AU2021102192A4

AU2021102192A4 - Remote sensing classification and identification method, device, equipment and storage medium for soil types

Info

Publication number: AU2021102192A4
Application number: AU2021102192A
Authority: AU
Inventors: Dejie Cui; Mengqi Duan; Xinwei LIU; Xiangyun Song; Guiling YU; Xiaoguang Zhang
Original assignee: Qingdao Agricultural University
Current assignee: Qingdao Agricultural University
Priority date: 2021-04-26
Filing date: 2021-04-26
Publication date: 2021-06-17
Anticipated expiration: 2029-04-26

Abstract

The invention applies to the field of computer technology and provides a remote sensing classification and recognition method, device, equipment and storage medium for soil types. The method comprises the following steps: Obtaining remote sensing images of the soil to be recognized; Extraction of remote sensing image texture features in the preset scale; Fusion of texture feature and remote sensing image to generate remote sensing classification image fused with texture feature; The remote sensing classification image is processed to determine the classification and recognition results of the soil to be identified. The invention provides a remote sensing classification and recognition method for soil type. By extracting the texture features that can be used to describe the spatial and geometric information of the soil in the remote sensing image, and fusing with the remote sensing image, the rich spatial and geometric information contained in the texture feature is used to assist the classification and recognition. The problem of insufficient spectral information in the classification and recognition of the existing technology is overcome, and the classification accuracy is improved, so as to ensure the remote sensing classification and recognition effect of the soil type. 1/8 FIGURES S102 Obtaining remote sensing image of soil to be identified 11 F r S104 According to the preset texture feature extraction rules and the texture features of the remote sensing image at the preset scale S106 The texture features are fused with the remote sensing image to generate a remote sensing classification image fused with the texture features S108 According to the preset classificationand recognition rules, the remote sensing classification image is processed to determine the classification and recognition results of the soil to be recognized Figure 1

Description

1/8

FIGURES

S102 Obtaining remote sensing image of soil to be identified

11 F r S104 According to the preset texture feature extraction rules and the texture features of the remote sensing image at the preset scale

S106 The texture features are fused with the remote sensing image to generate a remote sensing classification image fused with the texture features

S108 According to the preset classificationand recognition rules, the remote sensing classification image is processed to determine the classification and recognition results of the soil to be recognized

Figure 1

Remote sensing classification and identification method, device, equipment and

storage medium for soil types

TECHNICAL FIELD The invention belong to that technical field of computers, and particularly relate to a method, a device, equipment and a storage medium for remotely sensing classification and identification of soil types. BACKGROUND Spatial distribution of soil types is an important basis for precision agriculture and land resource management. Traditional mapping of soil types is mainly done through field investigation and indoor mapping, which has been widely used for many years. However, the traditional method has long cycle, high cost, complicated procedure, strong subjectivity and low precision of the final soil map. Digital mapping technology is proposed to update soil data quickly and conveniently. With the development of remote sensing technology, it has become a new way and method of digital soil mapping to obtain soil attributes and soil types data from remote sensing images. However, most of the variable indexes extracted by remote sensing technology at present reflect the spectral characteristics of soil types. For some soil types with the same spectral characteristics, it is difficult to identify them separately by spectral characteristics, which leads to the low classification accuracy and poor classification effect of the existing remote sensing technology. It can be seen that the existing remote sensing technology for classification and recognition of soil types still has the technical problems of low classification accuracy and poor classification effect. SUMMARY The embodiment of the invention aims to provide a method for classifying and identifying soil types by remote sensing, and aims to solve the technical problems that the existing remote sensing technology for classifying and identifying soil types still has low classification accuracy and poor classification effect. The embodiment of the invention is realized as follows: the remote sensing classification and recognition method for soil types comprises the following steps:

Acquiring remote sensing images of soil to be identified; Extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; Fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; And processing the remote sensing classified image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized. Another object of the embodiment of the present invention is to provide a remote sensing classification and recognition device for soil types, which comprises: The remote sensing image acquisition unit is used for acquiring the remote sensing image of the soil to be identified; A texture feature extraction unit, configured to extract texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; The fusion unit is used for fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; A classification recognition unit, configured to process the remote sensing classification image according to a preset classification recognition rule to determine a classification recognition result of the soil to be recognized Another object of an embodiment of the present invention is to provide a computer device, which comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor is caused to execute the steps of the soil type remote sensing classification and recognition method. Another object of an embodiment of the present invention is to provide a computer readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the processor is caused to execute the steps of the soil type remote sensing classification and recognition method described above. According to the remote sensing classification and recognition method for soil types provided by the invention, texture features which can be used for describing soil spatial and geometric information in remote sensing images are extracted and fused with spectral features originally contained in the remote sensing images to obtain remote sensing classification images fused with texture features, so that in the process of classifying and recognizing the remote sensing classification images, rich spatial and geometric information contained in the texture features can be used for auxiliary classification and recognition. The method overcomes the problem of insufficient spectral information in the prior art that only spectral features are used for classification and recognition, and improves the classification accuracy, thereby ensuring the remote sensing classification and recognition effect of soil types. BRIEF DESCRIPTION OF THE FIGURES Brief description of the drawings Fig. 1 is a step flow chart of a remote sensing classification and recognition method for soil types provided by an embodiment of the present invention; Fig. 2 is a step flow chart of another soil type remote sensing classification and recognition method provided by an embodiment of the present invention; Fig. 3 is a step flow chart of another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention; Fig. 4 is a step flow chart of yet another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention; Fig. 5 is a step flow chart of another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention; Fig. 6(a)- Fig. 6(c) are schematic diagrams of classification and recognition effects obtained by three different remote sensing classification and recognition methods of soil types; Fig. 7 is a structural schematic diagram of a soil type remote sensing classification and recognition device provided by an embodiment of the present invention;

Fig. 8 is an internal structure diagram of a computer device for implementing a method for classifying and identifying soil types by remote sensing provided by an embodiment of the present invention. DESCRIPTION OF THE INVENTION In order to make the object, technical solutions and advantages of the present invention more clearly understood, the present invention is described in further detail hereinafter in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are used only to explain the present invention and are not intended to limit the present invention. Fig. 1 shows a flow chart of steps of a remote sensing classification identification method of soil types provided by embodiments of the present invention, specifically comprising the following steps. S102, acquiring a remote sensing image of the soil to be identified. In the embodiment of the present invention, the remote sensing image is usually composed of a panchromatic band with a resolution of 15 m (band 8: 0.50-0.68m) and ten multispectral bands with a resolution of 30 m. Usually, in order to simplify the operation, the first seven visible spectral bands are selected as research objects. S104, extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule. In an embodiment of the present invention, the texture feature is a feature that expresses spatial structure and geometric information, and the texture feature at least includes one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation. As a preferred embodiment of the present invention, the texture features are referred to as information entropy. Information entropy is chosen as texture feature, and the subsequent classification effect is better. In an embodiment of the present invention, there are many methods for extracting texture features at a preset scale, and the present invention will not be specifically described here. Please refer to Figure 4 and its explanation for a specific preferred feasible scheme. S106, fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features.

In an embodiment of the invention, texture features are fused with the remote sensing image to generate a remote sensing classified image fused with texture features, wherein texture features describe spatial structure and geometric shape information of soil, while remote sensing images contain spectral information of soil, so the remote sensing classified image fused with texture features contains both spatial structure and geometric shape information of soil and spectral information of soil. S108, processing the remote sensing classified image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized. In the embodiment of the invention, by further classifying and identifying the remote sensing classified images containing the spatial structure, geometric shape information and spectral information of the soil, the types of each soil in the region can be determined according to the remote sensing classified images, thereby obtaining the classification and identification results of the soil to be identified. According to the remote sensing classification and recognition method for soil types provided by the invention, texture features which can be used for describing soil spatial structure and geometric form information in remote sensing images are extracted and fused with spectral features originally contained in remote sensing images to obtain remote sensing classification images fused with texture features, so that in the process of classifying and recognizing remote sensing classification images, rich spatial and geometric information contained in texture features can be utilized to assist in classification and recognition. The method overcomes the problem of insufficient spectral information in the prior art that only spectral features are used for classification and recognition, and improves the classification accuracy, thereby ensuring the remote sensing classification and recognition effect of soil types. As shown in Fig. 2, it is a step flow chart of another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention, which is detailed as follows. In the embodiment of the present invention, the difference from the step flow chart of a soil type remote sensing classification and recognition method shown in Fig. 1 is that the step S104 specifically comprises:

S204, extracting texture features of the remote sensing image at preset multiple scales according to preset texture feature extraction rules. In the embodiment of the invention, considering that different texture features contain different spatial and geometric information at different scales, by acquiring texture features at multiple scales, it can be comprehensively ensured that the extracted different texture features contain the most spatial and geometric information at a specific scale, thereby further improving the effect of subsequent classification. As a preferred embodiment of the present invention, the preset scale is at least one of x5 scale, 15x15 scale, 21x21 scale and 23x23 scale, and 5x5 scale, 15x15 scale, 21x21 scale and 23x23 scale are optimally selected,Under these scales, the texture features achieve the best classification effect for different soil types As shown in Fig. 3, it is a step flow chart of another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention, which is detailed as follows. The embodiment of the present invention is different from the step flow chart of a soil type remote sensing classification and recognition method shown in Fig. 1 in that after step S102, it further comprises: S302, performing radiation correction and geometric correction on the remote sensing image to generate a processed remote sensing image. In the embodiment of the present invention, radiation correction and geometric correction can be understood as preprocessing remote sensing images, which can also improve the subsequent classification effect, and the present invention will not be specifically explained. As shown in Fig. 4, it is a step flow chart of another method for remote sensing classification and recognition of soil types provided by an embodiment of the present invention, which is detailed as follows. In the embodiment of the present invention, the difference from the step flow chart of a soil type remote sensing classification and recognition method shown in Fig. 1 is that the step S104 specifically comprises: S402, extracting texture features of the remote sensing image at a preset scale based on the gray level co-occurrence matrix.

In the embodiment of the present invention, texture features are extracted through gray level co-occurrence matrix. Specifically, gray level co-occurrence matrix is a list of relative frequencies of pixel pairs separated by a certain distance in a given direction, which records the appearance times of different pixel pairs, and also records the relative positions and spatial information of pixels, so that the extracted texture features have better effects. As shown in Fig. 5, it is a step flow chart of another soil type remote sensing classification and recognition method provided by an embodiment of the present invention, which is detailed as follows. In this embodiment of the present invention, the difference from the step flow chart of a soil type remote sensing classification and recognition method shown in Fig. 1 is that the step S108 specifically comprises: S502, processing the remote sensing classified image based on the maximum likelihood classification method to determine the classification recognition result of the soil to be recognized. In an embodiment of the present invention, the classification method uses the maximum likelihood classification method, which can classify previously unrecognized pixels into one with the highest similarity probability among all categories according to statistical information such as the mean and standard deviation of soil types in remote sensing data. In order to further understand the technical scheme provided by the present invention, the principle of the soil type remote sensing classification and recognition method provided by the present invention will be described below in combination with the specific experimental process, and the details are as follows. 1. Determine the study area: Considering that pingdu city is a typical plain and hilly area in Jiaodong Peninsula, which is rich in soil types, it provides better conditions for soil classification research. Therefore, we chose pingdu city in the east of Shandong Province as the study area. It is located between 36 28 ' 15 "-37 02 '46" north latitude and 119 31 ' 30 "-120 19 ' 13" east longitude, with a total area of 3 175.63 km2,It is the largest county-level city in Shandong Province. Pingdu city belongs to continental climate in warm temperate zone and semi humid monsoon zone of East Asia, with four distinct seasons. The annual average temperature is 11.9°C, the annual average precipitation is 680 mm, and the sunshine hours are about 2700 hours. Its terrain is generally high in the north and low in the south. There are thirteen soil subtypes in the study area from mountainous area to plain, which are leaching cinnamon soil, cinnamon soil, coastal saline soil, neutral stony soil, neutral coarse boned soil, coarse-boned soil, brown soil, moist brown soil, moist tidal soil, tidal soil and Shajiang black soil (according to the classification of soil genesis in China). Among them, there are three sub-types of soil: Shajiang black soil, brown soil and fluvo-aquic soil, with the largest area and the widest distribution. 2. Collection and preprocessing of remote sensing data; In order to cover the whole study area, two Landsat8 OLI satellite images were acquired. The image consists of a panchromatic band (band 8: 0.50-0.68 m) with 15 m resolution and ten multispectral bands with 30 m resolution. To simplify the calculation, the first seven visible spectral bands were selected to participate in the study. The collected Landsat 8 OLI image data are corrected by radiation and geometry, and then the two images are spliced by histogram matching. Finally, the images in the study area with spatial resolution of 30 m, that is, the remote sensing images of the soil to be identified, are obtained by cutting with ArcGIS10.2. 3. Extracting texture features from remote sensing images Extracting texture features by selecting gray level co-occurrence matrix. Gray level co-occurrence matrix is a list of relative frequencies of pixel pairs which are calculated based on spatial co-occurrence matrix and separated by a certain distance in a given direction,It records the times of different pixel pairs, and also considers the relative positions and spatial information of pixels. In order to avoid the repetition of information represented by texture features, the following texture features are considered: mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation. Different texture features reflect different image information, so it is necessary to select appropriate texture features according to the actual classification needs to ensure that the texture features involved in classification can effectively improve the separability of soil types. The eight texture features extracted above are processed to generate their own texture mean curves, and then descriptive statistical analysis is carried out. The texture features with high separability for each soil type can be selected by comparing the curve characteristics and descriptive statistical characteristics. The research shows that homogeneity, information entropy, second moment and correlation can distinguish leaching cinnamon soil, brown soil and fluvo-aquic soil, and the texture feature values of different soil types are relatively dispersed, with good classification effect, among which information entropy has the best degree of dispersion. The residual texture features have certain classification effect, but the effect is weaker than homogeneity, information entropy, second moment and correlation. Therefore, information entropy is used as texture feature for fusion classification in subsequent experiments. 4. Determine the window size of the best scale information entropy of soil type Twelve different windows (from 3x3 to 25x25 odd windows) are used to extract texture features, that is, information entropy. Each window size will be fused with spectral data to create a new image data.

In order to quantitatively study the indexability of soil types under texture features with different window sizes, Jeffries-Matusita distance method was used to analyze the indexability of training samples. J-M distance is a measure of statistical separability between class pairs, which calculates the average difference between density functions of two classes. Based on a set of statistical features such as mean, variance and covariance matrix, this method calculates the degree of separation between two classes in band combination. The value range of J-M distance is 0-2, and the larger the value, the higher the separability between soil types. If this value is between 1.9 and 2, the two classes have good separability, and the obtained classes will be separated accurately. If the value is between 1.7 and 1.9, the separability between the two classes is good; If it is less than 1.7, the separability is poor. The J-M analysis results are obtained by statistical spectral data and separability of soil training samples in texture feature images of different windows. With the increase of texture feature window, the average separability of each soil type increases gradually, but the increment decreases gradually with the increase of texture feature window. When the information entropy window is 21x21, the average separability reaches the maximum and remains unchanged. On this basis, the window with texture feature information entropy of 21x21 is selected as the optimal single-scale texture feature for soil classification.

Although the overall accuracy and Kappa coefficient reach the highest values when the windows of texture features are 21x21, it is impossible for each soil type to reach the highest accuracy at the same time. Therefore, the posterior probability method is used to determine the best information entropy extraction window for each soil type, that is, the best extraction window is determined by the information entropy of different windows. The results show that when the classification accuracy of each soil type reaches the highest accuracy, the corresponding window sizes are 5x5 (fluvo-aquic soil), 15x15 (Shajiang black soil, brown soil), 21x21 (wet fluvo-aquic soil, brown soil) and 23x23 (cinnamon soil), so the multi-scale window of information entropy characteristic parameter combination is 5x5. However, considering that adding the number of texture features will significantly increase the number of input variables, thus increasing the calculation time, in practice, we can select any suitable scale from the above scales according to the actual situation to extract texture features, so as to ensure the calculation effect and efficiency. 5. Soil remote sensing classification with multi-scale texture feature parameters based on optimal window fusion The information entropy extracted from 5x5, 15x15, 21x21 and 23x23 windows is selected, combined and then fused into new remote sensing data for soil classification. According to the maximum likelihood classification, that is, according to the statistical information such as the mean and standard deviation of soil types in remote sensing data, the previously unrecognized pixels are classified into one with the highest similarity probability among all categories. In order to show the difference between the soil type remote sensing classification and recognition method provided by the present invention and the prior art, the classification and recognition effects obtained by adopting the above three different soil type remote sensing classification and recognition methods of separate classification of remote sensing image spectral data, fusion classification of single-scale texture features and remote sensing image spectral data and fusion classification of multi-scale texture features and remote sensing image spectral data are drawn as follows, specifically as shown in Figure 6 (a) Figure 6(c). Fig. 6(a) shows the effect of separate classification of remote sensing image spectral data, Fig. 6(b) shows the effect of fusion classification of single-scale texture features and remote sensing image spectral data, and Fig. 6(c) shows the effect of fusion classification of multi-scale texture features and remote sensing image spectral data. Comparing the classification effect diagram with the real data, it can be found that the classification accuracy based on remote sensing image spectral data alone is 40.99%, and high quality results cannot be obtained. In the process of fusion classification of single scale texture features and remote sensing image spectral data, the single-scale window is 21x21, and the texture features choose information entropy. At this time, the classification accuracy is 63.46%, which is 22.47% higher than that of spectral data. This is because the fusion of texture features into spectral data can make use of its rich spatial and geometric information, overcome the problem of insufficient spectral information of images, and improve the classification accuracy of soil types. Furthermore, in the process of fusion classification of multi-scale texture features and remote sensing image spectral data, the selected multi-scales are 5x5, 15x15, 21x21 and 23x23, and the information entropy is also selected for texture features. At this time, the classification accuracy reaches 74.32%, which is improved by 10.96% compared with the information entropy classification result extracted by the best single-scale window, because of the fusion of multi-scale texture features. Soil types with the same spectral characteristics but different properties can be identified. Therefore, the multi-scale fusion method of optimal texture feature parameters of each soil type can improve the classification accuracy of remote sensing images more than the multi-scale fusion method of optimal single-scale texture feature parameters of each soil type. As shown in Fig. 7, it is a structural schematic diagram of a soil type remote sensing classification and recognition device provided by an embodiment of the present invention, which specifically includes the following units: The remote sensing image acquisition unit 710 is configured to acquire a remote sensing image of the soil to be identified. In the embodiment of the present invention, the remote sensing image is usually composed of a panchromatic band with a resolution of 15 m (band 8: 0.50-0.68m) and ten multispectral bands with a resolution of 30 m. Usually, in order to simplify the operation, the first seven visible spectral bands are selected as research objects. A texture feature extraction unit 720, configured to extract texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule. In an embodiment of the present invention, the texture feature is a feature that expresses spatial geometric information, and the texture feature at least includes one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation. As a preferred embodiment of the present invention, the texture features are referred to as information entropy. Information entropy is chosen as texture feature, and the subsequent classification effect is better. In an embodiment of the present invention, there are many methods for extracting texture features at a preset scale, and the present invention will not be specifically described here,Please refer to Figure 4 and its explanation for a specific preferred feasible scheme. A fusion unit 730, configured to fuse the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features. In an embodiment of the invention, texture features are fused with the remote sensing image to generate a remote sensing classified image fused with texture features, wherein texture features describe spatial and geometric information of soil, while remote sensing images contain spectral information of soil, so the remote sensing classified image fused with texture features contains both spatial and geometric information of soil and spectral information of soil. A classification recognition unit 740, configured to process the remote sensing classification image according to a preset classification recognition rule to determine a classification recognition result of the soil to be recognized. In the embodiment of the invention, the type of each soil in the region can be determined according to the remote sensing classified image by further classifying and identifying the remote sensing classified image containing the spatial, geometric shape information and spectral information of the soil, so as to obtain the classification and identification result of the soil to be identified. According to the remote sensing classification and recognition device for soil types provided by the invention, texture features which can be used for describing soil spatial and geometric information in remote sensing images are extracted and fused with spectral features originally contained in the remote sensing images to obtain remote sensing classification images fused with texture features, so that in the process of classifying and recognizing the remote sensing classification images, rich spatial and geometric information contained in the texture features can be used for auxiliary classification and recognition, The method overcomes the problem of insufficient spectral information in the prior art that only spectral features are used for classification and recognition, and improves the classification accuracy, thereby ensuring the remote sensing classification and recognition effect of soil types. Fig. 8 shows an internal structure diagram of a computer device in one embodiment. As shown in Fig. 8, the computer equipment includes a processor, a memory, a network interface, an input device and a display screen which are connected through a system bus. The memory includes a nonvolatile storage medium and an internal memory. The non volatile storage medium of the computer equipment stores an operating system, and can also store a computer program. When the computer program is executed by a processor, the processor can realize the remote sensing classification and recognition method of soil types. A computer program can also be stored in the internal memory, and when the computer program is executed by the processor, the processor can execute the remote sensing classification and recognition method of soil types. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a trackball or a touch pad arranged on the shell of the computer equipment, and an external keyboard, touch pad or mouse. It can be understood by those skilled in the art that the structure shown in Fig. 8 is only a block diagram of some structures related to the scheme of this application, and does not constitute a limitation on the computer equipment to which the scheme of this application applies, the specific computer equipment may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements. In one embodiment, the soil type remote sensing classification and recognition device provided by this application can be implemented in the form of a computer program, and the computer program can be run on a computer device as shown in Fig. 8. The memory of the computer equipment can store various program modules constituting the remote sensing classification and recognition device for soil types, such as a remote sensing image acquisition unit 710, a texture feature extraction unit 720, a fusion unit 730 and a classification and recognition unit 740 shown in Fig. 7. A computer program composed of each program module enables the processor to execute the steps in the soil type remote sensing classification and recognition method of each embodiment of the application described in this specification. For example, the computer device shown in Fig. 8 can execute step S102; through the remote sensing image acquisition unit 710 in the soil type remote sensing classification and recognition device shown in Fig. 7; The computer device can execute step S104; through the texture feature extraction unit 720, the computer device may execute step S106 through the fusion unit 730 .

In one embodiment, a computer device is provided, which comprises a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to realize the following steps: Acquiring remote sensing images of soil to be identified; Extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; Fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; And processing the remote sensing classified image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized. In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the processor is caused to execute the following steps: Acquiring remote sensing images of soil to be identified; Extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; Fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; And processing the remote sensing classified image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized. It should be understood that although each step in the flowchart of each embodiment of the present invention is displayed in sequence as indicated by the arrow, these steps are not necessarily executed in sequence as indicated by the arrow. Unless explicitly stated herein, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Furthermore, at least a part of the steps in each embodiment may include a plurality of sub-steps or stages, which may not necessarily be completed at the same time, but may be executed at different times, and the execution sequence of these sub-steps or stages may not necessarily be carried out sequentially, but may be executed alternately or alternately with other steps or at least a part of sub-steps or stages of other steps. Those of ordinary skill in the art can understand that all or part of the flow of the method for realizing the above embodiments can be completed by instructing related hardware through a computer program, which can be stored in a nonvolatile computer readable storage medium, and the program can include the flow of the above embodiments. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include nonvolatile and/or volatile memory. The nonvolatile memory may include read-only memory (ROM), programmable ROM(PROM), electrically programmable ROM(EPROM), electrically erasable programmable ROM(EEPROM) or flash memory. The volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM(SRAM), dynamic RAM(DRAM), synchronous DRAM(SDRAM), double data rate SDRAM(DDRSDRAM), enhanced SDRAM(ESDRAM), synchronous link DRAM (SLDRAM), memory bus (Rambus) direct RAM(RDRAM), and direct memory bus. The technical features of the above-mentioned embodiments can be combined arbitrarily,In order to make the description concise, all possible combinations of the technical features in the above-mentioned embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, they should be considered as the scope recorded in this specification. The above examples only express several embodiments of the present invention, and their descriptions are specific and detailed, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for those of ordinary skill in the field, without departing from the concept of the present invention, several modifications and improvements can be made, which belong to the protection scope of the present invention. Therefore, the scope of protection of the patent of the present invention shall be subject to the appended claims. The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the scope of protection of the present invention.

Claims

THE CLAIMS DEFINING THE INVENTION ARE AS FOLLOWS: 1. A remote sensing classification and recognition method for soil types, comprising: Acquiring remote sensing images of soil to be identified; Extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean,

variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; Fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; And processing the remote sensing classified image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized.
2. The remote sensing classification and recognition method of soil type according to claim 1, wherein the step of extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule specifically comprises: According to a preset texture feature extraction rule, extracting texture features of

the remote sensing image at preset multiple scales.
3. The remote sensing classification and recognition method of soil type according to claim 1, characterized in that after the step of obtaining the remote sensing image of the soil to be recognized, the method further comprises: And performing radiation correction and geometric correction on the remote sensing image to generate a processed remote sensing image.
4. The remote sensing classification and recognition method of soil type according to claim 1, wherein the step of extracting texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule specifically comprises: And extracting texture features of the remote sensing image at a preset scale based on the gray level co-occurrence matrix.
5. The remote sensing classification and recognition method of soil type according to claim 1, wherein the step of processing the remote sensing classification image according to a preset classification and recognition rule to determine the classification and recognition result of the soil to be recognized specifically comprises:

And processing the remote sensing classified images based on the maximum likelihood classification method to determine the classified recognition result of the soil to be recognized.
6. The remote sensing classification and recognition method of soil types according to claim 1, wherein the texture features are expressed as information entropy.
7. The remote sensing classification and recognition method of soil type according to claim 1, wherein in the step of extracting texture features of the remote sensing image at preset scale according to preset texture feature extraction rules, the preset scale is at least one of 5x5 scale, 15x15 scale, 21x21 scale and 23x23 scale.
8. A remote sensing classification and recognition device for soil types, comprising: The remote sensing image acquisition unit is used for acquiring the remote sensing image of the soil to be identified; A texture feature extraction unit, configured to extract texture features of the remote sensing image at a preset scale according to a preset texture feature extraction rule; The texture features at least include one of mean, variance, homogeneity, contrast, similarity, information entropy, second moment and correlation; The fusion unit is used for fusing the texture features with the remote sensing image to generate a remote sensing classified image fused with texture features; And the classification recognition unit is used for processing the remote sensing classification image according to a preset classification recognition rule to determine the classification recognition result of the soil to be recognized.
9. A computer device, characterized by comprising a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the soil type remote sensing classification and recognition method according to any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the processor is caused to execute the steps of the soil type remote sensing classification and recognition method according to any one of claims 1 to 7.

FIGURES 1/8

Figure 1