A method and apparatus for linear target determination based on remote sensing images

Abstract

This application discloses a method and apparatus for determining linear targets based on remote sensing images. Based on remote sensing images and an algorithm for automatic identification of linear targets, candidate linear targets are obtained. By comparing the difference between the spectral average value of the candidate linear targets and the spectral average value of the linear target samples, if the difference is less than or equal to a preset spectral threshold, the candidate linear targets are extracted as archaeological linear targets, thereby realizing the automatic and accurate identification and extraction of archaeological linear targets based on remote sensing images.

Description

Technical Field

[0001] This application relates to the field of remote sensing archaeology technology, and in particular to a method and apparatus for determining linear targets based on remote sensing images. Background Technology

[0002] With the rapid development of remote sensing technology, remote sensing images have been widely used in various fields. In the current field of archaeological research, obtaining basic information about linear archaeological sites such as ancient city walls from remote sensing images is one of the main objectives of current remote sensing archaeological research.

[0003] Currently, obtaining basic information about archaeological targets from remote sensing images is mostly achieved through manual visual interpretation and computer-automated interpretation. Manual visual interpretation involves personnel directly observing or using interpretation instruments to acquire information about specific target features on remote sensing images. Computer-automated interpretation involves identifying and classifying the attributes of information in remote sensing images, recognizing the actual features corresponding to the image information, and extracting feature information. However, manual visual interpretation requires comprehensive intervention from archaeological researchers in image preparation, processing, and interpretation. Computer-automated interpretation, on the other hand, requires researchers to pre-determine which features are archaeological targets on the remote sensing images. Both methods still require significant manual intervention from professional archaeological researchers, consuming substantial time, manpower, and resources. Summary of the Invention

[0004] This application provides a method and apparatus for determining linear targets based on remote sensing images. It designs an algorithm for automatic identification and extraction of linear archaeological targets, reduces the intervention of archaeological researchers, and achieves the effect of automatically and accurately identifying and extracting real linear archaeological targets from remote sensing images.

[0005] In a first aspect, this application provides a method for determining linear targets based on remote sensing images, the method comprising:

[0006] The spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples in the first image are obtained. The first image is an image after preprocessing an archaeological remote sensing image. The archaeological remote sensing image includes at least one archaeological linear target. The spectral values ​​include the spectral mean and spectral standard deviation.

[0007] Based on the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples, the first image is sequentially segmented and geometrically filtered to obtain the second image, which includes at least one candidate linear target.

[0008] Obtain the spectral average value of the third image of each candidate linear target of the at least one candidate linear target in the second image, wherein the third image of each candidate linear target includes the pixels of the candidate linear target in the second image;

[0009] If the spectral mean of the third image matches the spectral mean of the archaeological linear target sample, then the candidate linear target corresponding to the third image is determined to be an archaeological linear target.

[0010] Optionally, the spectral mean of the third image is matched with the spectral mean of the archaeological linear target sample, including:

[0011] The difference between the average spectral value of the third image and the average spectral value of the archaeological linear target sample is less than or equal to a preset spectral threshold.

[0012] Optionally, based on the spectral values ​​of the archaeological linear target samples and the spectral values ​​of the archaeological nonlinear target samples, the first image is sequentially subjected to image segmentation and geometric filtering to obtain the second image, including:

[0013] Calculate the spectral separation index of at least one band based on the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample;

[0014] The segmented band is determined based on the spectral separation index of at least one band, and the segmented band is one of the at least one bands.

[0015] The first image is segmented according to the segmentation bands to obtain the segmentation result image;

[0016] Geometric filtering is applied to the segmentation result image to obtain the second image.

[0017] Optionally, a spectral separation index for at least one band is calculated based on the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample, including:

[0018] The spectral difference is obtained based on the spectral average of the linear archaeological target samples and the spectral average of the nonlinear archaeological target samples.

[0019] The total spectral value is obtained based on the spectral standard deviation of the linear archaeological target samples and the spectral standard deviation of the nonlinear archaeological target samples.

[0020] Calculate the spectral separation index for at least one band based on the spectral difference and the total spectral value.

[0021] Optionally, the segmented bands are determined based on the spectral separation index of at least one band, including:

[0022] The band corresponding to the maximum value of the spectral separation index of at least one band is determined as the segmented band.

[0023] Optionally, after determining that the candidate linear target corresponding to the third image is an archaeological linear target, the method further includes:

[0024] The position coordinates of the archaeological linear target in the parameter space are extracted according to the line detection algorithm, and the line segment is drawn according to the position coordinates to obtain the fourth image.

[0025] Optionally, obtaining the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples in the first image includes:

[0026] Obtain at least one image containing an archaeological linear target from the first image, as an archaeological linear target sample;

[0027] Obtain at least one image that does not contain archaeological linear targets from the first image, as an archaeological nonlinear target sample;

[0028] The spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples are statistically analyzed using spatial analysis methods to obtain the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples.

[0029] Optionally, obtaining the spectral average of the third image for each candidate linear target in the second image (at least one candidate linear target in the second image) includes:

[0030] Obtain the third image corresponding to each candidate linear target in the second image;

[0031] The average spectral value of the third image corresponding to each candidate linear target is obtained by statistically analyzing the spatial analysis method.

[0032] Optionally, the method further includes:

[0033] Obtain remote sensing images acquired by sensors;

[0034] The remote sensing image is preprocessed to obtain the first image. The preprocessing operations include radiometric calibration, atmospheric correction and geometric correction.

[0035] Secondly, this application provides an apparatus for linear target determination based on remote sensing images, the apparatus comprising:

[0036] The first obtaining unit is used to obtain the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples in the first image. The first image is an image after preprocessing an archaeological remote sensing image. The archaeological remote sensing image includes at least one archaeological linear target. The spectral values ​​include the spectral mean and spectral standard deviation.

[0037] The first processing unit is used to perform image segmentation and geometric filtering on the first image sequentially based on the spectral values ​​of the archaeological linear target sample and the spectral values ​​of the archaeological nonlinear target sample to obtain a second image, the second image including at least one candidate linear target;

[0038] The second obtaining unit is used to obtain the spectral average value of the third image of each candidate linear target in the second image, wherein the third image of each candidate linear target includes the pixel points of the candidate linear target in the second image;

[0039] The second processing unit is used to determine the candidate linear target corresponding to the third image as an archaeological linear target if the spectral average value of the third image matches the spectral average value of the archaeological target sample.

[0040] Optionally, the second processing unit is specifically used for:

[0041] Calculate the difference between the spectral mean of the third image and the spectral mean of the archaeological linear target samples. If the difference is less than or equal to a preset spectral threshold, then the candidate linear target corresponding to the third image is an archaeological linear target.

[0042] Optionally, the first processing unit is specifically used for:

[0043] The spectral difference is obtained based on the spectral average of the linear archaeological target samples and the spectral average of the nonlinear archaeological target samples.

[0044] The total spectral value is obtained based on the spectral standard deviation of the linear archaeological target samples and the spectral standard deviation of the nonlinear archaeological target samples.

[0045] Calculate the spectral separation index for at least one band based on the spectral difference and the total spectral value;

[0046] The band corresponding to the maximum value of the spectral separation index of at least one band is determined as the segmented band, and the segmented band is one of the bands in at least one band.

[0047] The first image is segmented according to the segmentation bands to obtain the segmentation result image;

[0048] Geometric filtering is applied to the segmentation result image to obtain the second image.

[0049] Optionally, the device further includes:

[0050] The third processing unit is used to extract the position coordinates of the archaeological linear target in the parameter space according to the line detection algorithm, and draw the line segment according to the position coordinates to obtain the fourth image.

[0051] Optionally, the first obtaining unit is specifically used for:

[0052] Obtain at least one image containing an archaeological linear target from the first image, as an archaeological linear target sample;

[0053] Obtain at least one image that does not contain archaeological linear targets from the first image, as an archaeological nonlinear target sample;

[0054] The spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples are statistically analyzed using spatial analysis methods to obtain the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples.

[0055] Optionally, the second obtaining unit is specifically used for:

[0056] Obtain the third image corresponding to each candidate linear target in the second image;

[0057] The average spectral value of the third image corresponding to each candidate linear target is obtained by statistically analyzing the spatial analysis method.

[0058] Optionally, the device further includes:

[0059] The fourth processing unit is used to perform preprocessing operations on the remote sensing images obtained by the sensor to obtain the first image.

[0060] Thirdly, this application provides a device for linear target determination based on remote sensing images, the device including a memory and a processor:

[0061] Memory is used to store computer programs;

[0062] The processor is used to execute the method provided in the first aspect above according to the computer program.

[0063] Fourthly, this application also provides a computer-readable storage medium for storing a computer program for performing the method provided in the first aspect above.

[0064] Therefore, this application has the following beneficial effects:

[0065] This application provides a method and apparatus for determining linear targets based on remote sensing images. The method identifies candidate linear targets in the remote sensing image through image segmentation and geometric filtering. Then, by comparing the spectral average of the candidate linear targets with the spectral average of archaeological linear target samples, the candidate linear targets are extracted as archaeological linear targets. The technical solution of this application obtains the spectral values ​​of archaeological target samples and non-archaeological target samples from the remote sensing image. Image segmentation and geometric filtering are performed based on the archaeological and non-archaeological targets to extract candidate linear targets from the remote sensing image. The spectral average of the candidate linear targets is then obtained. Finally, the spectral average of the candidate linear targets is compared with the spectral average of the archaeological target samples. If the spectral average of the candidate linear target matches the spectral average of an archaeological linear target sample that is clearly an archaeological target, then the linear target is determined to be an archaeological target. This achieves automatic identification and extraction of archaeological linear targets from remote sensing images, reduces the intervention of archaeological researchers, and provides highly consistent identification results. Attached Figure Description

[0066] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0067] Figure 1 This is a flowchart illustrating a method for determining linear targets based on remote sensing images, as described in an embodiment of this application.

[0068] Figure 2 This is a schematic diagram of an archaeological linear target image determined in a method for determining linear targets based on remote sensing images according to an embodiment of this application.

[0069] Figure 3 This is a schematic diagram of an archaeological linear target image extracted by a method for determining linear targets based on remote sensing images in an embodiment of this application.

[0070] Figure 4 This is a flowchart illustrating an embodiment of a method for determining linear targets based on remote sensing images, as described in this application.

[0071] Figure 5 This is a schematic diagram of the structure of a linear target determination device 500 based on remote sensing images in an embodiment of this application;

[0072] Figure 6 This is a schematic diagram of the structure of a device 600 for linear target determination based on remote sensing images, as described in an embodiment of this application. Detailed Implementation

[0073] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0074] In the embodiments of this application, the word "first" in names such as "first obtaining unit" and "first processing unit" is only used for name identification and does not represent the first in order. The same rule applies to "second," "third," etc.

[0075] Currently, in the field of remote sensing archaeology, research on the baseline information (size, location, and structure) of linear archaeological targets, both domestically and internationally, mainly focuses on visual interpretation or human-computer interactive interpretation of multi-source remote sensing images. However, manual visual interpretation requires comprehensive intervention from archaeological researchers in image preparation, image processing, and image interpretation. Computer interpretation also requires archaeological researchers to pre-determine which parts of the remote sensing image are archaeological sites during the computer interpretation process. These research methods all require a significant amount of manual intervention from archaeological researchers in obtaining linear archaeological targets, consuming substantial time, manpower, and resources.

[0076] In this embodiment of the application, the method for determining linear targets based on remote sensing images first obtains the spectral values ​​of linear targets and nonlinear targets in a first image. Based on the obtained spectral values, the archaeological remote sensing image is segmented and geometrically filtered to obtain a second image containing candidate linear targets. Then, the spectral average value of a third image for each candidate linear target in the second image is obtained. Finally, if the spectral average value of the third image matches the spectral average value of the archaeological linear target sample, the candidate linear target corresponding to the third image is determined to be an archaeological linear target.

[0077] Specifically, the method may include, for example, obtaining a first image based on remote sensing images through image preprocessing; obtaining the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples in the first image through visual interpretation; performing image segmentation and geometric filtering on the first image sequentially based on the spectral values ​​of the archaeological linear target samples and archaeological nonlinear target samples to obtain a second image containing at least one candidate linear target; obtaining the spectral average value of a third image for each candidate linear target in the second image based on the second image using spatial analysis methods; determining whether the difference between the spectral average value of the third image and the spectral average value of the archaeological linear target samples is less than or equal to a preset spectral threshold; if the difference is less than or equal to the preset spectral threshold, then determining the candidate linear target corresponding to the third image as an archaeological linear target. The method provided in this application can automatically identify and extract archaeological linear targets from remote sensing images by comparing the magnitude of the spectral average value of the linear target samples and the spectral average value of the candidate linear targets, reducing the intervention of archaeological researchers.

[0078] To facilitate understanding of the specific implementation of the method for determining linear targets based on remote sensing images provided in the embodiments of this application, the following description will be provided in conjunction with the accompanying drawings.

[0079] It should be noted that the main body implementing the method for determining linear targets based on remote sensing images can be the apparatus for determining linear targets based on remote sensing images provided in the embodiments of this application. This apparatus can be carried in an electronic device or a functional module of an electronic device. The electronic device in the embodiments of this application can be any device capable of implementing the method for determining linear targets based on remote sensing images in the embodiments of this application, such as an Internet of Things (IoT) device.

[0080] Figure 1 This is a flowchart illustrating a method for determining linear targets based on remote sensing images, provided as an embodiment of this application. This method can be applied to an apparatus for determining linear targets based on remote sensing images, such as... Figure 5 The illustrated apparatus 500 for linear target determination based on remote sensing images, or the apparatus for linear target determination based on remote sensing images may also be integrated into... Figure 6 The functional modules in the device 600 for linear target determination based on remote sensing images are shown.

[0081] like Figure 1 As shown, the method includes the following steps:

[0082] S101: Obtain the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample in the first image. The first image is an image after preprocessing the archaeological remote sensing image. The archaeological remote sensing image includes at least one archaeological linear target. The spectral values ​​include the spectral mean and the spectral standard deviation.

[0083] Remote sensing images typically refer to images acquired by satellite sensors. Due to their macroscopic, objective, and convenient nature, remote sensing images have been applied in various research fields, such as archaeology. In this application's embodiments, the remote sensing image refers to an image acquired from an archaeological scene. The first image can be an image obtained after preprocessing the remote sensing image of the archaeological scene. Preprocessing operations can include, but are not limited to, one or more of radiometric calibration, atmospheric correction, and geometric correction. For example, if the sensor includes a function to perform atmospheric correction on the acquired remote sensing image, then the preprocessing operation of the device for determining linear targets based on the remote sensing image may not include atmospheric correction.

[0084] Currently, in the field of archaeological research, remote sensing images are also used to obtain information on linear targets such as ancient city walls. To enable computers to automatically obtain linear archaeological targets, the method provided in this application first acquires the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples from a first image. Then, based on the spectral values ​​of the archaeological linear target samples and archaeological nonlinear target samples, a second image containing at least one candidate linear target is obtained. Next, the average spectral value of the candidate linear targets is acquired. By determining whether the average spectral value of each candidate linear target matches the average spectral value of the archaeological linear target samples, if the average spectral value of the candidate linear target matches the average spectral value of the archaeological linear target samples, then the candidate linear target is determined to be an archaeological linear target. Therefore, this application embodiment, through step S101, obtains the spectral values ​​of archaeological linear target samples and archaeological nonlinear samples, providing a prerequisite for subsequently acquiring candidate linear targets.

[0085] As an example, obtaining the spectral values ​​of the archaeological linear target sample in the first image in S101 can include: one case, obtaining an image of a linear target belonging to the archaeological linear target from the first image through visual interpretation, as the archaeological linear target sample, and statistically analyzing the spectral mean and standard deviation of the archaeological linear target sample according to spatial analysis methods, as the spectral values ​​of the archaeological linear target sample in S101; another case, obtaining images of multiple linear targets belonging to the archaeological linear target from the first image through visual interpretation, as multiple archaeological linear target samples, and statistically analyzing the spectral mean and standard deviation of the multiple archaeological linear target samples respectively according to spatial analysis methods, then calculating the average of the spectral mean of the multiple archaeological linear target samples, and calculating the average of the spectral standard deviation of the multiple archaeological linear target samples, as the spectral values ​​of the archaeological linear target samples in S101.

[0086] As an example, obtaining the spectral values ​​of the archaeological nonlinear target sample in the first image in S101 can include: one case, obtaining an image of a target that does not belong to the archaeological linear target from the first image through visual interpretation, as an archaeological nonlinear target sample, and statistically analyzing the spectral mean and standard deviation of the archaeological nonlinear target sample according to spatial analysis methods, as the spectral value of the archaeological nonlinear target sample in S101; another case, obtaining images of multiple targets that do not belong to the archaeological linear target from the first image through visual interpretation, as multiple archaeological nonlinear target samples, and statistically analyzing the spectral mean and standard deviation of the multiple archaeological nonlinear target samples respectively according to spatial analysis methods, then calculating the average of the spectral mean of the multiple archaeological nonlinear target samples, and calculating the average of the spectral standard deviation of the multiple archaeological nonlinear target samples, as the spectral value of the archaeological nonlinear target sample in S101.

[0087] S102: Based on the spectral values ​​of the archaeological linear target sample and the spectral values ​​of the archaeological nonlinear target sample, the first image is sequentially processed by image segmentation and geometric filtering to obtain a second image, the second image including at least one candidate linear target.

[0088] As an example, S102 may include: S1021, calculating the spectral separation index of at least one band based on the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample; S1022, determining the band corresponding to the maximum value of the spectral separation index of at least one band as the segmentation band; S1023, performing image segmentation on the first image based on the segmentation band, segmenting the first image into candidate linear targets as targets and archaeological nonlinear targets as background, and obtaining a segmentation result image; S1024, performing geometric filtering on the segmentation result image to remove the archaeological nonlinear targets as background in the segmentation result image, and obtaining a second image.

[0089] It should be noted that in S1022, since the maximum spectral separation index represents the minimization of intra-class variance and the maximization of inter-class variance of the archaeological linear target sample and the non-archaeological linear target sample in the corresponding spectral band or band operation, using the band corresponding to the maximum spectral separation index for image segmentation can effectively separate the archaeological linear target sample and the archaeological non-linear target sample in that band or band operation. Therefore, the band corresponding to the maximum spectral separation index is determined as the separation band.

[0090] Specifically, in step S1021, the spectral separation index of at least one band is calculated based on the spectral values ​​of the linear and nonlinear archaeological target samples. For example, this may include: calculating the spectral separation index of each band using steps S10211 to S10213 as follows: S10211, calculating the difference between the average spectral value of the linear archaeological target sample and the average spectral value of the nonlinear archaeological target sample in the band, as the spectral difference value in the band; S10212, calculating the difference between the standard deviation of the spectral values ​​of the linear and nonlinear archaeological target samples, as the total spectral value in the band; S10213, calculating the spectral separation index of the band based on the spectral difference value and the total spectral value in the band.

[0091] S103: Obtain the spectral average value of the third image of each candidate linear target of the at least one candidate linear target in the second image, wherein the third image of each candidate linear target includes the pixels of the candidate linear target in the second image.

[0092] As an example, obtaining the spectral average value of the third image of each candidate linear target of the at least one candidate linear target in the second image in S103 may include: obtaining a third image containing each candidate linear target based on the second image, and statistically analyzing the spectral average value of the third image corresponding to each candidate linear target according to a spatial analysis method to obtain the spectral average value corresponding to the candidate linear target.

[0093] S104: If the spectral average value of the third image matches the spectral average value of the archaeological linear target sample, then the candidate linear target corresponding to the third image is determined to be an archaeological linear target.

[0094] Wherein, the spectral average value of the third image matches the spectral average value of the linear target sample, for example, the difference between the spectral average value of the third image and the spectral average value of the linear target sample is less than or equal to a preset spectral threshold.

[0095] In some implementations, after determining that the candidate linear target corresponding to the third image is an archaeological linear target, the position coordinates of the archaeological linear target in the parameter space are extracted on the third image according to the line detection algorithm, and a straight line segment is drawn according to the position coordinates to obtain the fourth image.

[0096] It should be noted that the method provided in this application embodiment can be understood as a model for determining linear targets based on remote sensing images. Therefore, a first image can be input into this model, and the model outputs a fourth image, which includes the archaeological linear targets present in the first image. For example, Figure 2 The first image shown is input to the linear target determination model based on remote sensing images; the fourth output image can be found in [reference needed]. Figure 3 As shown, Figure 3 Line segments A and B, as marked in the text, represent the archaeological linear targets determined by implementing the method provided in the embodiments of this application.

[0097] As can be seen, the method of this application embodiment obtains candidate linear targets based on remote sensing images and an algorithm for automatically identifying linear targets. By comparing the difference between the spectral average value of the candidate linear targets and the spectral average value of the archaeological linear target samples, if the difference is less than or equal to a preset spectral threshold, the candidate linear target is extracted as an archaeological linear target. This achieves accurate identification and extraction of archaeological linear targets from remote sensing images. Furthermore, the method has a high degree of automation and high accuracy in the identification results.

[0098] To make the methods provided in the embodiments of this application clearer and easier to understand, the following is combined with... Figure 4 A specific example of this method is then provided.

[0099] S401: Acquire raw remote sensing images.

[0100] The acquired raw remote sensing images can be high-resolution remote sensing images of the area containing soil or vegetation markers indicating the archaeological linear targets. Soil markers specifically refer to linear archaeological targets on the high-resolution remote sensing image that indicate areas of bare ground cover, while vegetation markers specifically refer to linear archaeological targets on the high-resolution remote sensing image that indicate areas of vegetation cover. Preferably, archaeological targets indicated by soil markers were captured at any time in autumn, and archaeological targets indicated by vegetation markers were preferably captured at any time in spring. The images contain four multispectral bands (blue, green, red, and near-infrared) and one panchromatic band, preferably in the red band with wavelengths of 0.63 μm to 0.69 μm and in the near-infrared band with wavelengths of 0.77 μm to 0.89 μm. The image resolution is preferably 0.5 m to 2 m, more preferably 0.5 m to 1 m, and the area corresponding to each pixel is 0.25 square meters (m²). 2 )~4m 2 The preferred size is 0.25m. 2 ~1m 2 .

[0101] As an example, S401 may include: acquiring multi-band high-resolution remote sensing images of the study area taken by the GF-1 satellite, obtaining 8m resolution multispectral image data in four bands (blue, green, red, and near-infrared) and 2m resolution panchromatic image data. Image data

[0102] S402: Preprocess the original remote sensing image to obtain a first image.

[0103] To improve the accuracy of linear target recognition and extraction, preprocessing of the original remote sensing images is necessary.

[0104] As an example, S402 may include: performing radiometric correction (including radiometric calibration and atmospheric correction) on the multispectral and panchromatic data of GF-1 image data using the atmospheric radiative transfer 6S model. The radiometric calibration method preferably uses the formula: L(λ) = Gain·DN + Bias, where L(λ) is the radiance value at the sensor entrance pupil, Gain is the gain coefficient, DN is the observed gray value, and Bias is the bias coefficient. Then, orthorectification is performed on the radiometrically corrected image data using the RPC file included with the GF-1 image data. Finally, the Gram-Schmidt algorithm is used to perform image fusion of the multispectral and panchromatic data to obtain 2m resolution color fusion data in four bands: blue, green, red, and near-infrared.

[0105] S403: Obtain archaeological linear target samples containing archaeological linear targets and archaeological nonlinear target samples not containing archaeological linear targets based on the first image.

[0106] To facilitate subsequent calculation of the spectral separation index, it is preferable that the archaeological linear target samples and archaeological nonlinear target samples for each band are sampled at the same location, and that the number of samples is equal.

[0107] S404: Obtain the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample.

[0108] S405: Perform image segmentation on the first image based on the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample to obtain a segmentation result image.

[0109] In order to distinguish between linear and nonlinear archaeological targets in the first image, image segmentation is required.

[0110] As an example, S405 may include: S4051, firstly, it is necessary to determine the segmented bands, for example, the segmented bands can be determined based on the spectral separation index, and the formula for calculating the spectral separation index can be: SSI = (μ 1(λ) -μ 2(λ) ) / (σ 1(λ) +σ 2(λ) ), where μ 1(λ) and μ 2(λ) σ represents the average spectral values ​​of the linear and nonlinear archaeological target samples, respectively. 1(λ) and σ 2(λ) The standard deviations of the spectral values ​​of the archaeological target samples and the non-archaeological target samples are respectively represented. The band with the largest spectral separation index among each band is selected as the segmentation band. This ensures that the intra-class variance of the archaeological linear target samples and the archaeological nonlinear target samples in the first image is minimized and the inter-class variance is maximized in the segmentation band. S4052, the first image is processed using an adaptive gray-scale thresholding algorithm to obtain the segmentation result image. The optimal threshold in the adaptive gray-scale thresholding algorithm is one that can produce separation performance between the foreground class (archaeological linear targets) and the background class (archaeological nonlinear targets). The maximum inter-class variance method is better used to characterize the threshold.

[0111] S406: Perform geometric filter processing on the segmentation result image to obtain a second image.

[0112] To remove archaeological nonlinear targets that serve as background elements from the segmentation results, the segmentation results need to be processed using a geometric filter.

[0113] As an example, S406 may include: First, calculating the area of ​​each segmented patch and the aspect ratio of its circumscribed rectangle in the segmented image; then, comparing whether the area of ​​each segmented patch and the aspect ratio of its circumscribed rectangle are within the threshold values ​​for the area parameter and the aspect ratio parameter of the circumscribed rectangle determined by archaeological experts' experience; if the patches are within the threshold values, they are retained; then, candidate linear targets are selected to obtain the second image.

[0114] S407: Obtain the spectral average of the third image for each candidate linear target in the second image.

[0115] S408: Calculate the difference between the spectral average value of the third image and the spectral average value of the archaeological linear target sample.

[0116] To distinguish archaeological linear targets from other linear targets in remote sensing images, it is necessary to determine whether the spectral mean is within a certain range. Therefore, it is necessary to calculate the difference between the spectral mean of the third image and the spectral mean of the archaeological linear target sample.

[0117] S409: If the difference is less than or equal to a preset spectral threshold, then the candidate linear target corresponding to the third image is determined to be an archaeological linear target.

[0118] The preset spectral threshold can be flexibly set according to the attributes of the remote sensing image.

[0119] S410: Extract the position parameters of the archaeological linear target according to the line detection algorithm, and draw a straight line segment according to the position parameters to obtain the fourth image.

[0120] In order to extract archaeological linear targets, a line detection algorithm is needed to display the archaeological linear targets in the image.

[0121] As an example, S410 may include: obtaining the Hough peak point of the archaeological linear target corresponding to the third image according to the Hough transform line detection algorithm, calculating the position information of (x,y) using the (ρ,θ) coordinates of the peak point parameter space, drawing the corresponding straight line of the archaeological linear target according to the position information, and obtaining the endpoint information after drawing the straight line. The straight line and the endpoints can be displayed by color. For example, the straight line segment can be pink, the starting point of the endpoint can be yellow, and the ending point of the endpoint can be red.

[0122] This application provides a method for determining linear targets based on remote sensing images. The method involves preprocessing the acquired remote sensing images to obtain a first image; obtaining spectral values ​​of archaeological linear target samples (belonging to archaeological linear targets) and archaeological nonlinear target samples (not belonging to archaeological linear targets) based on the first image; then, performing image segmentation and geometric filtering on the first image based on the spectral values ​​of the archaeological linear target samples and the archaeological nonlinear target samples to obtain a second image; obtaining the average spectral value of a third image for each candidate linear target (at least one candidate linear target) based on the second image; finally, calculating the difference between the average spectral value of the third image and the average spectral value of the archaeological linear target samples. If the difference is less than or equal to a preset spectral threshold, the candidate linear target corresponding to the third image is an archaeological linear target, and a line detection algorithm is used to mark the archaeological linear target in the first image, thereby obtaining a fourth image.

[0123] This application's embodiments have designed an algorithm for the automatic identification and extraction of linear archaeological targets, enabling remote sensing archaeological exploration and surveys to have a high degree of automation, thereby reducing the need for intervention by archaeological researchers and achieving the goal of automatically identifying and extracting linear archaeological targets based on remote sensing images, with high accuracy in the identification results.

[0124] See Figure 5 This application provides an apparatus 500 for determining linear targets based on remote sensing images. The apparatus 500 includes:

[0125] The first obtaining unit 501 is used to obtain the spectral values ​​of the archaeological linear target sample and the archaeological nonlinear target sample in the first image. The first image is an image after preprocessing the archaeological remote sensing image. The archaeological remote sensing image includes at least one archaeological linear target. The spectral values ​​include the spectral mean and the spectral standard deviation.

[0126] The first processing unit 502 is used to perform image segmentation and geometric filtering on the first image sequentially based on the spectral values ​​of the archaeological linear target sample and the spectral values ​​of the archaeological nonlinear target sample to obtain a second image, wherein the second image includes at least one candidate linear target.

[0127] The second obtaining unit 503 is used to obtain the spectral average value of the third image of each candidate linear target of at least one candidate linear target in the second image, wherein the third image of each candidate linear target includes the pixel points of the candidate linear target in the second image;

[0128] The second processing unit 504 is used to determine the candidate linear target corresponding to the third image as an archaeological linear target if the spectral average value of the third image matches the spectral average value of the archaeological target sample.

[0129] Optionally, the second processing unit 504 is specifically used for:

[0130] Calculate the difference between the spectral mean of the third image and the spectral mean of the archaeological linear target samples. If the difference is less than or equal to a preset spectral threshold, then the candidate linear target corresponding to the third image is an archaeological linear target.

[0131] Optionally, the first processing unit 502 is specifically used for:

[0132] The spectral difference is obtained based on the spectral average of the linear archaeological target samples and the spectral average of the nonlinear archaeological target samples.

[0133] The total spectral value is obtained based on the spectral standard deviation of the linear archaeological target samples and the spectral standard deviation of the nonlinear archaeological target samples.

[0134] Calculate the spectral separation index of the at least one band based on the spectral difference and the total spectral value;

[0135] The band corresponding to the maximum value of the spectral separation index of at least one band is determined as the segmented band, and the segmented band is one of the bands in at least one band.

[0136] The first image is segmented according to the segmentation bands to obtain the segmentation result image;

[0137] Geometric filtering is applied to the segmentation result image to obtain the second image.

[0138] Optionally, the device 500 further includes:

[0139] The third processing unit is used to extract the position coordinates of the archaeological linear target in the parameter space according to the line detection algorithm, and draw the line segment according to the position coordinates to obtain the fourth image.

[0140] Optionally, the first obtaining unit 501 is specifically used for:

[0141] Obtain at least one image containing an archaeological linear target from the first image, as an archaeological linear target sample;

[0142] Obtain at least one image that does not contain archaeological linear targets from the first image, as an archaeological nonlinear target sample;

[0143] The spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples are statistically analyzed using spatial analysis methods to obtain the spectral values ​​of archaeological linear target samples and archaeological nonlinear target samples.

[0144] Optionally, the second obtaining unit 503 is specifically used for:

[0145] Obtain the third image corresponding to each candidate linear target in the second image;

[0146] The average spectral value of the third image corresponding to each candidate linear target is obtained by statistically analyzing the spatial analysis method.

[0147] Optionally, the device 500 further includes:

[0148] The fourth processing unit is used to perform preprocessing operations on the remote sensing images obtained by the sensor to obtain the first image.

[0149] This application also provides an apparatus 600 for linear target determination based on remote sensing images, such as... Figure 6 As shown, the device 600 includes a memory 601 and a processor 602:

[0150] Memory 601 is used to store computer programs;

[0151] Processor 602 is used to execute the above according to the computer program. Figure 1 or Figure 4 The methods provided.

[0152] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that all or part of the steps in the methods of the above embodiments can be implemented by means of software plus a general-purpose hardware platform. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as a read-only memory (ROM) / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0153] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the device embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. The device embodiments described above are merely illustrative. Modules described as separate components may or may not be physically separate. Components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the objectives of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0154] The above description is merely an exemplary implementation of this application and is not intended to limit the scope of protection of this application.

Claims

1. A method for linear target determination based on remote sensing images, characterized in that, The method comprises the following steps: obtaining spectral values of an archaeological linear target sample and spectral values of an archaeological nonlinear target sample in a first image, the first image being an image obtained by preprocessing an archaeological remote sensing image, the archaeological remote sensing image comprising at least one archaeological linear target, and the spectral values comprising spectral mean values and spectral standard deviations; sequentially performing image segmentation and geometric filtering on the first image based on the spectral values of the archaeological linear target sample and the spectral values of the archaeological nonlinear target sample to obtain a second image, the second image comprising at least one candidate linear target; wherein the step of sequentially performing image segmentation and geometric filtering on the first image based on the spectral values of the archaeological linear target sample and the spectral values of the archaeological nonlinear target sample to obtain a second image comprises the following steps: obtaining spectral difference values according to the spectral mean values of the archaeological linear target sample and the spectral mean values of the archaeological nonlinear target sample; obtaining spectral total values according to the spectral standard deviations of the archaeological linear target sample and the spectral standard deviations of the archaeological nonlinear target sample; calculating spectral separation degree indexes of the at least one waveband according to the spectral difference values and the spectral total values; determining a segmentation waveband according to the spectral separation degree indexes of the at least one waveband, the segmentation waveband being one of the at least one waveband; performing image segmentation on the first image according to the segmentation waveband to obtain a segmentation result image; and performing the geometric filtering on the segmentation result image to obtain the second image; obtaining spectral mean values of a third image of each candidate linear target of the at least one candidate linear target in the second image, the third image of each candidate linear target comprising pixel points of the candidate linear target in the second image; if the spectral mean values of the third image match the spectral mean values of the archaeological linear target sample, determining that the candidate linear target corresponding to the third image is an archaeological linear target.

2. The method of claim 1, wherein, the spectral mean values of the third image match the spectral mean values of the archaeological linear target sample, comprising: a difference between the spectral mean values of the third image and the spectral mean values of the archaeological linear target sample is less than or equal to a preset spectral threshold.

3. The method of claim 1, wherein, the step of determining a segmentation waveband according to the spectral separation degree indexes of the at least one waveband, comprising: determining, as the segmentation waveband, a waveband corresponding to a maximum value in the spectral separation degree indexes of the at least one waveband.

4. The method according to any one of claims 1 to 3, characterized in that, after determining that the candidate linear target corresponding to the third image is an archaeological linear target, the method further comprises: extracting position coordinates of the archaeological linear target in a parameter space according to a straight line detection algorithm, and drawing a straight line segment according to the position coordinates to obtain a fourth image.

5. The method according to any one of claims 1 to 3, characterized in that, the step of obtaining spectral values of an archaeological linear target sample and spectral values of an archaeological nonlinear target sample in a first image, comprising: obtaining at least one image comprising an archaeological linear target from the first image as the archaeological linear target sample; obtaining at least one image not comprising an archaeological linear target from the first image as the archaeological nonlinear target sample; According to a spatial analysis method, the spectral values of the archeological linear target samples and the spectral values of the archeological nonlinear target samples are statistically analyzed to obtain the spectral values of the archeological linear target samples and the spectral values of the archeological nonlinear target samples.

6. The method of claim 1, wherein, The method further comprises: The method further comprises: The method further comprises:

7. The method of claim 1, wherein, The method further comprises: The method further comprises: The method further comprises:

8. An apparatus for linear target determination based on remote sensing images, characterized in that, The device comprises a memory and a processor, and the processor is configured to execute a program stored in the memory to run the method according to any one of claims 1-7. The computer readable storage medium is configured to store a computer program, and the computer program is configured to execute the method according to any one of claims 1-7. The device comprises a memory and a processor, and the processor is configured to execute a program stored in the memory to run the method according to any one of claims 1-7. The computer readable storage medium is configured to store a computer program, and the computer program is configured to execute the method according to any one of claims 1-7. ​ ​ 9. An apparatus for linear object extraction based on remote sensing images, characterized by, ​ 10. A computer-readable storage medium, characterized in that, ​

Images