High-resolution forestry remote sensing image detail enhancement method

Abstract

The invention discloses a high-resolution forestry remote sensing image detail enhancement method, relates to the technical field of image enhancement, solves the problem that feature points existing in a contour area are not confirmed and sharpened, and locks salient points by calculating gray level change features of pixel points and neighborhoods. The optimal to-be-determined line with the minimum feature difference is screened, and it is ensured that sharpening operation is focused on the regional core contour; based on the sharpening strategy of the regional gray average value, the sharpening intensity can be adaptively adjusted, edge overexposure or detail distortion is avoided, and the method is particularly suitable for fine processing of irregular targets (such as special-shaped crowns and winding rivers) in forestry images.

Description

Technical Field

[0001] The present invention relates to the technical field of image enhancement, in particular to a method for enhancing details of high-resolution forestry remote sensing images. Background Art

[0002] High-resolution forestry remote sensing images refer to image data acquired through aerial or space remote sensing platforms with high spatial resolution (usually ground resolution of 1 meter or less, or even up to sub-meter level), specifically used for monitoring and analysis in the forestry field.

[0003] Application Publication No. CN117575953B relates to a method for detail enhancement of high-resolution forestry remote sensing images. The method first adjusts the initial reference window size based on the texture eigenvalues ​​and color eigenvalues ​​of the pixels to obtain the adjusted reference window size of each pixel. The edge value of each color channel of each pixel is obtained based on the multimodal degree value and mutation degree value of the color channel of the pixel. The enhancement coefficient of each pixel is obtained based on the edge degree value and gradient value. Finally, a forestry remote sensing detail-enhanced image is obtained. The invention improves the detail performance of forestry remote sensing images by targetedly enhancing pixels with edge performance effects.

[0004] When the forestry remote sensing image is undergoing detail enhancement processing, there are generally a large number of gradient pixel points in the corresponding area contour. If simultaneous enhancement or sharpening is performed, the numerical value will also be changed simultaneously, and the most characteristic contour points inside it cannot be highlighted. Therefore, this application is based on several gradient pixel points in the corresponding contour area. By confirming the corresponding line to be sharpened, the point with more obvious features is locked, and then based on the overall performance characteristics of the line to be sharpened, the optimal point to be sharpened is locked and sharpening is performed to achieve the optimal contour display effect. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a method for detail enhancement of high-resolution forestry remote sensing images, which solves the problem of not confirming and sharpening the feature points existing in the contour area.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for enhancing details of high-resolution forestry remote sensing images, comprising the following steps:

[0007] Grayscale processing is performed on the high-resolution forestry remote sensing images of the designated area to confirm the grayscale images;

[0008] According to the set grayscale standard range, adjust the grayscale range in the grayscale image original image, and gradually process the points to confirm the preliminary refinement of the image, including:

[0009] Confirm the grayscale value H associated with different pixels from the converted grayscale image i , where i represents different pixels, and based on the different grayscale values ​​associated with different pixels, the minimum and maximum grayscale values ​​are determined [H i min, H i max];

[0010] Then, based on the set grayscale standard range [Fmin, Fmax], the required constants a and b are determined, using: a=(H i max-H i min)÷(Fmax-Fmin) lock constant a, then based on: b=Fmin-a×H i min lock constant b;

[0011] Then use: T i =a×H i +b locks the grayscale value T i , and based on the grayscale value T associated with each different pixel point i , adjust the grayscale value of the specified pixel point, and calibrate the adjusted grayscale image as the preliminary refined image. If T i >255, then this T i Adjust to 255;

[0012] Based on the number of associations of corresponding pixel values ​​in the preliminary refined image, the proportion feature is locked. Then, based on the determined proportion feature, the characteristic coefficient associated with the corresponding pixel point is locked. Then, based on the characteristic coefficient, the equilibrium value of the corresponding pixel point is locked. The pixel point is readjusted based on the equilibrium value to obtain a contrast-enhanced image. The confirmation method is as follows:

[0013] Confirm the grayscale values ​​associated with different pixels in the preliminary refined image and mark them as T i ;

[0014] And record the corresponding gray value T i The number of times C appears in the preliminary refined image k , where k represents different grayscale values, and based on the total number G of pixels in this preliminary refined image, we use: Z k =C k ÷G determines the proportion feature Z associated with its pixel point k ;

[0015] Sort the grayscale values ​​associated with different pixels in the preliminary refined image from small to large, confirm the value sequence, and remove the repeated values ​​in the value sequence. Confirm the characteristic coefficient associated with each grayscale value in the value sequence. Based on the position of the corresponding grayscale value, sum the proportion feature of the grayscale value and the proportion features of other grayscale values ​​at the front of the grayscale value to confirm the characteristic coefficient TZ belonging to the corresponding grayscale value. k ;

[0016] Different characteristic coefficients TZ confirmed based on different gray values k , using: JH k =(255×TZ k )+0.5 locks the balance value JH associated with the corresponding gray value k ;

[0017] Based on different equalization values ​​JH associated with different gray values k , adjust the pixel values ​​corresponding to the grayscale values ​​in the preliminary refined image, and the adjusted grayscale value is JH k , obtaining a contrast-enhanced image after contrast enhancement;

[0018] The Sobel algorithm is used to identify the gradient pixels in the contrast-enhanced image and to lock the area to be sharpened based on the specific areas where the gradient pixels are located, including:

[0019] Based on the confirmed contrast-enhanced image, the Sobel algorithm is used to confirm the horizontal and vertical gradients associated with different pixels in the image and calibrate them as HX i and SX i , where i represents different pixels;

[0020] use: Lock the integrated gradient ZH associated with the corresponding pixel point i , will satisfy: ZH i >The relevant pixel points of Y1 are recorded as gradient pixel points, where Y1 is the preset value;

[0021] The area covered by several continuous gradient pixel points is recorded as the area to be sharpened, and the several areas to be sharpened in the contrast enhanced image are confirmed in turn. i >The pixels related to Y1 are not marked;

[0022] Determine the change characteristics associated with different pixels in the area to be sharpened, and based on the change characteristics associated with different pixels in the same column, lock the salient points. Then, lock the line to be sharpened based on the salient points, and select the optimal line to be determined based on the difference characteristics associated with different lines to be sharpened. The specific method is as follows:

[0023] Based on the different gray values ​​D associated with different pixels in the area to be sharpened i , where i represents different pixels, confirm the change characteristics associated with the corresponding pixels: the gray value of this pixel is recorded as D1 i , the grayscale values ​​associated with the four groups of pixels above, below, left and right are recorded as D2 i 、D3 i 、D4 i and D5 i , the pixel points at the edge of the area to be sharpened are identified from the contrast-enhanced image by four groups of pixel points around it, using: BH i =D1 i ×4+D2 i ×(-1)+D3 i ×(-1)+D4 i ×(-1)+D5 i ×(-1) determines the change feature BH associated with this pixel i ;

[0024] Starting from the top pixel point of the area to be sharpened, three pixel points are selected from top to bottom as a value column, and the salient points in a single value column are identified. The change feature BH corresponding to the salient points i The maximum value is found among three pixels, and the salient points associated with each value column are calibrated in turn. If there are no more than three pixels in the corresponding value column, the point associated with the maximum value of the change feature is directly selected as the salient point. Several adjacent salient points determined in this stage are connected to confirm the line to be sharpened in this stage;

[0025] Starting from the second pixel point in each vertical column, select three pixels from top to bottom as a value column, and use the same method as above to confirm the line to be sharpened in this stage;

[0026] Determine the first pixel of the corresponding value column from top to bottom, and stop when the last pixel of the last selected value column belongs to the bottom pixel of the area to be sharpened, and determine the line to be sharpened associated with each stage in turn;

[0027] Based on the different lines to be sharpened, the pixels above the lines to be sharpened are recorded as upper pixel points, the pixels below the lines are recorded as lower pixel points, and the change features BH associated with the upper pixel points are recorded as i Perform summation to confirm the upper feature sum, and simultaneously process several lower pixel points to confirm the lower feature sum. Use the following formula: feature value = |upper feature sum - lower feature sum| to determine the feature value associated with the corresponding line to be sharpened. Based on the different feature values ​​associated with different lines to be sharpened, select the minimum value, and record the line to be sharpened associated with the minimum value as the optimal line to be determined.

[0028] Based on the overall grayscale value mean of the corresponding area to be sharpened, and then based on the numerical difference of the optimal line to be determined, the value to be adjusted is confirmed and sharpening enhancement is performed. The confirmation method is:

[0029] Based on the different grayscale values ​​associated with different pixels in the area to be sharpened, perform mean processing on them and determine the feature mean J;

[0030] And determine the gray value D associated with different pixel points in the optimal pending line i , using: D i +|D i -J|=RH i Lock the value RH to be adjusted associated with the corresponding pixel point i .

[0031] Preferably, when confirming the grayscale image, different weights are assigned to the RGB values ​​associated with different pixels in the remote sensing image, the grayscale values ​​associated with the corresponding pixels are locked, and the remote sensing image is converted into a grayscale image based on the locked grayscale values.

[0032] The present invention provides a method for enhancing details in high-resolution forestry remote sensing images. Compared with existing technologies, it has the following advantages:

[0033] This method calculates the grayscale variation characteristics of the pixel point and its neighborhood to lock the salient point, and selects the optimal pending line with the smallest feature difference, ensuring that the sharpening operation focuses on the core contour of the region. The sharpening strategy based on the regional grayscale mean can adaptively adjust the sharpening intensity to avoid edge overexposure or detail distortion. It is particularly suitable for the refined processing of irregular targets in forestry images (such as irregular tree crowns and winding rivers).

[0034] Grayscale adjustment, contrast enhancement, gradient sharpening and feature-driven edge optimization form a complete link, which can not only enhance the global visual effect (such as the overall light and dark levels of the forest), but also highlight local details (such as the texture of a single tree); compared with a single algorithm, this solution significantly improves the visual interpretation accuracy of forestry remote sensing images through phased and regional refined processing, providing a high-quality data foundation for applications such as forest resource monitoring and pest and disease identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0037] First embodiment

[0038] See also Figure 1 , this application provides a method for detail enhancement of high-resolution forestry remote sensing images, comprising the following steps:

[0039] Step 1: Grayscale the high-resolution forestry remote sensing image of the specified area, assign different weights to the RGB values ​​associated with different pixels in the image, lock the grayscale values ​​associated with the corresponding pixels, and convert the remote sensing image into a grayscale image based on the locked grayscale values. Specifically, grayscale processing of the relevant images is relatively common in the prior art, so it will not be described in detail here. Generally, the grayscale value is confirmed based on the corresponding RGB value, and its grayscale value = R×0.114+G×0.587+B×0.299. This method can confirm the grayscale value associated with the corresponding point, perform grayscale confirmation, and lock the converted grayscale image;

[0040] Step 2: Adjust the grayscale values ​​in the grayscale image. According to the set grayscale standard range, adjust the grayscale range in the original image, and gradually confirm the preliminary refinement of the image by point processing to complete the preliminary enhancement of the grayscale image. The specific adjustment methods are as follows:

[0041] Confirm the grayscale value H associated with different pixels from the converted grayscale image i , where i represents different pixels, and based on the different grayscale values ​​associated with different pixels, the minimum and maximum grayscale values ​​are determined [H i min, H i max];

[0042] Then, based on the set grayscale standard range [Fmin, Fmax], the required constants a and b are determined, using: a=(H i max-H i min)÷(Fmax-Fmin) lock constant a, then based on: b=Fmin-a×H i min lock constant b;

[0043] Then use: T i =a×H i +b locks the grayscale value T i, and based on the grayscale value T associated with each different pixel point i , adjust the grayscale value of the specified pixel point, and calibrate the adjusted grayscale image as the preliminary refined image. If T i >255, then this T i Adjust to 255,

[0044] Specifically, the set grayscale standard range [Fmin, Fmax] is wider than the interval of the determined grayscale image, and its grayscale value range is 0-255. Its grayscale standard range cannot exceed this interval value. Based on the set grayscale standard range, the specific difference between the corresponding image and this range is determined, so as to determine the grayscale value associated with each different pixel point and make relevant adjustments.

[0045] Step 3: Contrast enhancement is performed on the determined preliminary refined image. Based on the number of associations of corresponding pixel values ​​in the preliminary refined image, the proportion feature is locked. Then, based on the determined proportion feature, the characteristic coefficient associated with the corresponding pixel point is locked. Then, based on the characteristic coefficient, the equilibrium value of the corresponding pixel point is locked. The pixel point is readjusted based on the equilibrium value to obtain a contrast-enhanced image. Specifically, when enhancing the contrast, the contrast is adjusted based on the overall characteristics of the entire image, so that the corresponding pixel point has a stronger contrast intensity in the overall image, thereby achieving a better display effect. The specific method for determining the contrast-enhanced image is as follows:

[0046] Confirm the grayscale values ​​associated with different pixels in the preliminary refined image and mark them as T i ;

[0047] And record the corresponding gray value T i The number of times C appears in the preliminary refined image k , where k represents different grayscale values, and based on the total number G of pixels in this preliminary refined image, we use: Z k =C k ÷G determines the proportion feature Z associated with its pixel point k ;

[0048] Sort the grayscale values ​​associated with different pixels in the preliminary refined image from small to large, confirm the value sequence, and remove the repeated values ​​in the value sequence. Confirm the characteristic coefficient associated with each grayscale value in the value sequence. Based on the position of the corresponding grayscale value, sum the proportion feature of the grayscale value and the proportion features of other grayscale values ​​at the front of the grayscale value to confirm the characteristic coefficient TZ belonging to the corresponding grayscale value. kFor example, the proposed value sequence is {1, 2, 3, 4, 5, 6, 24, 27, 28, ...}. The characteristic coefficient of the gray value of 1 is the proportion feature of the gray value 1, and the characteristic coefficient of the gray value of 2 is the sum of the proportion features of the gray value 1 and the gray value 2. Similarly, the characteristic coefficient of the gray value of 24 is the sum of the proportion features of the gray values ​​1, 2, 3, 4, 5, 6, and 24.

[0049] Different characteristic coefficients TZ confirmed based on different gray values k , using: JH k =(255×TZ k )+0.5 locks the balance value JH associated with the corresponding gray value k ;

[0050] Based on different equalization values ​​JH associated with different gray values k , adjust the pixel values ​​corresponding to the grayscale values ​​in the preliminary refined image, and the adjusted grayscale value is JH k , obtaining a contrast-enhanced image after contrast enhancement;

[0051] Specifically, based on the relevant process of contrast enhancement, the balance values ​​associated with different grayscale points in the corresponding image are reconfirmed and adjusted to achieve a better contrast display effect, thereby achieving a contrast-enhanced image with stronger contrast, and achieving a better contrast display effect.

[0052] Step 4: Gradient confirmation is performed on the contrast-enhanced image obtained by processing. The Sobel algorithm is used to confirm the gradient pixels existing in the contrast-enhanced image, and the area to be sharpened is locked based on the specific areas where the gradient pixels are located. The specific method of locking is as follows:

[0053] Based on the confirmed contrast-enhanced image, the Sobel algorithm is used to confirm the horizontal and vertical gradients associated with different pixels in the image and calibrate them as HX i and SX i , where i represents different pixels. Specifically, the Sobel algorithm can process the pixel values ​​associated with the pixel and the surrounding pixels, and determine the gradient data associated with the corresponding pixel in combination with the preset weight factor. The gradient data includes horizontal gradient and vertical gradient. The weight factor is selected between -2 and 2. The pixels in the outermost circle of this contrast-enhanced image do not participate in the processing process because there are no other pixels around the outermost pixel.

[0054] use: Lock the integrated gradient ZH associated with the corresponding pixel point i , will satisfy: ZH i>Y1 related pixel points are recorded as gradient pixel points, where Y1 is a preset value, and its specific value is determined by the operator based on experience. Related pixel points that do not meet the value will not be marked;

[0055] The area covered by several continuous gradient pixel points is recorded as the area to be sharpened, and the several areas to be sharpened in the contrast enhanced image are confirmed in turn. Among them, the gradient pixel points will not appear alone, because the outline of a single area is confirmed by area, so several gradient pixel points are associated with different outline areas, and the corresponding outline area is composed of several gradient pixel points. Then, based on several gradient pixel points, the corresponding area to be sharpened can be locked, and then the relevant points of the area to be sharpened are sharpened, which can make the outline of the corresponding image clearer, the texture intensity higher, and the highlighting effect of the corresponding image better.

[0056] Step 5: Based on the determined area to be sharpened, determine the change characteristics associated with different pixels in the area to be sharpened, and based on the change characteristics associated with different pixels in the same column, lock the salient points. Then, lock the lines to be sharpened based on the salient points, and select the optimal line to be determined based on the difference characteristics associated with different lines to be sharpened. The specific method of selection is as follows:

[0057] Based on the different gray values ​​D associated with different pixels in the area to be sharpened i , where i represents different pixels, confirm the change characteristics associated with the corresponding pixels: the gray value of this pixel is recorded as D1 i , the grayscale values ​​associated with the four groups of pixels above, below, left and right are recorded as D2 i 、D3 i 、D4 i and D5 i , the pixel points at the edge of the area to be sharpened are identified from the contrast-enhanced image by four groups of pixel points around it, using: BH i =D1 i ×4+D2 i ×(-1)+D3 i ×(-1)+D4 i ×(-1)+D5 i ×(-1) determines the change feature BH associated with this pixel i ;

[0058] Starting from the top pixel point of the area to be sharpened, three pixel points are selected from top to bottom as a value column, and the salient points in a single value column are identified. The change feature BH corresponding to the salient points iThe maximum value is found among three pixels, and the salient points associated with each value column are calibrated in turn. If there are no more than three pixels in the corresponding value column, the point associated with the maximum value of the change feature is directly selected as the salient point. Several adjacent salient points determined in this stage are connected to confirm the line to be sharpened in this stage;

[0059] Starting from the second pixel point in each vertical column, select three pixels from top to bottom as a value column, and use the same method as above to confirm the line to be sharpened in this stage;

[0060] Determine the first pixel of the corresponding value column from top to bottom, and stop when the last pixel of the last selected value column belongs to the bottom pixel of the area to be sharpened, and determine the line to be sharpened associated with each stage in turn;

[0061] Based on the different lines to be sharpened, the pixels above the lines to be sharpened are recorded as upper pixel points, the pixels below the lines are recorded as lower pixel points, and the change features BH associated with the upper pixel points are recorded as i Perform summation to confirm the upper feature sum, and simultaneously process several lower pixel points to confirm the lower feature sum. Use the following formula: feature value = |upper feature sum - lower feature sum| to determine the feature value associated with the corresponding line to be sharpened. Based on the different feature values ​​associated with different lines to be sharpened, select the minimum value, and record the line to be sharpened associated with the minimum value as the optimal line to be determined.

[0062] The upper and lower feature changes of this optimal pending line are closest, and also belong to the relative center position of the corresponding area to be sharpened. The relative center position here belongs to the difference between the corresponding change features, rather than the position close to the center.

[0063] Step 6: Perform sharpening and enhancement processing on the optimal lines to be determined associated with different areas to be sharpened. Based on the overall grayscale value mean of the corresponding area to be sharpened and the numerical difference of the optimal lines to be determined, the values ​​to be adjusted are confirmed and sharpening and enhancement are performed. The specific method of confirmation is:

[0064] Based on the different grayscale values ​​associated with different pixels in the area to be sharpened, perform mean processing on them and determine the feature mean J;

[0065] And determine the gray value D associated with different pixel points in the optimal pending line i , using: D i +|D i -J|=RH i Lock the value RH to be adjusted associated with the corresponding pixel point i , completing the sharpening enhancement process of the contour edge of the corresponding area, making the edge contour of the entire image more prominent.

[0066] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0067] The above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for detail enhancement of high-resolution forestry remote sensing images, characterized in that: The following steps are involved: Grayscale processing is performed on the high-resolution forestry remote sensing images of the designated area to confirm the grayscale images; According to the set grayscale standard range, adjust the grayscale range in the original grayscale image, and gradually process the points to confirm the preliminary refinement of the image; Based on the number of associations of corresponding pixel values ​​in the preliminary refined image, the proportion feature is locked, and then the characteristic coefficient associated with the corresponding pixel point is locked based on the determined proportion feature. Then, based on the characteristic coefficient, the equilibrium value of the corresponding pixel point is locked, and the pixel point is readjusted based on the equilibrium value to obtain a contrast-enhanced image. The Sobel algorithm is used to identify the gradient pixels in the contrast-enhanced image, and the area to be sharpened is locked based on the specific areas where the gradient pixels are located. Determine the change characteristics associated with different pixels in the area to be sharpened, and based on the change characteristics associated with different pixels in the same column, lock the salient points. Then, lock the line to be sharpened based on the salient points, and select the optimal line to be determined based on the difference characteristics associated with different lines to be sharpened. Based on the overall grayscale value mean of the corresponding area to be sharpened and the numerical difference of the optimal line to be determined, the value to be adjusted is confirmed and sharpening enhancement is performed.

2. The method for enhancing details of high-resolution forestry remote sensing images according to claim 1, characterized in that: When confirming the grayscale image, different weights are assigned to the RGB values ​​associated with different pixels in the remote sensing image, the grayscale values ​​associated with the corresponding pixels are locked, and the remote sensing image is converted into a grayscale image based on the locked grayscale values.

3. The method for enhancing details of high-resolution forestry remote sensing images according to claim 1, characterized in that: The confirmation process of the preliminary refined image includes: Confirm the grayscale value H associated with different pixels from the converted grayscale image i , where i represents different pixels, and based on the different grayscale values ​​associated with different pixels, the minimum and maximum grayscale values ​​are determined [H i min, H i max]; Then, based on the set grayscale standard range [Fmin, Fmax], the required constants a and b are determined, using: a=(H i max-H i min)÷(Fmax-Fmin) lock constant a, then based on: b=Fmin-a×H i min lock constant b; Then use: T i =a×H i +b locks the grayscale value T i , and based on the grayscale value T associated with each different pixel point i , adjust the grayscale value of the specified pixel point, and calibrate the adjusted grayscale image as the preliminary refined image. If T i >255, then this T i Adjust to 255.

4. The method for enhancing details of high-resolution forestry remote sensing images according to claim 1, characterized in that: The contrast-enhanced image is confirmed by: Confirm the grayscale values ​​associated with different pixels in the preliminary refined image and mark them as T i ; And record the corresponding gray value T i The number of times C appears in the preliminary refined image k , where k represents different grayscale values, and based on the total number G of pixels in this preliminary refined image, we use: Z k =C k ÷G determines the proportion feature Z associated with its pixel point k ; Sort the grayscale values ​​associated with different pixels in the preliminary refined image from small to large, confirm the value sequence, and remove the repeated values ​​in the value sequence. Confirm the characteristic coefficient associated with each grayscale value in the value sequence. Based on the position of the corresponding grayscale value, sum the proportion feature of the grayscale value and the proportion features of other grayscale values ​​at the front of the grayscale value to confirm the characteristic coefficient TZ belonging to the corresponding grayscale value. k ; Different characteristic coefficients TZ confirmed based on different gray values k , using: JH k =(255×TZ k )+0.5 locks the balance value JH associated with the corresponding gray value k ; Based on different equalization values ​​JH associated with different gray values k , adjust the pixel values ​​corresponding to the grayscale values ​​in the preliminary refined image, and the adjusted grayscale value is JH k , and obtain the contrast-enhanced image after contrast enhancement.

5. The method for enhancing details of high-resolution forestry remote sensing images according to claim 1, characterized in that: The locking method of the area to be sharpened includes: Based on the confirmed contrast-enhanced image, the Sobel algorithm is used to confirm the horizontal and vertical gradients associated with different pixels in the image and calibrate them as HX i and SX i , where i represents different pixels; use: Lock the integrated gradient ZH associated with the corresponding pixel point i , will satisfy: ZH i >The relevant pixel points of Y1 are recorded as gradient pixel points, where Y1 is the preset value; The area covered by several continuous gradient pixel points is recorded as the area to be sharpened, and the several areas to be sharpened in the contrast enhanced image are confirmed in turn.

6. The method for enhancing details of high-resolution forestry remote sensing images according to claim 5, characterized in that: Not satisfied with ZH i >The pixels related to Y1 are not marked.

7. The method for enhancing details of high-resolution forestry remote sensing images according to claim 1, characterized in that: The specific method for selecting the optimal undetermined line is: Based on the different gray values ​​D associated with different pixels in the area to be sharpened i , where i represents different pixels, confirm the change characteristics associated with the corresponding pixels: the gray value of this pixel is recorded as D1 i , the grayscale values ​​associated with the four groups of pixels above, below, left and right are recorded as D2 i 、D3 i 、D4 i and D5 i , the pixel points at the edge of the area to be sharpened are identified from the contrast-enhanced image by four groups of pixel points around it, using: BH i =D1 i ×4+D2 i ×(-1)+D3 i ×(-1)+D4 i ×(-1)+D5 i ×(-1) determines the change feature BH associated with this pixel i ; Starting from the top pixel point of the area to be sharpened, three pixel points are selected from top to bottom as a value column, and the salient points in a single value column are identified. The change feature BH corresponding to the salient points i The maximum value is found among three pixels, and the salient points associated with each value column are calibrated in turn. If there are no more than three pixels in the corresponding value column, the point associated with the maximum value of the change feature is directly selected as the salient point. Several adjacent salient points determined in this stage are connected to confirm the line to be sharpened in this stage; Starting from the second pixel point in each vertical column, select three pixels from top to bottom as a value column, and use the same method as above to confirm the line to be sharpened in this stage; Determine the first pixel of the corresponding value column from top to bottom, and stop when the last pixel of the last selected value column belongs to the bottom pixel of the area to be sharpened, and determine the line to be sharpened associated with each stage in turn; Based on the different lines to be sharpened, the pixels above the lines to be sharpened are recorded as upper pixel points, the pixels below the lines are recorded as lower pixel points, and the change features BH associated with the upper pixel points are recorded as i Perform summation to confirm the upper feature sum, and simultaneously process several lower pixel points to confirm the lower feature sum. Use: eigenvalue = |upper feature sum - lower feature sum| to determine the eigenvalue associated with the corresponding line to be sharpened, and based on the different eigenvalues ​​associated with different lines to be sharpened, select the minimum value, and record the line to be sharpened associated with the minimum value as the optimal line to be determined.

8. The method for enhancing details of high-resolution forestry remote sensing images according to claim 7, characterized in that: The method for confirming the value to be adjusted is: Based on the different grayscale values ​​associated with different pixels in the area to be sharpened, perform mean processing on them and determine the feature mean J; And determine the gray value D associated with different pixel points in the optimal pending line i , using: D i +|D i -J|=RH i Lock the value RH to be adjusted associated with the corresponding pixel point i .

Images