Mining area vegetation detection method based on satellite remote sensing image

Abstract

The invention relates to the technical field of image processing, in particular to a mining area vegetation detection method based on a satellite remote sensing image, and the detection method comprises the steps: firstly collecting a remote sensing image of a mining area, and evaluating the feature complexity of the remote sensing image and each channel image; then, calculating the importance degree of each channel according to the remote sensing image and the feature complexity and contrast of each channel image of the remote sensing image; then, extracting the feature vectors of the remote sensing image and the image of each channel, and calculating the weight of each channel in combination with the importance degree of each channel; updating the pixel value of each pixel point in the remote sensing image according to the weight of each channel to obtain an updated remote sensing image; and finally, segmenting the updated remote sensing image by using an OTSU algorithm to obtain a vegetation region. According to the method, the accuracy of the image segmentation result is improved, and accurate detection of the mining vegetation area is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a mining area vegetation detection method based on satellite remote sensing images. Background Art

[0002] While obtaining resources, mining activities inevitably have a negative impact on the ecological environment around the mining area, among which vegetation damage is a key issue. Since vegetation growth is an important indicator of ecological environment quality, accurate detection of vegetation in mining areas is of great significance for assessing the degree of ecological environment damage.

[0003] The prior art usually collects remote sensing images of mining areas, then segments the remote sensing images, and obtains vegetation areas based on the segmentation results to achieve vegetation detection. The OTSU algorithm is a common image segmentation method. For example, the Chinese patent document with the existing announcement number CN107705283B discloses a particle and bubble collision detection method based on OTSU image segmentation. The method includes: firstly optimizing the image by using image preprocessing operations, and segmenting bubbles and particles by the OTSU algorithm; then fitting particles and bubbles into circles by combining the least squares method, and finally tracking the positions of all particles and detecting whether particles and bubbles collide.

[0004] However, when using the above OTSU algorithm to segment remote sensing images of mining areas, since a large area of ​​the mining area is covered by vegetation, the grayscale histogram of the remote sensing image usually does not show two clearly identifiable peaks, that is, it presents a non-bimodal shape. For such remote sensing images with non-bimodal grayscale histograms, it is difficult for the OTSU algorithm to accurately determine the threshold for distinguishing vegetation from other parts during segmentation, which affects the segmentation effect. This makes it impossible for the algorithm to accurately and completely segment the vegetation area from the entire remote sensing image, affecting the accuracy and reliability of the vegetation detection results. Summary of the invention

[0005] In order to solve the problem that there is a large area of ​​complex vegetation in the remote sensing image of the mining area, the grayscale histogram of the remote sensing image presents a non-bimodal shape, the OTSU algorithm has a poor segmentation effect on the remote sensing image, and affects the accuracy and reliability of the vegetation detection result, the present invention proposes a mining area vegetation detection method based on satellite remote sensing images, comprising: Collect remote sensing images of the mining area and evaluate the feature complexity of the remote sensing images and their channel images; According to the feature complexity of the remote sensing image and its channel images, as well as the contrast of each channel image, the importance of each channel is calculated ,in, For the The importance of each channel, is the feature complexity of the remote sensing image, For remote sensing images The feature complexity of the channel image, For remote sensing images The contrast of the channel image, is a linear normalization function; Extract the feature vectors of the remote sensing image and its channel images, and calculate the weight of each channel of the remote sensing image based on the importance of each channel: ,in, For remote sensing images The weight of each channel, is the feature vector of the remote sensing image and The cosine similarity between the feature vectors of the channel images, is the natural exponential function, is the total number of channels; For the pixel points in the remote sensing image, the pixel values ​​of the pixel points are updated according to the weights of each channel to obtain an updated remote sensing image. The updated remote sensing image is segmented by the OTSU algorithm, and the vegetation area is determined according to the segmentation results to realize vegetation detection in the mining area.

[0006] The above technical solution can quantitatively describe the remote sensing image and its channel images from multiple dimensions by evaluating the feature complexity of the remote sensing image and its channel images, which is helpful to distinguish different types of ground objects, because for vegetation detection in mining areas, different types of vegetation and other ground objects (such as bare land, mining facilities, etc.) have different feature complexity. And further, the feature complexity and contrast of each channel image are comprehensively considered to calculate the importance of each channel. The feature complexity reflects the complex characteristics of the image, and the contrast can reflect the degree of difference between different areas in the image. Combining the two can more comprehensively measure the importance of each channel of the remote sensing image for vegetation detection. And further, by calculating the cosine similarity between the feature vector of the remote sensing image and the feature vector of each channel image, the similarity between them can be measured, and the weight of each channel is calculated in combination with the previously calculated importance. The weight can accurately reflect the relative importance of each channel in the final vegetation detection, and can more accurately assign weights to different channels, improve the accuracy of the final detection results, and have better adaptability to vegetation detection in complex mining environments. Furthermore, by updating the pixel points according to the calculated weights, the importance and feature similarity information of different channels obtained from the previous analysis can be integrated into the pixel values ​​of the image to form an updated remote sensing image, so that the image contains information that is more conducive to vegetation detection. The weighted updated remote sensing image contains information that comprehensively considers the importance and feature similarity of each channel. Compared with the original image or the simply processed image, it is more conducive to highlighting the vegetation area and improving the accuracy of vegetation detection. The OTSU threshold segmentation method can find the optimal segmentation threshold according to the updated image, overcomes the shortcomings of the original OTSU algorithm in processing mining area remote sensing images (grayscale histogram is non-bimodal), and can more accurately segment the vegetation area, improve the reliability and accuracy of vegetation detection, and ultimately achieve effective detection of vegetation in mining areas, providing more reliable data support for evaluating the degree of ecological environmental damage in mining areas.

[0007] Preferably, the method for evaluating the feature complexity of the remote sensing image and its channel images is: For any image in the remote sensing image and its channel images, the local binary pattern is used to obtain the eigenvalue of each pixel in the image, and a histogram of the eigenvalues ​​is constructed. The horizontal axis of the histogram is different eigenvalues, and the vertical axis is the number of pixels corresponding to each eigenvalue. Calculate the feature complexity of this image: ;in, is the feature complexity of the image, is the variance of the number of pixels corresponding to all eigenvalues ​​in the histogram, is the maximum number of pixels corresponding to all eigenvalues ​​in the histogram. is the number of eigenvalues ​​that appear in the image, is the total number of eigenvalues.

[0008] The above technical solution can reflect the complexity of the image texture more comprehensively and meticulously by comprehensively analyzing the information of multiple dimensions of the image.

[0009] Preferably, a method for updating the pixel value of a pixel point according to the weight of each channel is: The pixel values ​​of the pixel points in each channel are directly weighted and summed according to the weights of each channel to update the pixel values ​​of the pixel points.

[0010] The above technical solution helps to improve the integrity and accuracy of image information. By integrating information from different channels, it can more accurately represent the information of pixel points and provide better input for subsequent OTSU threshold segmentation.

[0011] Preferably, another method for updating the pixel value of a pixel point according to the weight of each channel is: First, if the weight of a channel is the smallest, set the weight of the channel to 0; Then, the pixel values ​​of the pixel points in each channel are weighted and summed according to the weights of each channel to update the pixel values ​​of the pixel points.

[0012] The above technical solution excludes the channel with the smallest weight, which can reduce unnecessary calculations and reduce the noise or erroneous information caused by low-value or interfering channels, so that the final updated pixel value can more accurately reflect the characteristics of the vegetation area.

[0013] Preferably, the method for determining the vegetation area according to the segmentation result is: The pixel points with a pixel value of 0 in the segmentation result of the updated remote sensing image are regarded as vegetation pixels, and the area formed by all vegetation pixels is regarded as the vegetation area.

[0014] Preferably, the method for extracting the feature vectors of the remote sensing image and its channel images is: For any image in the remote sensing image and its channel images, a histogram of the pixel values ​​of the image is constructed. The horizontal axis of the histogram is different pixel values, and the vertical axis is the number of pixel points corresponding to each pixel value. The number of pixel points corresponding to all pixel values ​​in the image constitutes the feature vector of the image.

[0015] The above technical solution uses the number of pixel points corresponding to all pixel values ​​in the image as the elements of the feature vector. This method is simple and direct and can effectively capture the overall feature information of the image.

[0016] Preferably, the contrast of each channel image is any one of Weber contrast, Michelson contrast and root mean square contrast.

[0017] Preferably, the channels are a red channel, a green channel and a blue channel.

[0018] Preferably, the vegetation detection further comprises: After the vegetation area is determined, the ratio of the number of pixels in the vegetation area to the number of all pixels in the updated remote sensing image is taken as the vegetation coverage rate of the mining area.

[0019] The above technical solution can intuitively reflect the proportion of vegetation in the mining area in the entire area in a quantitative way, which is of great significance for comprehensively understanding the current status of the ecological environment in the mining area and detecting its changing trend over time.

[0020] The present invention has the following effects: The present invention analyzes remote sensing images from multiple dimensions through a unique method, and updates the pixel values ​​of pixel points in the remote sensing images, so that the updated images can better highlight the vegetation areas. The OTSU algorithm is used to accurately find the segmentation threshold based on the updated remote sensing images, thereby overcoming the defects of the traditional OTSU algorithm in segmenting remote sensing images, improving the accuracy of image segmentation, and being able to accurately segment the vegetation areas in the remote sensing images, making the vegetation detection results in the mining areas more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a schematic flow chart of the method of the present invention; Figure 2 is a remote sensing image of the present invention; Figure 3 is the grayscale histogram of the remote sensing image of the present invention; Figure 4 It is an initial segmentation effect diagram of the remote sensing image of the present invention; Figure 5 is a histogram of characteristic values ​​of the remote sensing image of the present invention; Figure 6 is a blue channel image of the remote sensing image of the present invention; Figure 7 is a green channel image of the remote sensing image of the present invention; Figure 8 is a red channel image of the remote sensing image of the present invention; Fig. 9 It is a segmentation effect diagram of the updated remote sensing image of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] Reference Figure 1 The present invention provides a mining area vegetation detection method based on satellite remote sensing images, comprising steps S1 to S7: S1: Collect remote sensing images of the mining area.

[0025] Drones are a flexible and cost-effective remote sensing platform that can carry cameras or sensors and are usually used for high-resolution data collection in small areas. Therefore, in this step, drones carry multispectral cameras to collect remote sensing images of mining areas. Multispectral cameras are key equipment for acquiring remote sensing images and are often used in areas such as crop growth assessment, coverage assessment, and water quality testing. Multispectral cameras can simultaneously acquire image information in multiple different bands, including the common red, green, blue, and near-infrared bands. Different objects have different reflective characteristics in these bands. For example, vegetation has a higher reflectivity in the near-infrared band. By recording these differences with multispectral cameras, rich data can be provided for subsequent analysis, thereby achieving accurate detection of vegetation coverage.

[0026] like Figure 2 As shown in the collected remote sensing images, the vegetation distribution is large and complex, such as Figure 3 As shown in , the grayscale histogram of the remote sensing image presents an obvious non-bimodal shape. Figure 4 As shown in the figure, the segmentation effect of the remote sensing image using the OTSU algorithm is poor. The black part represents the vegetation area, and the white part represents other land objects. It is difficult to accurately distinguish vegetation from other land objects. This shows that for such remote sensing images with non-bimodal grayscale histograms, the OTSU algorithm has certain limitations and cannot meet the needs of accurate segmentation of vegetation distribution.

[0027] Therefore, in the subsequent steps, the present invention first obtains each channel image of the remote sensing image. The remote sensing image is an image after grayscale processing. The pixel value of each pixel in the remote sensing image is the grayscale value of the pixel. The pixel value of each pixel in each channel image of the remote sensing image is equal to the channel value of the pixel. For example, in the red channel image, the pixel value of each pixel is the red channel value, and the same applies to the other channel images.

[0028] Then, by analyzing the feature complexity of the remote sensing image and its channel images, the importance of each channel is calculated according to the feature complexity of the remote sensing image and its channel images and the contrast of each channel image, the feature vector of the remote sensing image and its channel images is extracted, and the weight of each channel of the remote sensing image is calculated in combination with the importance of each channel. For the pixel points in the remote sensing image, the pixel values ​​of the pixel points are updated according to the weights of each channel to obtain the updated remote sensing image. The remote sensing image after threshold segmentation is updated, and the vegetation area is determined according to the segmentation result, which improves the segmentation effect to achieve accurate vegetation detection.

[0029] S2: Evaluate the feature complexity of remote sensing images and their channel images.

[0030] Generally speaking, because vegetation and open space in remote sensing images have different characteristics, image segmentation algorithms can be used to segment vegetation from images to achieve vegetation detection. However, the actual situation is more complicated. In the mining environment, there are often multiple plants in the same area, and the growth status of these plants is different. This leads to different pixel features corresponding to different plants in remote sensing images. For example, healthy green plants and plants affected by pests and diseases have different pixel features such as color and texture in remote sensing images, which increases the complexity of segmenting remote sensing images.

[0031] Taking the above into consideration, the present invention quantifies the complexity of remote sensing images. The greater the complexity, the more features of the pixels contained in the remote sensing image, and the more likely it is that the remote sensing image contains vegetation in multiple different states.

[0032] In one embodiment, for any image in the remote sensing image and each channel image thereof, the complexity can be evaluated according to the following method: Use the local binary pattern to obtain the eigenvalue of each pixel in the image, and construct a histogram of the eigenvalues. The horizontal axis of the histogram is the different eigenvalues, and the vertical axis is the number of pixels corresponding to each eigenvalue.

[0033] The complexity of the image is:

[0034] In this formula, is the feature complexity of the image, is the variance of the number of pixels corresponding to all eigenvalues ​​in the histogram, that is, the variance of the ordinates corresponding to all eigenvalues ​​(abscissas) of the histogram. is the maximum number of pixels corresponding to all eigenvalues ​​in the histogram. is the number of eigenvalues ​​that appear in the remote sensing image, is the total number of eigenvalues, .

[0035] In one embodiment, the method for evaluating the feature complexity of a remote sensing image is: First, the local binary mode is used to obtain the characteristic value of each pixel in the remote sensing image. Specifically, the local binary mode under eight-point sampling is used. This method can effectively capture the local texture features of the remote sensing image. Different texture structures correspond to different characteristic values, providing rich basic data for subsequent analysis. Implementers can flexibly select sampling methods according to actual conditions to meet the needs of image feature extraction in different scenarios.

[0036] Then, the characteristic values ​​of all pixels in the remote sensing image are counted, and the histogram of the characteristic values ​​is constructed according to the characteristic values ​​of all pixels, such as Figure 5 As shown in the figure, the horizontal axis of the histogram is different eigenvalues, and the vertical axis is the frequency of each eigenvalue in the remote sensing image, that is, the number of pixels corresponding to each eigenvalue. For example, the vertical axis of the histogram at the horizontal axis of 20 is 10, which means that in the remote sensing image, there are 20 pixels with an eigenvalue of 20. The histogram intuitively shows the distribution of eigenvalues ​​in remote sensing images, which provides a key basis for further analysis of the feature complexity of remote sensing images.

[0037] Finally, the feature complexity of the remote sensing image is calculated based on the above histogram information:

[0038] In this formula, is the feature complexity of the remote sensing image, is the variance of the number of pixels corresponding to all eigenvalues ​​in the histogram, that is, the variance of the ordinates corresponding to all eigenvalues ​​(abscissas) of the histogram. is the maximum number of pixels corresponding to all eigenvalues ​​in the histogram. is the number of eigenvalues ​​that appear in the remote sensing image, is the total number of eigenvalues. Since eight-point sampling is used here, the total number of eigenvalues .

[0039] For the formula : when When it is small, it means that the number of times each eigenvalue appears in the histogram of the eigenvalues ​​is relatively close, indicating that in the remote sensing image, the distribution of the eigenvalues ​​of the pixels is more dispersed, which implies that there may be a variety of different local features in the remote sensing image. For example, there are different types of vegetation, different forms of mines, exposed rocks and soil, and other ground objects intertwined in the remote sensing image, presenting complex and diverse textures, which makes the feature complexity of the remote sensing image greater.

[0040] On the contrary, when When it is large, it means that the number of times each eigenvalue appears in the histogram of the eigenvalues ​​varies greatly, which suggests that there may be prominent and relatively simple local features in the remote sensing image, such as large areas of single vegetation coverage or vast flat mining areas, which makes the feature complexity of the remote sensing image correspondingly smaller.

[0041] For the formula : Since the eigenvalues ​​of the local binary pattern represent the local features of the image, The larger it is, the fewer cases where different pixels in the remote sensing image have the same eigenvalue, which means that the remote sensing image contains more diverse local features and the feature complexity of the remote sensing image is greater. The smaller it is, the more cases where different pixels in the remote sensing image have the same eigenvalue, which means that the remote sensing image contains fewer local features, so the feature complexity of the remote sensing image is also smaller. Divide by , to achieve The normalization operation makes The value range is standardized to The interval makes it easier to compare feature complexity between different images.

[0042] For the formula : It is a key reference point for the frequency distribution of eigenvalues ​​in the histogram. In essence, it reflects which eigenvalue corresponds to the largest number of pixels among all the different eigenvalues, that is, the frequency of the dominant local features in the remote sensing image. For example, in a remote sensing image, if the eigenvalues ​​of some pixels correspond to It is very large, which means that the local feature represented by the eigenvalue is widely present in the remote sensing image, indicating that these pixels are large areas in the remote sensing image, such as a large area of ​​specific vegetation texture or a common mineral structure.

[0043] For the formula :This proportional relationship affects the calculation results of feature complexity. Smaller and When it is relatively large, The value of will be smaller, which shows that although there is a dominant local feature ( The larger the value, the more uniform the distribution of other eigenvalues ​​( The smaller it is), at this time The closer the value of is to 1, the final calculated feature complexity will be relatively large, which means that there are many different types of features in the remote sensing image. Although there is a feature that has an advantage in area, the advantage is not absolute, and other features are also widely present and evenly distributed. For example, in a remote sensing image of a mining area, there may be dense forest vegetation areas, exposed rock areas, reclaimed farmland areas, and some small mine pit buildings, etc., resulting in a high feature complexity of the remote sensing image.

[0044] On the contrary, if Larger and Relatively small, The value of will be larger, which indicates that there is a relatively unobtrusive local feature in the remote sensing image ( The smaller the value, the greater the difference in the distribution of other eigenvalues ​​( The larger the The closer the value is to 0, the final calculated feature complexity will be relatively small, which means that there are no prominent dominant local features in the remote sensing image, and the distribution of other feature values ​​is very different. For example, in some small-scale mining areas in remote sensing images, there are a small amount of vegetation, scattered rocks and small mines. Although the distribution of their feature values ​​is very different, the number of feature value pixels corresponding to each feature is not large, resulting in a low feature complexity of the remote sensing image.

[0045] In summary, the above technical scheme can accurately quantify the feature complexity of remote sensing images. The greater the feature complexity, the more different types of objects there are in the remote sensing images. On the contrary, the smaller the feature complexity, the simpler the types of objects in the remote sensing images.

[0046] Furthermore, the present invention also takes into account that, in general, the various colors in any image are composed of pixel values ​​of the red channel, the green channel, and the blue channel, and the pixel values ​​of different channels express different colors differently. Figure 6 , Figure 7 , Figure 8 As shown in Figure 6 and Figure 8 It can be seen that in remote sensing images, the pixel values ​​of vegetation pixels and soil pixels in the blue channel are very similar, but there are obvious differences between vegetation pixels and soil pixels in the red channel. Therefore, this difference can be used to assist vegetation detection.

[0047] First, obtain the feature complexity of each channel image of the remote sensing image (the same method as the method of obtaining the feature complexity of the remote sensing image). Take the red channel image as an example: First, use the local binary mode to obtain the feature value of each pixel in the red channel image. Specifically, the local binary mode under eight-point sampling is used. Then, count the feature values ​​of all pixels in the red channel image, and construct a histogram of the feature values ​​based on the feature values ​​of all pixels. The horizontal axis of the histogram is different feature values, and the vertical axis is the number of pixels corresponding to each feature value, that is, the frequency of each feature value in the red channel image. Finally, calculate the feature complexity of the red channel image based on the information of the histogram. The calculation formula is the same as the calculation formula of the feature complexity of the remote sensing image.

[0048] S3: Calculate the importance of each channel according to the feature complexity of the remote sensing image and its channel images, as well as the contrast of each channel image.

[0049] Contrast is the degree of difference between the light and dark parts of an image. A high-contrast channel image usually means that the difference between the light and dark parts in the channel image is larger, indicating that the channel contains more details and change information, while a low-contrast image is relatively bland, with a smaller difference between the light and dark parts, which is more likely to imply that the channel image contains less information and it is more difficult to extract useful features from it.

[0050] In remote sensing images, for different types of objects, their contrast in different channels is different. For example, vegetation may have a higher contrast in the near-infrared channel because the reflectivity of vegetation has unique characteristics in the near-infrared band, which is quite different from the surrounding soil, water, etc., thus forming a higher contrast. This high contrast helps to identify and distinguish vegetation from other objects, while some objects may have a lower contrast in some channels, which may mean that their features in this channel are not prominent and difficult to distinguish from other objects. In short, calculating contrast helps determine which channels are more conducive to the extraction and identification of object features.

[0051] In one embodiment, the method for obtaining the contrast of each channel image can be selected from Weber contrast, Michelson contrast and root mean square contrast, and Weber contrast is selected here.

[0052] Weber contrast is based on the brightness of the pixels in the image, and measures the contrast by calculating the relative difference in brightness between adjacent areas. It can keenly capture subtle brightness changes in the image and is suitable for scenes with high requirements for image detail contrast. For example, when distinguishing the subtle texture differences between vegetation and surrounding soil in remote sensing images, Weber contrast can play a good role and accurately present the detail contrast between the two. Michelson contrast focuses on the overall brightness distribution of the image, and performs well in measuring the overall brightness and darkness contrast of the image. It can intuitively reflect the relative proportional relationship between the brightest and darkest parts of the image. For example, when analyzing the contrast between large areas of water and land in remote sensing images, this method can be used to clearly present the significant brightness and darkness contrast between the two, which is of great significance for distinguishing macroscopic features. Root mean square contrast quantifies the discrete degree of pixel brightness values ​​in the image from a statistical point of view. This method is more sensitive to the comprehensive changes in brightness in the image and can reflect the overall contrast characteristics of the image. When analyzing remote sensing images containing multiple types of objects and complex changes in object brightness, root mean square contrast can provide comprehensive contrast information, which helps to grasp the image characteristics as a whole.

[0053] Furthermore, the contrast of the remote sensing image is obtained by selecting one of Weber contrast, Michelson contrast and root mean square contrast, and Weber contrast is selected here.

[0054] In one embodiment, the importance of each channel satisfies the following relationship:

[0055] In the formula, For the The importance of each channel, is the feature complexity of the remote sensing image, For remote sensing images The feature complexity of the channel image, For remote sensing images The contrast of the channel image (Weber contrast), is the linear normalization function, The values ​​of are 1, 2, and 3, representing the red channel image, the green channel image, and the blue channel image, respectively.

[0056] In this formula, For remote sensing images The contrast of the channel image. The larger the value, the better the remote sensing image. The stronger the contrast feature of the first channel, the The first channel is more likely to clearly reflect the difference between vegetation and other objects. In the vegetation detection task, the The more likely a channel is to provide critical information, the The greater the importance of a channel, The smaller the size, the higher the number of The weaker the contrast feature of the first channel, The more channels there are, the harder it is to highlight the difference between vegetation and other objects. In vegetation detection tasks, the The less a channel can provide key information, the The less important the channel is.

[0057] In this formula, The smaller, the The feature complexity of the first channel image is relatively low, which means that the texture features in the channel image are relatively simple and it is easier to highlight the difference between vegetation and other objects. The channel is more important for the detection of vegetation areas, and the corresponding importance On the contrary, The bigger, the The higher the feature complexity of the channel image, the more complex and diverse the texture features in the channel image are, and it is more difficult to clearly distinguish the vegetation area from other objects, thus reducing the importance of this channel for vegetation detection. Therefore, using right Correct the gamma transform method, specifically: when The more hours, The smaller the The closer it is to 0, the right Play an upward correction role, making the The greater the importance of the channel, for example, when the The contrast of each channel At a medium level, but due to Very small, that is, the features are simple, and it is easier to distinguish vegetation areas from other objects. By upward correction, the importance calculated based on the contrast of this channel increases, which is consistent with the actual situation that this channel is more valuable in vegetation detection.

[0058] when The bigger the The larger the The larger the value is, the right Play a downward correction role, making the For example, although the The contrast of each channel Higher, but due to It is also very large, that is, the features are too complex, which is not conducive to distinguishing vegetation areas from other landforms. By correcting downward, reducing the importance of this channel is also consistent with the previous logic.

[0059] In summary, this formula can accurately evaluate the importance of each channel in vegetation detection. In practical applications, different channels have different reflection capabilities for vegetation and other objects. This formula can comprehensively consider factors such as contrast and feature complexity to give each channel a reasonable degree of importance.

[0060] After determining the importance of each channel, in the subsequent process of updating the remote sensing image, different weights can be assigned according to the importance of each channel, so that the fused image can better highlight the vegetation information and improve the accuracy of vegetation detection.

[0061] S4: Extract feature vectors of remote sensing images and their channel images.

[0062] Remote sensing images are a complex type of geospatial data that contains a wealth of information about objects. The data contained in remote sensing images is high-dimensional and complex. By extracting feature vectors, data dimensionality reduction can be achieved, and feature vectors provide a quantifiable way to measure the similarities or differences between images. By calculating the distance or similarity between feature vectors, it is possible to determine whether different images belong to the same category, or to identify different types of objects in remote sensing images.

[0063] In one embodiment, the feature vector is extracted according to the following method: For remote sensing images, a histogram of the pixel values ​​of the remote sensing image is constructed. The horizontal axis of the histogram is different pixel values, and the vertical axis is the number of pixels corresponding to each pixel value. The number of pixels corresponding to all pixel values ​​in the remote sensing image constitutes the feature vector of the remote sensing image. For example, there are 10 pixels with a pixel value of 0, 11 pixels with a pixel value of 1, 12 pixels with a pixel value of 3, and 110 pixels with a pixel value of 255 in the remote sensing image (this is an abbreviation, and there will not be only these 4 pixel values ​​in the actual remote sensing image). The feature vector of the remote sensing image is .

[0064] For each channel image of the remote sensing image, the method of extracting feature vectors is the same as that of the remote sensing image. Take the red channel image as an example: construct a histogram of the pixel values ​​of the red channel image, the horizontal axis of the histogram is different pixel values, and the vertical axis is the number of pixel points corresponding to each pixel value. The number of pixel points corresponding to all pixel values ​​in the red channel image constitutes the feature vector of the red channel image.

[0065] S5: Determine the weight of each channel of the remote sensing image.

[0066] In one embodiment, the weights of each channel of the remote sensing image are calculated based on the following formula:

[0067] In this formula, For remote sensing images The weight of each channel, is the feature vector of the remote sensing image and The cosine similarity between the feature vectors of the channel images, is the natural exponential function, is the total number of channels, , For the The importance of a channel.

[0068] In this formula, The larger the value, the higher the remote sensing image The more beneficial the channel is to the segmentation of vegetation areas in remote sensing images, the The larger the weight of a channel, The smaller the size, the higher the number of The more unfavorable the channel is for the segmentation of vegetation areas in remote sensing images, the The smaller the weight of the channel.

[0069] In this formula, The size reflects the difference between the remote sensing image and the The difference in pixel value distribution between the channel images is as follows: The larger the The closer the pixel value distribution of a channel image is to the remote sensing image, the better the image segmentation task, especially the segmentation of vegetation areas. If such a channel image with a pixel value distribution similar to that of the remote sensing image itself is used for segmentation, the effect is likely to be similar to the effect of directly segmenting the grayscale image of the remote sensing image. For identifying and segmenting vegetation areas, the channel image does not bring additional, unique information or features, and the channel image does not have unique advantages. Therefore, in this case, according to the logic of the formula, the weight of the channel will be smaller. This is a reasonable setting, because if the information provided by a channel image is not much different from the remote sensing image itself, then its role in performing complex vegetation area segmentation tasks is relatively limited. Giving the channel a lower weight can avoid the channel from having too much influence on the final segmentation result.

[0070] On the contrary, when When the value of is small, this indicates that The pixel value distribution of the channel image is very different from that of the remote sensing image, which implies that the channel image may contain information different from the remote sensing image, which may be unique and valuable. For the segmentation task of vegetation area, when this channel that is very different from the remote sensing image is used for segmentation, a different segmentation effect will be produced (compared with directly segmenting the remote sensing image), which is more likely to highlight the unique features of the vegetation area in this channel. These features are not obvious in the remote sensing image, but can be clearly shown in this channel, which is very helpful for distinguishing vegetation areas from other landforms. Therefore, such a channel has more potential value for the segmentation of vegetation areas, so according to the logic of the formula, a larger weight will be assigned to this channel, so that it can play a more important role in subsequent image analysis and processing, helping to more accurately identify and segment vegetation areas.

[0071] In this formula, part, the natural exponential function plays a role of regulation and conversion here. The bigger the is a negative number with a large absolute value, so The value of is close to 0, which means that for channels similar to remote sensing images ( The larger the value, the greater its contribution to the weight calculation will be greatly suppressed, which is consistent with When it is larger, The lower the importance of a channel, the smaller the weight is set. The more hours, The absolute value of is small, making The value of is relatively large, making the The contribution of each channel in the weight calculation is amplified, which is consistent with When smaller, The higher the importance of a channel, the greater the logic of setting the weight.

[0072] In this formula, Part of it is a comprehensive consideration of the The importance of each channel and the influence of the feature similarity between the channel and the remote sensing image make the final weight calculation more comprehensive and accurate. The larger the size) and the more unique the information ( The smaller the value of this part, the larger the value of this part will be, while for channels that are neither important ( smaller) and lack of unique information ( The larger the channel, the smaller the value of this part will be.

[0073] The denominator of this formula plays a normalization role, ensuring that when allocating weights, the weights between different channels are relative, so that the weight allocation of different channels is carried out under overall consideration, and the weight will not be too large or too small to destroy the overall balance, making the weight allocation more reasonable.

[0074] In summary, this formula can reasonably assign different weights to each channel according to the characteristics of each channel and the potential contribution of each channel to vegetation region segmentation, providing a scientific basis for the subsequent use of multi-channel information for vegetation region segmentation tasks of remote sensing images. By giving greater weights to channels that are more conducive to vegetation region segmentation and giving smaller weights to channels that have little effect on vegetation region segmentation, multi-channel information can be used more effectively and the segmentation effect can be optimized.

[0075] S6: Update the pixel value of each pixel in the remote sensing image according to the weight of each channel to obtain an updated remote sensing image.

[0076] The feature differences of different objects in the updated remote sensing image will be more obvious. The channel information related to vegetation will be highlighted by the size of the weight, making the vegetation area more prominent in the updated image, which is conducive to subsequent accurate segmentation.

[0077] In one embodiment, a method of updating a remote sensing image is: The pixel values ​​of the pixels in each channel are directly weighted and summed according to the weights of each channel to update the pixel values ​​of the pixels. Specifically, for any pixel in the remote sensing image, the pixel value is updated according to the following weighted summation formula:

[0078] In this formula, is the updated pixel value of the pixel point. For this pixel point The pixel value of each channel, For remote sensing images The weight of each channel.

[0079] This operation can integrate the information of each pixel scattered in different channels to obtain a new pixel value that comprehensively reflects the characteristics of the pixel, which helps to explore the potential connection between information in different channels and more comprehensively represent the features of the objects represented by each pixel in the image.

[0080] In one embodiment, another method for updating a remote sensing image is: first, if the weight of a channel is the smallest, the weight of the channel is set to 0; then, the pixel values ​​of the pixel point in each channel are weighted and summed according to the weight of each channel to update the pixel value of the pixel point.

[0081] Specifically, for any pixel in the remote sensing image, its pixel value is updated according to the following formula:

[0082] In this formula, is the updated pixel value of the pixel point. For this pixel point The pixel value of each channel, For remote sensing images The weight of each channel. There will be a weight of 0 in this formula, which is equivalent to not considering the least important channel information when updating the pixel value.

[0083] This operation is because the channel with the smallest weight mainly reflects the details of the ground objects that are not related to the vegetation. By setting the weight of this channel to 0, it can avoid these irrelevant ground object details from interfering with the vegetation feature extraction, so that the subsequent analysis can focus more on key information.

[0084] S7: Accurately segment the vegetation area from the updated remote sensing image to complete vegetation detection.

[0085] In one embodiment, the classic image binarization algorithm OTSU is used to perform segmentation operations on the updated remote sensing image. The core principle of the OTSU algorithm is to automatically search for an optimal threshold by calculating the grayscale histogram of the image based on the principle of maximizing the inter-class variance, and divide the pixels in the image into two categories so that the variance between the two categories is maximized, thereby achieving effective image segmentation.

[0086] In the present invention, after the updated remote sensing image is segmented using the OTSU algorithm, the segmentation result is as follows: Fig. 9 As shown in , it only contains two pixel values, 0 and 1, 0 is white (bright) and 1 is black (dark). The pixel points with a pixel value of 0 in the segmentation result are determined as vegetation pixels, and the continuous area composed of all vegetation pixels is the vegetation area. In this way, the vegetation area can be clearly and accurately extracted from the remote sensing image of the complex mining area, and the vegetation detection of the mining area is completed through these operations.

[0087] In one embodiment, the vegetation detection further comprises: After extracting the vegetation area, the vegetation coverage rate of the mining area is also calculated. Specifically, the ratio of the number of pixels in the vegetation area to the number of all pixels in the updated remote sensing image is taken as the vegetation coverage rate of the mining area. This calculation result is of great value for the detection and analysis of the ecological environment in the mining area. A higher vegetation coverage rate often means that the ecological environment in the mining area is relatively good, and vegetation plays a positive role in maintaining water and soil, purifying the air, and regulating the climate; conversely, a lower vegetation coverage rate may imply that the mining area has potential risks such as ecological degradation and soil erosion. By calculating and detecting the vegetation coverage rate in a long-term and continuous manner, it can provide strong data support for the ecological restoration, environmental protection and sustainable development planning of the mining area.

[0088] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0089] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A mining area vegetation detection method based on satellite remote sensing images, characterized in that: include: Collect remote sensing images of the mining area and evaluate the feature complexity of the remote sensing images and their channel images; According to the feature complexity of the remote sensing image and its channel images, as well as the contrast of each channel image, the importance of each channel is calculated ,in, For the The importance of each channel, is the feature complexity of the remote sensing image, For remote sensing images The feature complexity of the channel image, For remote sensing images The contrast of the channel image, is a linear normalization function; Extract the feature vectors of the remote sensing image and its channel images, and calculate the weight of each channel of the remote sensing image based on the importance of each channel: ,in, For remote sensing images The weight of each channel, is the feature vector of the remote sensing image and The cosine similarity between the feature vectors of the channel images, is the natural exponential function, is the total number of channels; For the pixel points in the remote sensing image, the pixel values ​​of the pixel points are updated according to the weights of each channel to obtain an updated remote sensing image. The updated remote sensing image is segmented by the OTSU algorithm, and the vegetation area is determined according to the segmentation results to realize vegetation detection in the mining area.

2. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: The method for evaluating the feature complexity of remote sensing images and their channel images is: For any image in the remote sensing image and its channel images, the local binary pattern is used to obtain the eigenvalue of each pixel in the image, and a histogram of the eigenvalues ​​is constructed. The horizontal axis of the histogram is different eigenvalues, and the vertical axis is the number of pixels corresponding to each eigenvalue. Calculate the feature complexity of this image: ;in, is the feature complexity of the image, is the variance of the number of pixels corresponding to all eigenvalues ​​in the histogram, is the maximum number of pixels corresponding to all eigenvalues ​​in the histogram. is the number of eigenvalues ​​that appear in the image, is the total number of eigenvalues.

3. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: One method to update the pixel value of a pixel point according to the weight of each channel is: The pixel values ​​of the pixel points in each channel are directly weighted and summed according to the weights of each channel to update the pixel values ​​of the pixel points.

4. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: Another way to update the pixel value according to the weight of each channel is: First, if the weight of a channel is the smallest, set the weight of the channel to 0; Then, the pixel values ​​of the pixel points in each channel are weighted and summed according to the weights of each channel to update the pixel values ​​of the pixel points.

5. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: The method to determine the vegetation area based on the segmentation results is: The pixel points with a pixel value of 0 in the segmentation result of the updated remote sensing image are regarded as vegetation pixels, and the area formed by all vegetation pixels is regarded as the vegetation area.

6. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: The method for extracting the feature vectors of remote sensing images and their channel images is: For any image in the remote sensing image and its channel images, a histogram of the pixel values ​​of the image is constructed. The horizontal axis of the histogram is different pixel values, and the vertical axis is the number of pixel points corresponding to each pixel value. The number of pixel points corresponding to all pixel values ​​in the image constitutes the feature vector of the image.

7. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: The contrast of each channel image is any one of Weber contrast, Michelson contrast and root mean square contrast.

8. The mining area vegetation detection method based on satellite remote sensing images according to claim 1 is characterized in that: The channels are red channel, green channel and blue channel.

9. The mining area vegetation detection method based on satellite remote sensing images according to claim 5 is characterized in that: The vegetation detection further comprises: After the vegetation area is determined, the ratio of the number of pixels in the vegetation area to the number of all pixels in the updated remote sensing image is taken as the vegetation coverage rate of the mining area.

Images