] For example, the crop recognition system adjusts the vectors of the features to be recognized and the vectors of the sample crops through a series of geometric transformations (such as translation, rotation, scaling, affine transformation, etc.), so that the vectors of the features to be recognized and the vectors of the sample crops at the same geographical location are as close as possible in position.

[0141] Correspondingly, the distance generated by the adjustment is the warping distance.

[0142] Relatively speaking, the warping distance can help quantify the registration error. A smaller warping distance means higher registration accuracy, while a larger one indicates possible large errors or mismatching problems. By determining the warping distance as the similarity degree, the crop recognition system can accurately and effectively determine the difference between the feature to be recognized and the crop sample, that is, improve the accuracy and reliability of the determined similarity degree.

[0143] Step 22: Determine the recognition result according to the similarity degree and the farmland mask.

[0144] Exemplarily, after obtaining the similarity degree and the farmland mask, the similarity degree can be multiplied by the farmland mask to determine the recognition result based on the multiplication result.

[0145] For example, if the similarity degree of a certain pixel is 0.9 and the farmland mask is 0 (i.e., non-farmland), then the recognition result is that the feature on this pixel (i.e., the feature to be recognized) is a non-crop.

[0146] Another example, if the similarity degree of a certain pixel is 0.9 and the farmland mask is 1 (i.e., farmland), then the recognition result is that the feature on this pixel (i.e., the feature to be recognized) is a crop. And if the crop corresponding to this similarity degree is wheat, then the feature to be recognized is wheat.

[0147] Based on the above analysis of Step 21 and Step 22, on the basis of determining the correlation between the feature to be recognized and the sample crops, the crop recognition system can further combine the farmland mask to determine the recognition result, so as to improve the accuracy and reliability of the recognition result.

[0148] Combined with the above analysis, it can be known that the number of crop samples may be multiple. Then, correspondingly, the similarity degrees are also multiple, and one crop sample corresponds to one similarity degree.

[0149] Correspondingly, in this case, Step 22 may include: determining the maximum similarity degree from each similarity degree, and determining the recognition result according to the maximum similarity degree and the farmland mask.

[0150] Exemplarily, if the number of crop samples is 2, which are wheat and corn respectively. The similarity degree corresponding to wheat is A, and the similarity degree corresponding to corn is B, and A > B.

[0151] Then the crop recognition system can determine the recognition result according to the similarity A and the farmland mask. If the similarity A reaches the preset threshold and the farmland mask represents farmland, it can be determined that the recognition result is that the object to be recognized is a crop, specifically wheat.

[0152] Relatively speaking, the growth trends among the same crops are more fitting. Therefore, by determining the recognition result through the maximum similarity and the farmland mask, the crop recognition system can improve the accuracy and reliability of the determined recognition result.

[0153] Based on the above analysis, it can be known that the crop recognition system can determine the similarity based on the warping distance. In some embodiments, the steps for the crop recognition system to determine the warping distance may include the following steps 31 to 33:

[0154] Step 31: Determine the initial distance between the object to be recognized and the crop sample on the time series vector.

[0155] Exemplarily, the crop recognition system can obtain the image data vector of the object to be recognized on the time series and the image data vector of the crop sample on the time series, and determine the initial distance according to the obtained image data vectors corresponding to the object to be recognized and the crop sample on the time series respectively.

[0156] That is to say, the initial distance can be understood as the difference between the image data vectors of the object to be recognized and the crop sample on the time series.

[0157] For example, as Figure 6 shown, the time series multi-polarization SAR image includes the nth SAR image, where n is an integer greater than 1. P is the image data vector of the known pixel such as the crop sample on the time series. Q is the image data vector of the unknown pixel such as the object to be recognized on the time series.

[0158] Step 32: Correct the initial distance according to the logical time weight and the target distance to obtain the corrected distance.

[0159] Combined with the above analysis, it can be known that the logical time weight is determined based on the time characteristics, and the target distance is determined based on the spatial characteristics.

[0160] Therefore, in this step, the crop recognition system can correct the initial distance based on the time characteristics and the spatial characteristics to make the corrected distance relatively more accurate.

[0161] In some embodiments, the crop recognition system can calculate the corrected distance d based on Equation 4 i,j , Equation 4:

[0162]

[0163] where, ∣pi -q j ∣ is the initial distance, p i is the image data vector of the crop sample in the time series, q j is the image data vector of the feature to be recognized in the time series, w t (i, j) is the logical time weight and can be expressed based on Equation 5. Equation 5:

[0164]

[0165] Combined with the above analysis, is the target distance corresponding to the feature to be recognized in the x-th row and y-th column. N is the number of SAR images used to determine the corrected distance. The size of N is not limited in this embodiment and can be determined by the crop recognition system based on requirements, historical records, experiments, etc. For example, N = 5.

[0166] Step 33: Determine the warping distance according to the corrected distance.

[0167] Correspondingly, after obtaining the corrected distance, the crop recognition system can determine the warping distance based on the corrected distance.

[0168] Based on the above analysis of Steps 31 to 33, in this embodiment, the crop recognition system can correct and obtain the warping distance from the time characteristics and spatial characteristics. It can make the warping distance consider both the time characteristics and the spatial characteristics. Therefore, the accuracy and reliability of the determined warping distance can be improved.

[0169] In some embodiments, the corrected distance includes multiple matrix elements, and the matrix elements include a starting point and an ending point. Step 33 may include the following Steps 331 and 332:

[0170] Step 331: Based on the matrix elements, determine the minimum cumulative cost from the starting point to the ending point.

[0171] Exemplarily, the corrected distance can be represented in the form of a matrix, and the matrix includes multiple elements (i.e., matrix elements). One element can correspond to one pixel.

[0172] The matrix elements include the starting pixel (i.e., the element), which can be called the starting point. The matrix elements include the ending pixel, which can be called the ending point.

[0173] It can be understood that there are multiple ways from the starting point row to the ending point. Therefore, the corrected distance can be a set including multiple distances.

[0174] Correspondingly, the crop recognition system can respectively determine the cumulative cost corresponding to each distance from the starting point to the ending point, and determine the minimum cumulative cost from each cost.

[0175] Among them, the cumulative cost can be understood as the sum of all local "costs" accumulated on the path from one point (or pixel) to another point (or pixel).

[0176] Correspondingly, in this embodiment, the minimum cumulative cost can be understood as the minimum value of the sum of all local "costs" accumulated on the path from the starting point to the ending point. Or, the minimum cumulative cost can be understood as a relatively optimal solution found in the entire search space, that is, among all possible distance schemes, the one that results in the lowest cumulative cost is selected.

[0177] Exemplarily, the crop recognition system can obtain the cumulative cost Γ by using recursion and boundary, continuity, and monotonicity constraints. For example, the crop recognition system can calculate each cumulative cost based on Equation 6, Equation 6:

[0178]

[0179] Among them, m×n is a matrix of m rows multiplied by n columns, which can be represented by the matrix element D m×n denoted as, γ m,n is the element in the m-th row and n-th column.

[0180] Step 331: Determine the distance corresponding to the minimum cumulative cost as the warping distance.

[0181] Exemplarily, after obtaining the minimum cumulative cost, the crop recognition system can determine the distance corresponding to the minimum cumulative cost as the warping distance, thereby determining the similarity.

[0182] Continuing with the above example, the minimum cumulative cost can be represented by Equation 6, Equation 6:

[0183] γ i,j = d i,j + min{γ i-1,j , γ i-1,j-1 , γ i,j-1}

[0184] Among them, the matrix element γ i,j is the minimum cumulative cost from the starting point (1,1) to the point (i,j) in the matrix element D m×n . Then γ m,n is the minimum cumulative cost from the starting point (1,1) to the ending point (m,n). That is, the warping distance.

[0185] Combined with the above analysis of steps 331 and 332, in this embodiment, the crop recognition system determines the warping distance by taking the distance corresponding to the minimum cumulative cost. Considering the global information rather than relying only on local features, it can more accurately reflect the true differences between the ground objects to be recognized and the crop samples, thereby improving the accuracy of crop recognition. In addition, the minimum cumulative cost method helps to overcome partial matching errors caused by noise, occlusion, or other interference factors. Since it does not simply rely on the best match of a single point or a small area, but seeks the overall optimal solution, it has better resistance to outliers in the data.

[0186] It should be noted that the above examples are only used to exemplarily illustrate the possible implementation manners of the crop recognition method in this specification, and should not be construed as a limitation on the implementation manners of the crop recognition method in this specification. Exemplarily, based on the above technical concepts, some of the above technical features can be combined to obtain a new embodiment; new technical features can also be added based on the above examples to obtain a new embodiment; some technical features can also be reduced based on the above examples to obtain a new embodiment; some of the technical features in the above examples can be replaced with other technical features; or the order of some of the technical features in the above examples can be adjusted to obtain a new embodiment, etc., which will not be listed one by one here.

[0187] Based on the above technical concepts, this specification also provides a computer-readable non-transitory storage medium, in which at least one instruction set is stored. When the at least one instruction set is executed by a processor, the steps of the crop recognition method described in this specification are implemented.

[0188] In some possible embodiments, various aspects of this specification can also be implemented in the form of a program product, which includes program code. When the program product runs on the crop recognition system 300, the program code is used to cause the crop recognition system 300 to execute the steps of the crop recognition method described in this specification. The program product for implementing the above method can be a portable compact disc read-only memory (CD-ROM) including program code and can run on the crop recognition system 300. However, the program product of this specification is not limited to this. In this specification, the readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in combination with an instruction execution system. The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. The computer-readable storage medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, and this readable medium can send, propagate, or transmit a program for use by or in combination with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium can be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above. The program code for performing the operations of this specification can be written in any combination of one or more programming languages, including object-oriented programming languages - such as Java, C++, etc., and also including conventional procedural programming languages - such as the "C" language or similar programming languages. The program code can be executed entirely on the crop recognition system 300, partially on the crop recognition system 300, executed as an independent software package, partially on the crop recognition system 300 and partially on a remote crop recognition system, or entirely on a remote crop recognition system 300.

[0189] The above description has been made of specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims can be performed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the figures do not necessarily require a particular order or a sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0190] In summary, after reading this detailed disclosure, those skilled in the art will understand that the foregoing detailed disclosure may be presented only by way of example and may not be restrictive. Although not explicitly stated herein, those skilled in the art can understand that this specification is intended to encompass various reasonable changes, improvements, and modifications to the embodiments. These changes, improvements, and modifications are intended to be proposed by this specification and are within the spirit and scope of the exemplary embodiments of this specification.

[0191] Furthermore, certain terms in this specification have been used to describe embodiments of this specification. For example, "one embodiment", "an embodiment", and / or "some embodiments" mean that the specific features, structures, or characteristics described in connection with that embodiment may be included in at least one embodiment of this specification. Thus, it should be emphasized and understood that two or more references to "an embodiment" or "one embodiment" or "alternative embodiments" in various parts of this specification do not necessarily all refer to the same embodiment. Additionally, the specific features, structures, or characteristics may be appropriately combined in one or more embodiments of this specification.

[0192] It should be understood that in the foregoing description of the embodiments of this specification, for the purpose of helping to understand a feature and for the purpose of simplifying this specification, this specification combines various features in a single embodiment, drawing, or its description. However, this does not mean that the combination of these features is necessary. Those skilled in the art may very well mark out some of the devices as separate embodiments for understanding when reading this specification. That is to say, the embodiments in this specification can also be understood as an integration of multiple sub - embodiments. And the content of each sub - embodiment is also valid when it has fewer features than all the features of a single foregoing disclosed embodiment.

[0193] Every patent, patent application, published patent application, and other materials cited herein, such as articles, books, specifications, publications, documents, references, etc. (excluding any historical prosecution files associated therewith), are hereby incorporated by reference for all purposes relevant to this document, e.g., in the specification and claims of this document. However, in the event of any inconsistency or conflict between the descriptions, definitions, and / or terms of the above materials and those used in this document, the descriptions, definitions, and / or terms used in this document shall prevail.

[0194] Finally, it should be understood that the embodiments of the application disclosed herein are illustrative of the principles of the embodiments of this specification. Other modified embodiments are also within the scope of this specification. Accordingly, the embodiments disclosed in this specification are presented by way of example only and not by way of limitation. Those skilled in the art may adopt alternative configurations in accordance with the embodiments in this specification to implement the application in this specification. Therefore, the embodiments of this specification are not limited to the embodiments precisely described in the application.

Claims

1. A crop recognition method, comprising: Obtaining a time series of multi-polarized SAR images, the time series of multi-polarized SAR images including crop samples and features to be recognized; Determining a logical time weight according to the temporal backscattering coefficient of the time series of multi-polarized SAR images, wherein the logical time weight is used to distinguish the differences between features under different seasons and different phenological characteristics; Determining the category center to which the feature to be recognized belongs according to the spatial information of the time series of multi-polarized SAR images, and determining the target distance between the feature to be recognized and its category center; And Recognizing the feature to be recognized according to the logical time weight and the target distance to obtain a recognition result.

2. The method according to claim 1, wherein The method further comprises: Determining a farmland mask corresponding to at least one image in the time series of multi-polarized SAR images, the farmland mask being used to characterize the distribution information of farmland and non-farmland in the at least one image; And, the recognizing the feature to be recognized according to the logical time weight and the target distance to obtain a recognition result includes: recognizing the feature to be recognized according to the logical time weight, the target distance, and the farmland mask to obtain the recognition result.

3. The method according to claim 2, wherein The recognizing the feature to be recognized according to the logical time weight, the target distance, and the farmland mask to obtain the recognition result includes: Determining the similarity between the feature to be recognized and the crop sample according to the logical time weight and the target distance; and Determining the recognition result according to the similarity and the farmland mask.

4. The method according to claim 3, wherein, The determining the similarity between the feature to be recognized and the crop sample according to the logical time weight and the target distance includes: Determining a warping distance between the feature to be recognized and the crop sample according to the logical time weight and the target distance, wherein the warping distance is used to characterize the pixel offset for registering the feature to be recognized and the crop sample; and Determining the warping distance as the similarity.

5. The method according to claim 4, wherein, The determining the warping distance between the feature to be recognized and the crop sample according to the logical time weight and the target distance includes: Determining an initial distance between the feature to be recognized and the crop sample on the time series vector; Correcting the initial distance according to the logical time weight and the target distance to obtain a corrected distance; and Determining the warping distance according to the corrected distance.

6. The method according to claim 5, wherein The corrected distance includes a plurality of matrix elements, and the matrix elements include a starting point and an ending point; The determining the warping distance according to the corrected distance includes: Determining a minimum cumulative cost from the starting point to the ending point based on the matrix elements; And Determining the distance corresponding to the minimum cumulative cost as the warping distance.

7. The method according to any one of claims 3 to 6, wherein, The number of crop samples is multiple, and one crop sample corresponds to one similarity; the determining the recognition result according to the similarity and the farmland mask includes: Determining the maximum similarity from each similarity; and Determining the recognition result according to the maximum similarity and the farmland mask.

8. The method according to any one of claims 1 to 6, wherein Determining the target distance between the object to be recognized and the center of its category includes: Obtaining the respective time-series feature curves of the crop sample and the center of the category; and Determining the target distance when the respective time-series feature curves are warped and aligned.

9. The method according to any one of claims 2 to 6, wherein The method further includes: Preprocessing the time-series multi-polarization SAR image, where the preprocessing includes performing a first filtering process on the images in the time-series multi-polarization SAR image and performing a second filtering process on the image data of the same pixel point in the time-series multi-polarization SAR image at different times; And, the farmland mask is determined based on the result of the first filtering process, and the logical time weight and the target distance are determined based on the result of the second filtering process.

10. A crop recognition system, comprising: At least one storage medium storing at least one instruction set for crop recognition; At least one processor communicatively connected to the at least one storage medium, wherein when the at least one processor runs, it reads the at least one instruction set and executes the method according to any one of claims 1 to 9 according to the indication of the at least one instruction set.

Images