CN117541654A - Detail enhancement method for high-resolution remote sensing image - Google Patents
Detail enhancement method for high-resolution remote sensing image Download PDFInfo
- Publication number
- CN117541654A CN117541654A CN202410027006.1A CN202410027006A CN117541654A CN 117541654 A CN117541654 A CN 117541654A CN 202410027006 A CN202410027006 A CN 202410027006A CN 117541654 A CN117541654 A CN 117541654A
- Authority
- CN
- China
- Prior art keywords
- image
- remote sensing
- layer
- sensing image
- pyramid
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000012512 characterization method Methods 0.000 claims abstract description 52
- 238000005070 sampling Methods 0.000 claims abstract description 28
- 238000010606 normalization Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims 8
- 230000000694 effects Effects 0.000 abstract description 14
- 238000004458 analytical method Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000009466 transformation Effects 0.000 description 5
- 230000014759 maintenance of location Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
Abstract
The invention relates to the technical field of image enhancement, in particular to a detail enhancement method of a high-resolution remote sensing image. The method comprises the following steps: acquiring a remote sensing image of a farmland; obtaining complexity according to gray level distribution of pixel points in the remote sensing image, and setting a variance scale and a sampling step length corresponding to each layer in the pyramid image; downsampling the remote sensing image by using a variance scale, a sampling step length and different gamma parameters to obtain a corresponding downsampled image; according to gray distribution of pixel points, the number of characteristic points and position distribution in a downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer, detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image is obtained, optimal gamma parameters corresponding to each layer in the pyramid image are determined based on the detail characterization degree, and then the enhanced image is obtained. The invention improves the enhancement effect of farmland remote sensing images.
Description
Technical Field
The invention relates to the technical field of image enhancement, in particular to a detail enhancement method of a high-resolution remote sensing image.
Background
The high-resolution remote sensing image has wide application in various fields such as geographic information systems, agriculture, city planning, environment monitoring and the like. In the agricultural field, because the farmland remote sensing image is easily interfered by image acquisition equipment or external environment in the acquisition process, the acquired farmland remote sensing image has poor quality, so that the subsequent analysis result of crops in the farmland is affected, and in order to better identify the detailed information in the remote sensing image, an image enhancement technology is widely used. The traditional image enhancement method, such as histogram equalization, logarithmic transformation and power law transformation, can improve the contrast of the farmland remote sensing image, but can cause partial detail loss or increase of image noise in the farmland remote sensing image, so that the enhancement effect is poor, and the subsequent analysis result is influenced.
Disclosure of Invention
In order to solve the problem of poor enhancement effect of the existing method when the farmland remote sensing image is enhanced, the invention aims to provide a detail enhancement method of a high-resolution remote sensing image, which adopts the following technical scheme:
the invention provides a detail enhancement method of a high-resolution remote sensing image, which comprises the following steps:
acquiring a remote sensing image of a farmland;
obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region in the remote sensing image; setting a variance scale and a sampling step length corresponding to each layer of the remote sensing image in the pyramid image based on the complexity; performing downsampling processing on the remote sensing image by using the variance scale, the sampling step length and different gamma parameters to obtain a downsampled image under each gamma parameter corresponding to each layer; detecting the feature points of all the downsampled images to obtain corresponding feature points;
obtaining the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image according to the gray level distribution, the number of the characteristic points and the position distribution of the pixel points in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer; determining an optimal gamma parameter corresponding to each layer of the remote sensing image in the pyramid image based on the detail characterization degree;
and obtaining an enhanced image based on the optimal gamma parameter.
Preferably, the obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region in the remote sensing image includes:
for an x-th region in the remote sensing image: calculating gray difference and Euclidean distance of every two pixel points in the x-th area; the pixel point pair formed by two pixel points with gray level difference smaller than a preset difference threshold value and corresponding Euclidean distance smaller than a preset distance threshold value in the x-th area is marked as a target pixel point pair, and the probability of the target pixel point pair in the remote sensing image is calculated; substituting the probability into a calculation formula of entropy to obtain a space entropy corresponding to the x-th region;
and determining the normalization result of the variances of the spatial entropy corresponding to all the areas in the remote sensing image as the complexity of the remote sensing image.
Preferably, according to the gray distribution of pixel points, the number of feature points and the position distribution in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer, the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image is obtained, including:
for the jth gamma parameter corresponding to the kth layer in the pyramid image of the xth region in the remote sensing image:
performing convex hull detection on feature points in a downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image to obtain corresponding connected domains; the distance between each feature point in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the center point of the connected domain is recorded as a first distance corresponding to each feature point;
calculating the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image according to the space entropy corresponding to the xth region, the variance scale corresponding to the kth layer in the pyramid image, the number of characteristic points in the downsampled image of the xth region under each gamma parameter corresponding to the kth layer in the pyramid image and the first distance;
and obtaining the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image based on the feature characterization degree.
Preferably, the following formula is adopted to calculate the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image:
；
wherein,representing the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image,/o>Representing the variance scale corresponding to the kth layer in the pyramid image,/->Represents the spatial entropy corresponding to the xth region, e represents a natural constant, J represents the number of types of gamma parameters corresponding to each layer in the pyramid image, < + >>Representing the number of feature points in the downsampled image of the xth region at the jth gamma parameter corresponding to the kth layer in the pyramid image,representing the kth layer in the pyramid image other than the jth gamma parameter +.>The number of feature points in the downsampled image under the gamma parameters +.>Representing the mean value of the first distances corresponding to all feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image, < >>Representing the kth layer in the pyramid image other than the jth gamma parameter +.>Average value of first distances corresponding to all feature points in downsampled image under gamma parameters +.>Is a sine function +.>Is the circumference ratio.
Preferably, obtaining the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image based on the feature characterization degree includes:
and carrying out normalization processing on the characteristic characterization degree, and taking a normalization result as the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image.
Preferably, the determining, based on the detail characterization degree, the optimal gamma parameter corresponding to each layer of the remote sensing image in the pyramid image includes:
for the k-th layer of the pyramid image:
the product of the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the variance scale corresponding to the kth layer is recorded as the characteristic index of the jth gamma parameter corresponding to the kth layer in the pyramid image;
taking the sum of characteristic indexes of the jth gamma parameter corresponding to the kth layer in the pyramid image as the characteristic value of the jth gamma parameter corresponding to the kth layer;
and determining the gamma parameter corresponding to the maximum characteristic value as the optimal gamma parameter corresponding to the kth layer of the remote sensing image in the pyramid image.
Preferably, the acquiring of each region in the remote sensing image includes:
and dividing the remote sensing image to obtain at least two areas, wherein the sizes of all the areas are equal.
Preferably, the obtaining the enhanced image based on the optimal gamma parameter includes:
taking a downsampled image under the optimal gamma parameters corresponding to each layer in the pyramid image as a target image of each layer;
and obtaining an enhanced image based on the target images of all layers in the pyramid image.
Preferably, the setting the variance scale and the sampling step length corresponding to each layer of the remote sensing image in the pyramid image based on the complexity includes:
determining a variance scale corresponding to each layer of the remote sensing image in the pyramid image based on the complexity, wherein the complexity and the variance scale are in a negative correlation relationship, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is equal, and the variance scale corresponding to the next layer in the two adjacent layers is larger than the variance scale corresponding to the previous layer;
and determining a sampling step length based on the complexity, wherein the complexity and the sampling step length are in a negative correlation.
Preferably, the SITF operator is adopted to detect the feature points of all the downsampled images to obtain the corresponding feature points.
The invention has at least the following beneficial effects:
in order to solve the problem that details of a farmland remote sensing image are lost or image noise is increased after enhancement, a pyramid image denoising algorithm is selected to extract collected remote sensing images of the farmland in different scales, and then the images of the farmland in different scales are enhanced, and the variance scale, the sampling step length and gamma parameters influence the image downsampling result when the remote sensing images of the farmland are subjected to downsampling treatment in different scales; likewise, the more detail features, the smaller the sampling step should be, the more the corresponding layers, the more the different detail features need to be characterized at different layers of the pyramid; the invention combines the complexity to determine the corresponding variance scale and sampling step length of each layer of the remote sensing image in the pyramid image, and uses different gamma parameters to perform downsampling processing on the remote sensing image of the farmland, and adaptively determines the optimal gamma parameters corresponding to each layer of the pyramid image by analyzing the gray level distribution of pixel points, the number of characteristic points and the position distribution of each region in the downsampling image of each layer of the remote sensing image under each gamma parameter corresponding to each layer, thereby obtaining the enhanced image, enabling the images with different scales to amplify the corresponding characteristics, eliminating the problems of detail characteristic loss or noise increase and improving the enhancement effect of the remote sensing image of the farmland.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a detail enhancement method of a high-resolution remote sensing image according to an embodiment of the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the present invention to achieve the preset purpose, the following detailed description of a high-resolution remote sensing image detail enhancement method according to the present invention is given with reference to the accompanying drawings and the preferred embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the high-resolution remote sensing image detail enhancement method provided by the invention with reference to the accompanying drawings.
An embodiment of a detail enhancement method for a high-resolution remote sensing image comprises the following steps:
the embodiment provides a high-resolution remote sensing image detail enhancement method, as shown in fig. 1, which includes the following steps:
and S1, acquiring a remote sensing image of a farmland.
The specific scene aimed at by this embodiment is: when analyzing the growth condition of crops in a farmland, firstly, remote sensing images of the farmland are required to be acquired, but in consideration of the fact that the remote sensing images are easily influenced by image acquisition equipment or external factors in the acquisition process, the quality of the acquired remote sensing images is poor, and the subsequent analysis results are influenced, so that the embodiment processes the acquired remote sensing images by combining the characteristics of the acquired remote sensing images, improves the quality of the remote sensing images, and further can effectively improve the accuracy of the analysis results of the growth condition of the crops.
In the embodiment, firstly, a remote sensing image of a farmland to be analyzed is collected, histogram equalization processing is carried out on the collected remote sensing image of the farmland to be analyzed, the purpose is to amplify contrast of detail features, and the remote sensing image obtained after processing is recorded as the remote sensing image of the farmland. The histogram equalization process is the prior art, and will not be described in detail here.
Step S2, obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region in the remote sensing image; setting a variance scale and a sampling step length corresponding to each layer of the remote sensing image in the pyramid image based on the complexity; performing downsampling processing on the remote sensing image by using the variance scale, the sampling step length and different gamma parameters to obtain a downsampled image under each gamma parameter corresponding to each layer; and detecting the feature points of all the downsampled images to obtain corresponding feature points.
The part of the general remote sensing image which can represent the characteristics most is edge detail, and the clearer the edge detail is, the better the effect of the remote sensing image is correspondingly. In order to preserve this detail, it is necessary to enhance a significant portion thereof while also avoiding the effects of noise. The method comprises the steps of obtaining a detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image, determining a characteristic value of each gamma parameter corresponding to each layer in combination with a specific filtering scale after obtaining the detail characterization degree, and measuring an image enhancement effect under the corresponding enhancement parameter by the characteristic value so as to obtain the optimal enhanced image.
The selection of the image parameters of the golden sub-tower influences the selection of the gamma transformation enhancement parameters in the subsequent processing process, and the obtained scale is different due to different variances of Gaussian filtering, and the obtained layer numbers are different due to different sampling step sizes, so that proper parameters are required to be determined, and the processing effect of the subsequent images is improved. For the determination of parameters, the image features, mainly the complexity of the image, are combined, the more the image is complex, the more detail features are included, the more the detail features are, the variation of the Gaussian filter variance among different scales is reduced, the larger the variation is, and the more parts are filtered; likewise, the more detail features, the smaller the sampling step should be, the more the corresponding number of layers, the more the different detail features need to be characterized at different layers of the pyramid. In this embodiment, the remote sensing image of the farmland is first divided into a preset number of areas with the same size, in this embodiment, the preset number is 20, that is, the remote sensing image of the farmland is divided into 20 areas with the same size, and in a specific application, an implementer can set according to specific situations. And obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region in the remote sensing image of the farmland.
Specifically, for the x-th region in the remote sensing image of the farmland: calculating gray difference and Euclidean distance of every two pixel points in the x-th area; the pixel point pair formed by two pixel points with gray level difference smaller than a preset difference threshold value and corresponding Euclidean distance smaller than a preset distance threshold value in the x-th area is marked as a target pixel point pair, and the probability of the target pixel point pair in the remote sensing image is calculated; substituting the probability into a calculation formula of entropy to obtain the spatial entropy corresponding to the xth region. In this embodiment, the preset difference threshold is 20, and the preset distance threshold is 3, and in a specific application, the practitioner may set according to a specific situation. By adopting the method, the spatial entropy corresponding to each region in the remote sensing image of the farmland can be obtained. And determining the normalization result of the variances of the spatial entropy corresponding to all the areas in the remote sensing image as the complexity of the remote sensing image. In this embodiment, the variance of the spatial entropy is normalized by using a maximum-minimum normalization method, and as other embodiments, the variance may be normalized by using other normalization methods. The maximum and minimum normalization method is the prior art, and will not be described in detail here.
The higher the complexity of the remote sensing image of the farmland is, the smaller the corresponding Gaussian filter variance variation is when the remote sensing image is subjected to filter processing, and the smaller the sampling step length is. Based on the above, the embodiment determines the variance scale corresponding to each layer of the remote sensing image in the pyramid image based on the complexity, wherein the complexity and the variance scale are in a negative correlation relationship, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is equal, and the variance scale corresponding to the next layer in the two adjacent layers is larger than the variance scale corresponding to the previous layer; in this embodiment, when the complexity is greater than 0.9, the difference between the variance scales corresponding to two adjacent layers in the pyramid image is 0.1; when the complexity is more than 0.8 and less than or equal to 0.9, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.2; when the complexity is more than 0.7 and less than or equal to 0.8, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.3; when the complexity is more than 0.6 and less than or equal to 0.7, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.4; when the complexity is more than 0.5 and less than or equal to 0.6, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.5; when the complexity is more than 0.4 and less than or equal to 0.5, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.6; when the complexity is more than 0.3 and less than or equal to 0.4, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.7; when the complexity is more than 0.2 and less than or equal to 0.3, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.8; when the complexity is more than 0.1 and less than or equal to 0.2, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 0.9; when the complexity is less than or equal to 0.1, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is 1. Meanwhile, determining a sampling step length based on the complexity, wherein the complexity and the sampling step length are in a negative correlation; in this embodiment, when the complexity is greater than 0.4, the sampling step is made to be 1; when the complexity is less than or equal to 0.4, the sampling step length is 2; in a specific application, the practitioner may set up according to the specific circumstances.
The remote sensing image of the farmland is subjected to downsampling processing by utilizing the variance scale corresponding to each layer, the sampling step length and different gamma parameters to obtain a corresponding pyramid image, wherein each layer in the pyramid image corresponds to a downsampled image under each gamma parameter, namely, each layer in the pyramid image corresponds to a plurality of downsampled images, gamma parameters of the corresponding downsampled images of any layer in the pyramid image are different, the gamma parameters are respectively set to be 0.05, 0.1, 0.15, 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95 and 1, namely, each layer in the pyramid image corresponds to a downsampled image under the 20 gamma parameters, and each layer corresponds to 20 downsampled images. In a particular application, the practitioner can set the gamma parameters on a case-by-case basis.
In the embodiment, the SITF operator is adopted to detect the feature points of all the downsampled images in the pyramid image to obtain the corresponding feature points. The SITF operator is prior art and will not be described in detail here.
Step S3, according to the gray distribution of pixel points, the number of characteristic points and the position distribution of the pixel points in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer, obtaining the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image; and determining the optimal gamma parameters corresponding to each layer of the remote sensing image in the pyramid image based on the detail characterization degree.
The embodiment has obtained the downsampled image under each gamma parameter corresponding to each layer and the feature points in each downsampled image, and then determines the detail characterization degree of each region in the remote sensing image of the farmland under each gamma parameter corresponding to each layer in the pyramid image according to the gray distribution of the pixel points in the downsampled image, the number and the position distribution of the feature points of each region in the remote sensing image of the farmland under each gamma parameter corresponding to each layer.
The acquisition of the degree of detail characterization is based on different enhancement parameters at different scales, the first difference being to highlight features representing different edges at different scales, for example: on a scale image with smaller Gaussian filtering variance, the feature retention effect on the original image is better, and then the scale can highlight some more detailed features; on the contrary, on the scale image with larger Gaussian filtering variance, the feature retention effect on the original image is poor, only some more obvious high-frequency features can be retained at the moment, and the high-frequency features correspond to the edges with the most obvious features, so that the scale can highlight some more obvious features. Different nonlinear enhancement parameters are needed under different scales, and when the gamma parameter is between 0 and 1, the gray values of pixel points in the original image can be amplified under different scales, namely the contrast ratio is increased, so that the expressive force of the features is increased. When the gamma parameter is smaller than 1, the gamma transformation can improve the contrast of the low gray level part of the image, so that the details of the dark area are more obvious, the image enhancement for displaying the details of the dark area is facilitated, and some high gray level details can be lost; when the gamma parameter is greater than 1, the gamma transformation increases the contrast of the high gray level portion of the image, making the bright details more visible, facilitating the display of the bright details or reducing the overexposure phenomenon of the image, but possibly losing some of the low gray level details. Based on the above, in this embodiment, the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image is obtained according to the gray distribution of the pixel points, the number of the feature points and the position distribution in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer.
For the jth gamma parameter corresponding to the kth layer in the pyramid image of the xth region in the remote sensing image:
performing convex hull detection on feature points in a downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image to obtain corresponding connected domains; the distance between each feature point in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the center point of the connected domain is recorded as a first distance corresponding to each feature point; calculating the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image according to the space entropy corresponding to the xth region, the variance scale corresponding to the kth layer in the pyramid image, the number of characteristic points in the downsampled image of the xth region under each gamma parameter corresponding to the kth layer in the pyramid image and the first distance; and obtaining the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image based on the feature characterization degree. The specific calculation formula of the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image is as follows:
；
wherein,representing the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image,/o>Representing the variance scale corresponding to the kth layer in the pyramid image,/->Represents the spatial entropy corresponding to the xth region, e represents a natural constant, J represents the number of types of gamma parameters corresponding to each layer in the pyramid image, < + >>Representing the number of feature points in the downsampled image of the xth region at the jth gamma parameter corresponding to the kth layer in the pyramid image,representing the kth layer in the pyramid image other than the jth gamma parameter +.>Personal gammaNumber of feature points in downsampled images under horse parameters, +.>Representing the mean value of the first distances corresponding to all feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image, < >>Representing the kth layer in the pyramid image other than the jth gamma parameter +.>Average value of first distances corresponding to all feature points in downsampled image under gamma parameters +.>Is a sine function +.>Is the circumference ratio.
And carrying out normalization processing on the characteristic characterization degree, and taking the normalization result as the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image. In this embodiment, the feature characterization degrees are normalized by using a maximum and minimum normalization method, which is the prior art, and will not be described in detail here.
The higher the complexity of the remote sensing image of the farmland is, the more the corresponding details are, and the higher the corresponding detail characterization degree is. The larger the scale variance, the higher the corresponding filter strength and the lower the level of detail retention. The closer the gamma parameter is to 1, the less obvious the corresponding enhancement effect; the less the gamma parameter is close to 1, the darker detail can be amplified, namely the detail characteristic is amplified, so that the gamma parameter and the detail characterization degree are in negative correlation, and the embodiment uses a trigonometric function to quantify the relationship and stagger the angular frequency toThe purpose is to have its argument at 0 to1, and the dependent variable is also between 0 and 1. In the present embodiment, the number of feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the other jth gamma parameter are used ≡>The ratio of the number of the feature points in the downsampled image under the gamma parameters characterizes the change quantity of the feature points, so as to measure the amplified effect after enhancement, if the feature points become more, the number of the feature points under the enhanced parameters becomes more, and the value after summation becomes larger. The reflection of the region with the building or the tree in the remote sensing image of the farmland is obvious, and the feature points obtained by establishing the pyramid image are arranged in the DOG space to form a geometric image with a shape more similar to that of a regular geometric image. The larger the average value of the first distances corresponding to all the feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image is, and the larger the average value of the first distances corresponding to all the feature points in the downsampled image under other gamma parameters except the jth gamma parameter is, the larger the corresponding detail characterization degree is.
By adopting the method, the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image can be obtained, and then the embodiment determines the optimal gamma parameter corresponding to each layer in the pyramid image based on the detail characterization degree.
Specifically, for the k-th layer of the pyramid image: the product of the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the variance scale corresponding to the kth layer is recorded as the characteristic index of the jth gamma parameter corresponding to the kth layer in the pyramid image; taking the sum of characteristic indexes of the jth gamma parameter corresponding to the kth layer in the pyramid image as the characteristic value of the jth gamma parameter corresponding to the kth layer; and determining the gamma parameter corresponding to the maximum characteristic value as the optimal gamma parameter corresponding to the kth layer of the remote sensing image in the pyramid image. The higher the layer number of the pyramid image is, the larger the variance scale is, and the lower the corresponding detail characterization degree is, so that the weight of the pyramid image is correspondingly amplified, and the feature value is obtained by taking the weight as the weight.
By adopting the method, the optimal gamma parameters corresponding to each layer of the remote sensing image in the pyramid image can be obtained.
And S4, obtaining an enhanced image based on the optimal gamma parameter.
In the embodiment, in step S3, an optimal gamma parameter corresponding to each layer in a pyramid image of a remote sensing image is obtained, and then a downsampled image under the optimal gamma parameter corresponding to each layer in the pyramid image is used as a target image of each layer; the enhanced image is obtained based on the target images of all layers in the pyramid image, namely, the image reconstruction is carried out based on all the target images, the filtered image is obtained, and the interference of noise is eliminated by the filtered image, so that the accuracy of the subsequent analysis of the growth condition of crops based on the enhanced image is higher.
The method provided by the embodiment completes the enhancement processing of the remote sensing image of the farmland.
In order to solve the problem of loss of details or increase of image noise of the enhanced farmland remote sensing image, the method selects to use a pyramid image denoising algorithm to extract the collected farmland remote sensing image with different scales, further enhances the image with different scales, considers that when the farmland remote sensing image is subjected to downsampling processing with different scales, the variance scale, the sampling step length and the gamma parameter influence the downsampling result of the image, judges the complexity of the remote sensing image according to the gray level distribution condition of pixel points in each region of the farmland remote sensing image, and indicates that the greater the complexity of the remote sensing image is, the more the growth condition of crops in the farmland is, the more the detail information contained in the farmland remote sensing image is, the more the variation of Gaussian filtering variance among different scales is reduced, and the greater the variation is, and the more the filtered part is; likewise, the more detail features, the smaller the sampling step should be, the more the corresponding layers, the more the different detail features need to be characterized at different layers of the pyramid; therefore, the embodiment combines the complexity to determine the corresponding variance scale and sampling step length of each layer of the remote sensing image in the pyramid image, performs downsampling processing on the remote sensing image of the farmland by utilizing different gamma parameters, and adaptively determines the optimal gamma parameters corresponding to each layer of the pyramid image by analyzing the gray distribution of pixel points, the number of characteristic points and the position distribution of each region in the downsampled image of each region in each layer of the remote sensing image under each gamma parameter, thereby obtaining the enhanced image, enabling the images with different scales to amplify the corresponding characteristics, eliminating the problems of loss of detail characteristics or increase of noise, and improving the enhancement effect of the remote sensing image of the farmland.
It should be noted that: the foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. The detail enhancement method of the high-resolution remote sensing image is characterized by comprising the following steps of:
acquiring a remote sensing image of a farmland;
obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region in the remote sensing image; setting a variance scale and a sampling step length corresponding to each layer of the remote sensing image in the pyramid image based on the complexity; performing downsampling processing on the remote sensing image by using the variance scale, the sampling step length and different gamma parameters to obtain a downsampled image under each gamma parameter corresponding to each layer; detecting the feature points of all the downsampled images to obtain corresponding feature points;
obtaining the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image according to the gray level distribution, the number of the characteristic points and the position distribution of the pixel points in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer; determining an optimal gamma parameter corresponding to each layer of the remote sensing image in the pyramid image based on the detail characterization degree;
and obtaining an enhanced image based on the optimal gamma parameter.
2. The method for enhancing details of a high resolution remote sensing image according to claim 1, wherein obtaining the complexity of the remote sensing image according to the gray level distribution of the pixel points in each region of the remote sensing image comprises:
for an x-th region in the remote sensing image: calculating gray difference and Euclidean distance of every two pixel points in the x-th area; the pixel point pair formed by two pixel points with gray level difference smaller than a preset difference threshold value and corresponding Euclidean distance smaller than a preset distance threshold value in the x-th area is marked as a target pixel point pair, and the probability of the target pixel point pair in the remote sensing image is calculated; substituting the probability into a calculation formula of entropy to obtain a space entropy corresponding to the x-th region;
and determining the normalization result of the variances of the spatial entropy corresponding to all the areas in the remote sensing image as the complexity of the remote sensing image.
3. The method for enhancing details of a high resolution remote sensing image according to claim 2, wherein obtaining the detail characterization degree of each region in the remote sensing image under each gamma parameter corresponding to each layer in the pyramid image according to the gray level distribution, the number of feature points and the position distribution of the pixel points in the downsampled image of each region in the remote sensing image under each gamma parameter corresponding to each layer comprises:
for the jth gamma parameter corresponding to the kth layer in the pyramid image of the xth region in the remote sensing image:
performing convex hull detection on feature points in a downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image to obtain corresponding connected domains; the distance between each feature point in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the center point of the connected domain is recorded as a first distance corresponding to each feature point;
calculating the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image according to the space entropy corresponding to the xth region, the variance scale corresponding to the kth layer in the pyramid image, the number of characteristic points in the downsampled image of the xth region under each gamma parameter corresponding to the kth layer in the pyramid image and the first distance;
and obtaining the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image based on the feature characterization degree.
4. The method for enhancing details of a high-resolution remote sensing image according to claim 3, wherein the following formula is adopted to calculate the feature characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image:
；
wherein,representing the characteristic characterization degree of the xth region in the remote sensing image under the jth gamma parameter corresponding to the kth layer in the pyramid image,/o>Representing the variance scale corresponding to the kth layer in the pyramid image,/->Represents the spatial entropy corresponding to the xth region, e represents a natural constant, J represents the number of types of gamma parameters corresponding to each layer in the pyramid image, < + >>Representing the number of feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image, +.>Representing the kth layer in the pyramid image other than the jth gamma parameter +.>The number of feature points in the downsampled image under the gamma parameters +.>Representing the mean value of the first distances corresponding to all feature points in the downsampled image of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image, < >>Representing the kth layer in the pyramid image other than the jth gamma parameter +.>Average value of first distances corresponding to all feature points in downsampled image under gamma parameters +.>Is a sine function +.>Is the circumference ratio.
5. A method for enhancing details of a high resolution remote sensing image according to claim 3, wherein obtaining the detail characterization degree of the x-th region under the j-th gamma parameter corresponding to the k-th layer in the pyramid image based on the feature characterization degree comprises:
and carrying out normalization processing on the characteristic characterization degree, and taking a normalization result as the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image.
6. The method for enhancing details of a high-resolution remote sensing image according to claim 1, wherein determining an optimal gamma parameter corresponding to each layer of the remote sensing image in the pyramid image based on the detail characterizations comprises:
for the k-th layer of the pyramid image:
the product of the detail characterization degree of the xth region under the jth gamma parameter corresponding to the kth layer in the pyramid image and the variance scale corresponding to the kth layer is recorded as the characteristic index of the jth gamma parameter corresponding to the kth layer in the pyramid image;
taking the sum of characteristic indexes of the jth gamma parameter corresponding to the kth layer in the pyramid image as the characteristic value of the jth gamma parameter corresponding to the kth layer;
and determining the gamma parameter corresponding to the maximum characteristic value as the optimal gamma parameter corresponding to the kth layer of the remote sensing image in the pyramid image.
7. The method for enhancing details of a high resolution remote sensing image according to claim 1, wherein the obtaining of each region in the remote sensing image comprises:
and dividing the remote sensing image to obtain at least two areas, wherein the sizes of all the areas are equal.
8. The method of claim 1, wherein obtaining the enhanced image based on the optimal gamma parameters comprises:
taking a downsampled image under the optimal gamma parameters corresponding to each layer in the pyramid image as a target image of each layer;
and obtaining an enhanced image based on the target images of all layers in the pyramid image.
9. The method for enhancing details of a high-resolution remote sensing image according to claim 1, wherein the setting the variance scale and the sampling step size corresponding to each layer of the remote sensing image in the pyramid image based on the complexity comprises:
determining a variance scale corresponding to each layer of the remote sensing image in the pyramid image based on the complexity, wherein the complexity and the variance scale are in a negative correlation relationship, the difference value of the variance scales corresponding to two adjacent layers in the pyramid image is equal, and the variance scale corresponding to the next layer in the two adjacent layers is larger than the variance scale corresponding to the previous layer;
and determining a sampling step length based on the complexity, wherein the complexity and the sampling step length are in a negative correlation.
10. The method for enhancing the detail of the high-resolution remote sensing image according to claim 1, wherein the SITF operator is adopted to detect the feature points of all the downsampled images to obtain the corresponding feature points.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410027006.1A CN117541654A (en) | 2024-01-09 | 2024-01-09 | Detail enhancement method for high-resolution remote sensing image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410027006.1A CN117541654A (en) | 2024-01-09 | 2024-01-09 | Detail enhancement method for high-resolution remote sensing image |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117541654A true CN117541654A (en) | 2024-02-09 |
Family
ID=89790372
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410027006.1A Pending CN117541654A (en) | 2024-01-09 | 2024-01-09 | Detail enhancement method for high-resolution remote sensing image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117541654A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5592571A (en) * | 1994-03-08 | 1997-01-07 | The University Of Connecticut | Digital pixel-accurate intensity processing method for image information enhancement |
| DE602006020361D1 (en) * | 2006-11-24 | 2011-04-07 | Bosch Gmbh Robert | PROCESS, DEVICE AND COMPUTER PROGRAM FOR IMPROVING DETAILED VISIBILITY IN AN INPUT IMAGE |
| CN114418906A (en) * | 2021-12-02 | 2022-04-29 | 中国航空工业集团公司洛阳电光设备研究所 | Image contrast enhancement method and system |
| CN114494898A (en) * | 2021-12-24 | 2022-05-13 | 山东土地集团数字科技有限公司 | Remote sensing interpretation method and equipment for farmland protection |
| WO2022259072A1 (en) * | 2021-06-07 | 2022-12-15 | Precision Planting Llc | Systems and methods for providing field views including enhanced agricultural maps having a data layer and image data |
| CN115984134A (en) * | 2022-12-30 | 2023-04-18 | 黄河水利职业技术学院 | Intelligent enhancing method for remote sensing mapping image |
| CN116843584A (en) * | 2023-09-01 | 2023-10-03 | 浙江鼎立实业有限公司 | Image data optimization enhancement method |
-
2024
- 2024-01-09 CN CN202410027006.1A patent/CN117541654A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5592571A (en) * | 1994-03-08 | 1997-01-07 | The University Of Connecticut | Digital pixel-accurate intensity processing method for image information enhancement |
| DE602006020361D1 (en) * | 2006-11-24 | 2011-04-07 | Bosch Gmbh Robert | PROCESS, DEVICE AND COMPUTER PROGRAM FOR IMPROVING DETAILED VISIBILITY IN AN INPUT IMAGE |
| WO2022259072A1 (en) * | 2021-06-07 | 2022-12-15 | Precision Planting Llc | Systems and methods for providing field views including enhanced agricultural maps having a data layer and image data |
| CN114418906A (en) * | 2021-12-02 | 2022-04-29 | 中国航空工业集团公司洛阳电光设备研究所 | Image contrast enhancement method and system |
| CN114494898A (en) * | 2021-12-24 | 2022-05-13 | 山东土地集团数字科技有限公司 | Remote sensing interpretation method and equipment for farmland protection |
| CN115984134A (en) * | 2022-12-30 | 2023-04-18 | 黄河水利职业技术学院 | Intelligent enhancing method for remote sensing mapping image |
| CN116843584A (en) * | 2023-09-01 | 2023-10-03 | 浙江鼎立实业有限公司 | Image data optimization enhancement method |
Non-Patent Citations (3)
| Title |
|---|
| 刘攀 等: "基于自适应伽马转换函数的遥感图像增强", 电脑知识与技术, no. 03, 25 January 2013 (2013-01-25), pages 129 - 133 * |
| 向阳: "基于编码解码的高分辨率遥感图像变化检测", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》, no. 01, 15 January 2021 (2021-01-15), pages 028 - 198 * |
| 郭红山 等: "基于萤火虫算法的农业遥感图像增强研究", 浙江农业学报, no. 06, 25 June 2016 (2016-06-25), pages 181 - 186 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104200442B (en) | Non-local mean MRI image denoising method based on improved canny rim detections | |
| CN102968798B (en) | SAR (Synthetic Aperture Radar) image sea-land segmentation method based on wavelet transform and OTSU threshold | |
| Hou | Automatic segmentation of retinal blood vessels based on improved multiscale line detection | |
| CN115841434B (en) | Infrared image enhancement method for gas concentration analysis | |
| CN116188465B (en) | Crop growth state detection method based on image processing technology | |
| TWI624807B (en) | Iterative analysis of medical images | |
| CN116110053B (en) | Container surface information detection method based on image recognition | |
| CN113837974B (en) | NSST domain power equipment infrared image enhancement method based on improved BEEPS filtering algorithm | |
| CN106127209B (en) | A kind of objects' contour extracting method based on local edge feature integration | |
| He et al. | Retinal image denoising via bilateral filter with a spatial kernel of optimally oriented line spread function | |
| CN116245880B (en) | Electric vehicle charging pile fire risk detection method based on infrared identification | |
| CN116664457B (en) | Image processing method for enhancing denoising | |
| CN116416252A (en) | Method for detecting sedimentation image of wastewater in boehmite production process | |
| CN113822279B (en) | Infrared target detection method, device, equipment and medium based on multi-feature fusion | |
| CN116740579B (en) | Intelligent collection method for territorial space planning data | |
| CN109035228A (en) | A kind of radioscopic image processing method of non-uniform thickness component | |
| CN102222325A (en) | Method for detecting object contour based on phase characteristic | |
| CN115994870B (en) | Image processing method for enhancing denoising | |
| CN117274293A (en) | Accurate bacterial colony dividing method based on image features | |
| CN117541654A (en) | Detail enhancement method for high-resolution remote sensing image | |
| CN115908155A (en) | NSST domain combined GAN and scale correlation coefficient low-illumination image enhancement and denoising method | |
| CN112734666B (en) | SAR image speckle non-local mean suppression method based on similarity value | |
| CN108765448A (en) | A kind of shrimp seedling analysis of accounts method based on improvement TV-L1 models | |
| Zhao et al. | Edge detection and processing method of plant root image based on joint wavelet transform | |
| Yuan et al. | Automatic detection of solar features in HSOS full-disk solar images using guided filter |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |