An investigation method of an unmanned aerial vehicle for afforestation verification

Abstract

The present invention discloses a survey method for afforestation verification by using an unmanned aerial vehicle, which relates to the technical field of afforestation verification and solves the technical problems that a large number of images taken by unmanned aerial vehicles are directly used without being processed, which will interfere with subsequent analysis, and the obtained images are not systematically analyzed, which will cause errors in verification. The present invention enhances images through mean filtering denoising and histogram equalization, effectively improves image quality, reduces noise interference with subsequent feature extraction, matching and other links, extracts feature points based on the construction of DOG images, uses Euclidean distance to scientifically screen pre-selected images, further subdivides and matches according to the number and distribution of feature points, and sequentially splices to obtain a spliced ​​image. Based on the spliced ​​image, GIS software can be used to flexibly use manual delineation or automatic classification to measure the area of ​​the region, accurately quantify vegetation coverage using the NDVI formula, judge the vegetation growth and coverage according to the set interval, and accurately identify the vegetation type according to the image features.

Description

Technical Field

[0001] The present invention relates to the technical field of forestation verification, and specifically to an investigation method for forestation verification by an unmanned aerial vehicle (UAV). Background Art

[0002] With the development of UAV technology, it has become a trend to use UAVs equipped with cameras to obtain images for auxiliary verification. In the work of forestation verification, traditional verification methods mostly rely on manual on-site investigation, which is time-consuming, laborious, and inefficient. Moreover, it is difficult to conduct comprehensive, detailed, and accurate verification of large-scale forestation areas. Therefore, UAVs are also applied to forestation verification work.

[0003] Currently, visible light cameras are commonly used to obtain high-resolution color images for UAV photography, and some are combined with multispectral cameras to obtain multi-band images for auxiliary analysis of information such as vegetation growth vitality. In the field of image processing, there are already conventional operation technologies such as mean filtering for image denoising and histogram equalization for image enhancement. At the same time, in terms of image feature point extraction and matching, there are methods such as constructing Difference of Gaussian (DOG) images to extract feature points and comparing feature point descriptor vectors based on the Euclidean distance formula to determine the matching degree. Image stitching and subsequent analysis methods such as using GIS software, calculating vegetation indices, and identifying tree species based on the stitched images are also gradually applied to related field work. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides an investigation method for forestation verification by an unmanned aerial vehicle, which solves the problems that a large number of images taken by the UAV are directly used without being processed, which will interfere with subsequent analysis, and at the same time, the obtained images are not systematically analyzed, which will cause errors in verification.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An investigation method for forestation verification by an unmanned aerial vehicle, which specifically includes the following steps:

[0006] Step 1, obtain verification information and at the same time obtain the captured images of the UAV;

[0007] Step 2, perform preprocessing operations on the obtained captured images to obtain preprocessed captured images;

[0008] Step 3, extract the image feature points corresponding to the obtained preprocessed captured images, construct DOG images, select feature points based on the similarity values of the pixel points of the preprocessed captured images, and at the same time screen the preprocessed captured images according to the feature points to obtain preselected images;

[0009] Step 4: Analyze the obtained preselected images. By segmenting the preselected images and comprehensively selecting based on the number and distribution of feature points in the segmented preselected images, matching images are obtained. At the same time, the matching images are stitched to obtain a stitched image;

[0010] Step 5: Conduct an investigation and analysis of the verification area based on the obtained stitched image, and generate verification information by verifying three aspects of the verification area, namely the area, vegetation coverage rate, and vegetation types.

[0011] As a further solution of the present invention, the verification information in Step 1 includes the verification area and the UAV flight path.

[0012] As a further solution of the present invention, the preprocessing operations in Step 2 include image denoising and image enhancement processing.

[0013] As a further solution of the present invention, the specific method for extracting image feature points in Step 3 is as follows:

[0014] Obtain all the preprocessing captured images and label them as i, where i = 1, 2, …, j, and j represents the number label of the preprocessing captured images. Then, obtain any group of preprocessing captured images as the standard and denote it as the analysis object. Construct the DoG image of the analysis object. At the same time, obtain the corresponding pixel points of the DoG image and label them as n, where n = 1, 2, …, m, and m represents the number label of the pixel points. Calculate the pixel value of pixel point n and denote it as Hn. Then, sort the pixel values Hn of the pixel points from small to large, and select the pixel point with the smallest pixel value Hn as the feature point and denote it as u;

[0015] And so on, analyze and select the feature points of all preprocessing captured images i and denote them as iu.

[0016] As a further solution of the present invention, the specific method for screening the preprocessing captured images to obtain preselected images in Step 3 is as follows:

[0017] Obtain the preprocessing captured image corresponding to the label i = 1 and denote it as the image to be analyzed, and obtain the corresponding feature point 1u. Then, calculate and analyze it with the image to be analyzed in turn according to the order of labels i = 2, 3, …, j. The specific calculation method is as follows:

[0018] Obtain the feature point 1u of the image to be analyzed, and obtain the descriptor vector of the feature point 1u and denote it as A = (a1, a2, …, a o ), then obtain the feature point corresponding to the preprocessing captured image with the label i = 2 and denote it as 2u. Similarly, obtain the corresponding descriptor vector and denote it as B = (b1, b2, …, b o ), and at the same time, according to the Euclidean distance formula The characteristic point distance d is calculated, and then the distance d is compared with a threshold value to screen the preprocessed captured images to obtain preselected images, and the value of the threshold is set by the operator.

[0019] As a further solution of the present invention, the specific method for comparing the distance with the threshold value in step three is:

[0020] If the distance d is greater than the preset value, it means that the preprocessed captured image with the corresponding label does not match the image to be analyzed. On the contrary, if the distance d is less than the preset value, it means that the preprocessed captured image with the corresponding label matches the image to be analyzed. At the same time, the corresponding preprocessed captured image is obtained and recorded as the preselected image;

[0021] And so on, analyze the preprocessed captured images with all labels, and select the preselected images.

[0022] As a further solution of the present invention, the specific method for obtaining the spliced image in step four is:

[0023] Obtain all the preselected images and the images to be analyzed, and evenly divide the preselected images and the images to be analyzed into nine equal parts to obtain segmented images. Then, match the preselected segmented images with the segmented images to be analyzed according to the segmentation order, obtain the number and distribution of characteristic points in the preselected segmented images, and at the same time obtain the number and distribution of characteristic points in the segmented images to be analyzed. If the number of characteristic points in the preselected segmented images is the same as that in the segmented images to be analyzed, mark the preselected segmented images as matching images. On the contrary, if any group is different, do not select;

[0024] And so on, analyze according to the label to obtain the matching images, and at the same time sort the obtained matching images in order to obtain the spliced image.

[0025] As a further solution of the present invention, the specific method for generating verification information in step five is:

[0026] Obtain the spliced image, and at the same time measure the area of the corresponding verification area in the spliced image based on GIS software, and generate area information;

[0027] Secondly, use the spectral information of the image of the corresponding verification area of the spliced image, calculate NDVI to evaluate the vegetation coverage, and at the same time judge the obtained NDVI to obtain the NDVI judgment interval. The judgment interval is specifically set by the operator according to historical data, and match the NDVI of the current verification area with the judgment interval, and at the same time generate vegetation coverage information;

[0028] Then, identify the afforestation tree species according to the characteristics such as the color, texture and shape of the image of the spliced image, and at the same time generate vegetation type information.

[0029] Beneficial effects

[0030] The present invention provides a survey method for afforestation verification using a drone. Compared with the prior art, it has the following beneficial effects:

[0031] The present invention enhances images through mean filtering denoising and histogram equalization, effectively improves image quality, reduces noise interference on subsequent feature extraction, matching and other links, enhances image details and contrast, and facilitates accurate analysis. It extracts feature points based on the construction of DOG images, uses Euclidean distance to scientifically screen pre-selected images, and further subdivides and matches according to the number and distribution of feature points, splices them in sequence to form a complete and practical spliced ​​image, highly restores the scene of the verification area, and lays a solid foundation for subsequent verification and analysis. Based on the spliced ​​image, with the help of GIS software, manual delineation or automatic classification can be flexibly used to accurately measure the area of ​​the area; the NDVI formula is used to accurately quantify vegetation coverage, and the vegetation growth and coverage situation is scientifically judged according to the set interval; the vegetation type is accurately identified according to the image characteristics, and the afforestation area is comprehensively and accurately checked from the spatial scale to the vegetation growth state and vegetation composition, which greatly improves the efficiency and quality of afforestation verification work and reduces manpower input and errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a diagram of the steps of the present invention. DETAILED DESCRIPTION

[0033] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] For example, see Figure 1 The present application provides a method for investigating afforestation using a drone, which method specifically includes the following steps:

[0035] Step 1: Obtain verification information and images taken by the drone at the same time. The verification information includes the verification area and the drone route. The most common way to take drone images is to use a visible light camera, which can obtain high-resolution color images. In some cases, it will be combined with a multispectral camera. The multispectral camera can obtain images in multiple bands, such as red, green, blue, and near-infrared bands. For example, the reflectivity of vegetation in the near-infrared band is higher. By analyzing the images in this band, the growth vitality of vegetation can be evaluated.

[0036] Step 2: Perform preprocessing operations on the acquired captured images to obtain preprocessed captured images. The preprocessing operations include denoising and enhancement of the captured images. Specifically, image denoising is performed by the mean filtering method, and image enhancement is performed by histogram equalization. The above operations are all existing technologies and will not be elaborated here.

[0037] Step 3: Extract the image feature points corresponding to the obtained preprocessed captured images. Construct a Difference of Gaussian (DOG) image, and select feature points based on the similarity values of the pixel points of the preprocessed captured images. At the same time, screen the preprocessed captured images according to the feature points to obtain preselected images.

[0038] Obtain all the preprocessed captured images and label them as i, where i = 1, 2, …, j. Here, j represents the number label of the preprocessed captured images, and the labels here are random labels. Then, obtain any group of preprocessed captured images as the standard and record it as the analysis object. Construct the DoG image of the analysis object. The DoG image is obtained by subtracting Gaussian blurred images of different scales. Specifically, assume we have an image pyramid, which consists of multiple groups (octaves), and each group contains multiple layers (layers). Within a group, the standard deviation of the Gaussian kernel between adjacent layers is different. For example, the standard deviation of the Gaussian kernel of the first layer is σ, the second layer is kσ, and the third layer is k 2 σ, where k is the scale factor, usually taking the value of The DoG image is the result of subtracting adjacent two-layer Gaussian blurred images, that is, DoG(x, y, σ) = G(x, y, kσ) - G(x, y, σ), where G(x, y, σ) is the Gaussian blurred image with a standard deviation of σ at the coordinates (x, y). At the same time, obtain the pixel points corresponding to the DoG image and label them as n, where n = 1, 2, …, m. Here, m represents the number label of the pixel points, and calculate the pixel value of the pixel point n and record it as Hn. Then, sort the pixel values Hn of the pixel points from small to large, and select the pixel point with the smallest pixel value Hn as the feature point and record it as u;

[0039] And so on, analyze and select the feature points of all preprocessed captured images i and record them as iu. At the same time, obtain the preprocessed captured image corresponding to the label i = 1 and record it as the image to be analyzed, and obtain the corresponding feature point 1u. Then, calculate and analyze it in sequence with the image to be analyzed according to the order of the labels i = 2, 3, …, j. The specific calculation method is as follows:

[0040] Obtain the feature point 1u of the image to be analyzed, and obtain the descriptor vector of the feature point 1u and record it as A = (a1, a2, …, a o )). Then, obtain the feature point corresponding to the preprocessed captured image with the label i = 2 and record it as 2u. Similarly, obtain the corresponding descriptor vector and record it as B = (b1, b2, …, b o), and at the same time according to the Euclidean distance formula The distance d of the feature points is calculated, and then the distance d is compared with the threshold value, and the value of the threshold is set by the operator;

[0041] If the distance d is greater than the preset value, it means that the preprocessed captured image with the corresponding label does not match the image to be analyzed. On the contrary, if the distance d is less than the preset value, it means that the preprocessed captured image with the corresponding label matches the image to be analyzed. At the same time, the corresponding preprocessed captured image is obtained and recorded as the preselected image. By analogy, all preprocessed captured images with labels are analyzed, and the preselected images are selected;

[0042] Step 4: Analyze the obtained preselected images. By segmenting the preselected images and comprehensively selecting according to the number and distribution of feature points in the segmented preselected images, the matching images are obtained. At the same time, the matching images are stitched to obtain the stitched image.

[0043] All preselected images and images to be analyzed are obtained. Here, the images to be analyzed and the preselected images are obtained from the same group of analyzes. Specifically, they are obtained by distance comparison. The preselected images and the images to be analyzed are equally divided into nine parts to obtain segmented images, and the segmented images include preselected segmented images and images to be analyzed segmented images. Then, the preselected segmented images are matched with the images to be analyzed segmented images in the segmentation order. The matching method is as follows: obtain the number and distribution of feature points in the preselected segmented image, and at the same time obtain the number and distribution of feature points in the image to be analyzed segmented image. If the number of feature points in the preselected segmented image is the same as that in the image to be analyzed segmented image, the preselected segmented image is marked as the matching image. On the contrary, if any group is different, it is not selected;

[0044] By analogy, the matching images are obtained by analyzing according to the labels. At the same time, the obtained matching images are sorted in order to obtain the stitched image. Specifically, for example, after analyzing the image to be analyzed with label 1, the obtained matching image is the preprocessed image with label 5. Immediately, taking label 5 as the image to be analyzed for analysis, the matching image of the image to be analyzed 5 is analyzed and selected. By analogy, the final stitched image is obtained.

[0045] Step 5: Based on the obtained stitched image, investigate and analyze the verification area, and generate verification information by verifying the three aspects of the area, vegetation coverage rate, and vegetation types of the verification area.

[0046] Obtain the spliced image, and at the same time measure the area of the corresponding verification area in the spliced image based on GIS software, and generate area information. Specifically, manual delineation or automatic classification methods can be used. Manual delineation is to outline the boundaries of forestation by mouse in the software according to the characteristics such as the color and texture of the image, and the software will automatically calculate the area. Automatic classification is to use algorithms such as machine learning to separate the forestation area from the background according to the spectral characteristics of the image, and then calculate the area;

[0047] Secondly, use the spectral information of the image of the corresponding verification area of the spliced image to evaluate the vegetation coverage by calculating the vegetation index (such as the normalized difference vegetation index - NDVI). The calculation formula of NDVI is (near-infrared band reflectance - red band reflectance) / (near-infrared band reflectance + red band reflectance). The value of NDVI ranges from -1 to 1, and the higher the value, the higher the vegetation coverage. At the same time, judge the obtained NDVI, and the specific judgment method is: obtain the NDVI judgment interval, and the judgment interval is specifically set by the operator according to historical data, and match the NDVI of the current verification area with the judgment interval, and generate vegetation coverage information at the same time;

[0048] Then, identify the tree species for afforestation according to the characteristics such as the color, texture and shape of the image of the spliced image, and generate vegetation species information at the same time. Different tree species have different performances on the image. For example, the crown shapes of broad-leaved trees such as poplars are relatively broad, while the crown shapes of coniferous trees such as pine trees are relatively slender. At the same time, the leaf colors of different tree species also have differences on the image. In the growing season, the leaves of broad-leaved trees are lighter green, and the leaves of coniferous trees are darker green.

[0049] Example 2. As the second example of the present invention, it is implemented on the basis of Example 1, and the difference from Example 1 is as follows:

[0050] The method of identifying vegetation species is different. Specifically, the spectral characteristics of multi-spectral images can also be combined to identify tree species. By establishing a tree species spectral characteristic library, classification algorithms are used to determine the types and distribution of tree species for afforestation.

[0051] Example 3. As the third example of the present invention, the key lies in combining the implementation processes of Example 1 and Example 2.

[0052] At the same time, the content not described in detail in this specification belongs to the prior art well known to those skilled in the art.

[0053] The above examples are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred examples, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An investigation method for a drone to conduct verification of forestation and afforestation, characterized in that, The method specifically includes the following steps: Step 1, obtain verification information and at the same time obtain the captured images of the drone; Step 2, perform preprocessing operations on the obtained captured images to obtain preprocessed captured images; Step 3, extract the image feature points corresponding to the obtained preprocessed captured images. By constructing a DOG image and selecting feature points based on the similarity values of the pixel points of the preprocessed captured images, and at the same time screen the preprocessed captured images according to the feature points to obtain preselected images; Step 4, analyze the obtained preselected images. By segmenting the preselected images and comprehensively selecting according to the number and distribution of feature points in the segmented preselected images to obtain matching images, and at the same time splice the matching images to obtain spliced images; Step 5, conduct an investigation and analysis of the verification area based on the obtained spliced images, and generate verification information by verifying three aspects of the area, vegetation coverage rate, and vegetation types of the verification area; The specific method for extracting image feature points in Step 3 is: Obtain all preprocessed captured images and label them as i, and i = 1, 2,..., j, where j represents the number label of the preprocessed captured images. Then obtain any group of preprocessed captured images as the standard and denote it as the analysis object, construct the DoG image of the analysis object, and at the same time obtain the corresponding pixel points of the DoG image and label them as n, and n = 1, 2,..., m, where m represents the number label of the pixel points, and calculate the pixel value of pixel point n and denote it as Hn. Then sort the pixel values Hn of the pixel points from small to large, and select the pixel point with the smallest pixel value Hn as the feature point and denote it as u; And so on, analyze and select the feature points of all preprocessed captured images i and denote them as iu.

2. The investigation method of the UAV for afforestation verification according to claim 1, characterized in that, The verification information in Step 1 includes the verification area and the drone flight path.

3. The investigation method of an unmanned aerial vehicle for afforestation verification according to claim 1, characterized in that, The preprocessing operations in Step 2 include image denoising and image enhancement processing.

4. The investigation method of a drone for afforestation verification according to claim 1, characterized in that, The specific method for screening the preprocessed captured images to obtain preselected images in Step 3 is: Obtain the preprocessed captured image corresponding to the label i = 1 and denote it as the image to be analyzed, and obtain the corresponding feature point 1u. Then calculate and analyze it with the image to be analyzed in sequence according to the labels i = 2, 3,..., j, and the specific calculation method is: Obtain the feature point 1u of the image to be analyzed, and obtain the descriptor vector of the feature point 1u, denoted as A=(a1, a2, …, a o ), then obtain the feature point corresponding to the preprocessed captured image with label i = 2, denoted as 2u. Similarly, obtain the corresponding descriptor vector, denoted as B=(b1, b2, …, b o ). At the same time, according to the Euclidean distance formula Calculate the feature point distance d. Then compare the distance d with the threshold to screen the preprocessed captured image to obtain the preselected image, and the value of the threshold is set by the operator.

5. The investigation method of an unmanned aerial vehicle for afforestation verification according to claim 4, characterized in that, The specific method for comparing the distance with the threshold in Step 3 is: If the distance d is greater than the preset value, it means that the preprocessed captured image corresponding to the label does not match the image to be analyzed. On the contrary, if the distance d is less than the preset value, it means that the preprocessed captured image corresponding to the label matches the image to be analyzed, and at the same time obtain the corresponding preprocessed captured image and denote it as the preselected image; And so on, analyze all preprocessed captured images with labels and select preselected images.

6. The investigation method of a drone for afforestation verification according to claim 1, wherein The specific method for obtaining the spliced image in Step 4 is: Obtain all preselected images and images to be analyzed, and evenly divide the preselected images and images to be analyzed into nine equal parts to obtain segmented images. Then, match the preselected segmented images with the segmented images to be analyzed according to the segmentation order, obtain the number and distribution of feature points in the preselected segmented images, and at the same time obtain the number and distribution of feature points in the segmented images to be analyzed. If the number of feature points in the preselected segmented images is the same as that in the segmented images to be analyzed, mark the preselected segmented images as matching images. Otherwise, if any group is different, do not select them; And so on, analyze according to the label to obtain matching images, and at the same time sort the obtained matching images in order to obtain a spliced image.

7. The investigation method of a drone for afforestation verification according to claim 1, characterized in that, The specific method for generating verification information in step five is as follows: Obtain the spliced image, and at the same time measure the area of the corresponding verification area in the spliced image based on GIS software, and generate area information; Secondly, use the spectral information of the image corresponding to the verification area of the spliced image, evaluate the vegetation coverage by calculating NDVI, and at the same time judge the obtained NDVI to obtain the NDVI judgment interval. The judgment interval is specifically set by the operator according to historical data, and match the NDVI of the current verification area with the judgment interval, and at the same time generate vegetation coverage information; Then, identify the afforestation tree species according to the color, texture and shape characteristics of the spliced image, and at the same time generate vegetation type information.

Images