A method for assisting archaeological investigation by registering historical remote sensing images

Abstract

The application discloses a kind of methods for archeological investigation assisted by historical remote sensing image registration, comprising: obtaining historical image, and preprocessing historical remote sensing image, obtaining preprocessed image to be registered, the preprocessing includes splicing and merging to historical remote sensing image, correcting direction and optimizing toning;Obtain the reference image corresponding to the range of image to be registered, use reference image to the preprocessed image to be registered preliminary registration, obtain preliminary registration result;According to image proofreading standard, preliminary registration result is reviewed: if it passes review, preliminary registration result is used as final registration result output;Otherwise, according to local readjustment method, preliminary registration result is optimized until it passes review, and the result after passing review is used as final registration result output.The method of the application has high value in social, economic and technical aspects, and can be used for subsequent link archeological investigation and analysis, and has practicality.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of archaeology, and particularly relates to a method for assisting archaeological investigation by using historical remote sensing image registration. BACKGROUND

[0002] Archaeological investigation is an important part of archaeology, which refers to finding and determining the location and range of ancient relics through various methods and means, and providing a basis for further archaeological excavation or archaeological research. The purpose of archaeological investigation is to reveal the living conditions, cultural characteristics, social organization, and historical changes of ancient humans, so as to enrich and perfect the historical memory of mankind. For example, the discovery of the Liangzhu ancient dam near the Liangzhu ancient city site provides many clues for archaeologists to understand the origin and development of Liangzhu culture and to understand the superb water conservancy technology and management ability of Liangzhu ancestors.

[0003] At present, the methods of archaeological investigation mainly include: ① literature research; ② ground survey; and ③ remote sensing detection. The first two methods are most widely used, but there are also some problems, for example:

[0004] ① Time-consuming and labor-intensive: archaeological investigation often requires a lot of time and manpower, especially in complex terrain and widely distributed sites, it is difficult to complete the investigation task quickly and effectively.

[0005] ② Destructive: archaeological investigation often causes some damage to the site, especially drilling, trial digging and other methods that directly contact the site, which may affect the integrity of the site.

[0006] ③ Inaccurate: the process of archaeological investigation is often disturbed by various factors, such as natural factors such as surface vegetation coverage, soil, and hydrology, as well as changes in place names recorded in literature, inaccurate direction records, and changes in topography, which may also be affected by subjective experience and lead to differences in judgment, which may lead to inaccurate and unreliable identification and positioning of the site.

[0007] In recent years, with the development of remote sensing detection means, using historical remote sensing images for archaeological investigation has become an innovative and efficient way of archaeological assistance. However, due to camera attitude, lens distortion or other optical factors, historical remote sensing images usually do not have accurate geographic coordinates and have geometric distortion, which cannot be matched with the actual geographic location. Therefore, it is necessary to perform geographic registration before use. The registration of historical remote sensing images refers to selecting an image with accurate geographic coordinates as the reference image, taking the historical remote sensing image as the image to be registered, and eliminating the distortion in the historical remote sensing image by selecting the same control points, so as to facilitate subsequent efficient archaeological investigation and accurate field reconnaissance by archaeologists.

[0008] Compared with the two methods mentioned above, the advantages of using historical remote sensing image registration method for archaeological investigation mainly include:

[0009] ①Efficient: Remote sensing images cover a wide range, and there is no need for extensive fieldwork, greatly saving time and manpower.

[0010] ②Non-destructive: Using historical image registration method for archaeological site investigation can avoid contact damage to the site and maintain the original appearance and integrity of the site.

[0011] ③Accurate: Using historical image registration method can use the differences between the site and the surrounding environment to accurately locate and position the suspected location and range of the site.

[0012] ④Dynamic: Using historical image registration method can use images from different periods to avoid changes caused by modern human activities, engineering, geological changes, and other factors that may obscure the characteristics of some sites.

[0013] Currently, image processing software has built-in computer automatic registration modules, but there are some unavoidable problems when using them for historical remote sensing image registration. For example, automatic registration is generally suitable for medium resolution (10-30m) images, but when the image resolution is high, the automatic registration effect is not good, and the historical remote sensing images used have a resolution of 0.6m. Automatic registration is generally suitable for cases where there is movement, rotation, or equal proportion deformation, but most historical remote sensing images not only have movement and rotation, but also have unequal proportion deformation such as stretching, compression, and twisting. The historical remote sensing images and modern images have a large age difference, resulting in significant changes in terrain and topography, and many control points change or even disappear. Therefore, if modern images are used as reference images, the changes in ground objects are large and the computer cannot identify the control points. If the reference image and the image to be registered have different resolutions or different center wavelengths, it will affect the automatic registration result. Therefore, if Tianditu is used as the reference image, the difference in resolution makes it difficult to automatically identify the control points. Therefore, the existing automatic registration module is not suitable for the registration of high-resolution historical remote sensing images.

[0014] In summary, using historical remote sensing image registration method for archaeological investigation is a new, effective, and promising method with broad application prospects. It can provide more information and basis for the archaeological investigation process, improve the efficiency of archaeology, and provide more possibilities for the exploration of ancient history of human civilization in China and even the world. SUMMARY

[0015] The purpose of the present application is to solve the problems in the prior art and provide a method for assisting archaeological investigation using historical remote sensing image registration.

[0016] In order to achieve the above-mentioned purposes, the present application specifically adopts the following technical solutions.

[0017] The present application provides a method for assisting archaeological investigation by registering historical remote sensing images, which comprises the following steps:

[0018] S1. Obtain a historical remote sensing image for archaeological investigation as a to-be-registered image, obtain a reference image corresponding to the range of the to-be-registered image, and pre-process the to-be-registered image with reference to the reference image to obtain a pre-processed to-be-registered image.

[0019] The pre-processing includes splicing and merging, correcting the direction, and optimizing the color of the to-be-registered image. The splicing and merging need to satisfy that the position sequence of the partitioned to-be-registered image is correct and all features in the overlapping part of the boundary transition line of the adjacent partitioned to-be-registered image are completely coincident. The optimization of the color needs to satisfy that the color parameters of the to-be-registered image are consistent with those of the reference image. The correction of the direction needs to satisfy that the direction of the to-be-registered image is consistent with that of the reference image.

[0020] S2. When the magnification states of the pre-processed to-be-registered image and the reference image are the same, select the positions of the same features as homonymic control points from the reference image and the pre-processed to-be-registered image, respectively, distribute the selected homonymic control points according to the distribution density judgment standard, and then preliminarily register according to the similarity or geometric relationship between the features to obtain a preliminary registration result. The features corresponding to the homonymic control points need to satisfy that the geographic coordinates, shape, and size are fixed.

[0021] S3. Review the preliminary registration result according to the image correction standard. If the review is passed, the preliminary registration result is taken as the final registration result and output, and the final registration result after the review is used to analyze and identify the archaeological site points by interpretation method, otherwise, the preliminary registration result is optimized by the local readjustment method of calibrating the homonymic control points until the review is passed.

[0022] The specific process of the local readjustment method is as follows: for the preliminary registration result, first check whether the selected homonymic control points are correct. If correct, the selected homonymic control points are retained, and if incorrect, the incorrect homonymic control points are deleted and then reselected, and new homonymic control points are added in the local area where the features deviate, until all the selected homonymic control points are correct.

[0023] On the basis of the above-mentioned scheme, each step can be realized in the following preferred specific manner.

[0024] As a preferred, in step S1, the reference image is the remote sensing image that is closest to the to-be-registered image in terms of age and resolution.

[0025] As preferred, in step S2, the distribution density judgment criterion is as follows: the first distance range, the second distance range and the third distance range of the preset numerical range are sequentially decreased, and it is pre-judged whether the same name control point corresponds to a mountain or a plain: if it is a mountain, one selected point is set in the first distance range; if it is a plain, one selected point is set in the second distance range, and the plain is divided into several plain areas, each of which is preset with a respective dense degree threshold value, and the total number of dense lines in each plain area is compared with the dense degree threshold value in sequence: if it is greater than or equal to, one selected point is set in the third distance range, otherwise, no processing is needed, until all the plain areas are traversed, and the dense line is a river, a road and an edge splicing place of the image to be registered.

[0026] As preferred, the first distance range is 4.4-6.2 square kilometers, the second distance range is 3.2-4.4 square kilometers, and the third distance range is 1.5-3.2 square kilometers.

[0027] As preferred, in step S2, the preliminary registration result is derived by an image registration software.

[0028] As preferred, the image registration software adopts Global Mapper.

[0029] As preferred, in step S3, the image correction standard is whether the preliminary registration result is completely coincided with the reference image, and whether there is a blurred or ghosted part in the preliminary registration result.

[0030] As preferred, in step S2, the same name control point is selected at a color transition or a line turning point.

[0031] As preferred, in step S2, the same name control point is selected at a road intersection, a bridge midpoint, a river intersection, a building corner or a dike edge corner.

[0032] Compared with the prior art, the present application has the following beneficial effects:

[0033] 1) Social effect

[0034] The method can provide more information and basis for archaeology, help archaeologists reveal the living conditions, cultural characteristics, social organization and historical changes of ancient humans, thereby enriching and perfecting the historical memory of human beings, carrying forward the excellent traditional Chinese culture and enhancing the national pride and cultural self-confidence.

[0035] 2) Economic effect

[0036] The method can save time and manpower of archaeological investigation, reduce the cost of large-area archaeological investigation, improve the efficiency and effect of archaeological investigation, provide a basis for archaeological excavation and protection, and create conditions for the development of cultural tourism and cultural industry.

[0037] 3) Technical effects

[0038] The method can promote the application and innovation of remote sensing technology in the field of archaeology, provide a model case of using historical image registration to assist archaeological investigation for practitioners in the archaeological industry, and clarify the basic principles contained therein, thereby guiding the conduct of other similar archaeological work, improving the quality and quantity of archaeological data, enhancing the analysis and interpretation capabilities of archaeological data, and promoting the development of archaeology. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a flowchart of the present application;

[0040] Figure 2 is a schematic diagram of the divided image in the CORONA DS1105-1086DA130 of the present embodiment;

[0041] Figure 3 is a schematic diagram of the spliced image of the present embodiment;

[0042] Figure 4 is a schematic diagram of the present embodiment for determining the angle of rotation of the image to be registered by comparing the reference image;

[0043] Figure 5 is a comparison diagram of the present embodiment before and after color adjustment of the image to be registered;

[0044] Figure 6 is a schematic diagram of the present embodiment for selecting the same name control points at the bridge;

[0045] Figure 7 is a schematic diagram of the present embodiment for selecting the same name control points at the road intersection;

[0046] Figure 8 is a schematic diagram of the present embodiment for selecting the same name control points;

[0047] Figure 9 is a comparison diagram of the present embodiment for selecting the same name control points in the mountainous area, near the image edge, and at the edge;

[0048] Figure 10 is a schematic diagram of the registration result of the present embodiment;

[0049] Figure 11 is a schematic diagram of the present embodiment for adjusting the area that does not completely coincide after superimposing the registration result and the reference image. DETAILED DESCRIPTION

[0050] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. Technical features in the various embodiments of the present invention can be combined accordingly without mutual conflict.

[0051] In a preferred embodiment of the present invention, a method for assisting archaeological surveys using historical remote sensing image registration is provided. The purpose is to summarize and generalize the team's many leading and successful experiences in assisting archaeological surveys using historical remote sensing image registration, and to propose a method and detailed processing guidelines for assisting archaeological surveys using historical remote sensing image registration. This aims to save more time and energy for large-scale archaeological surveys that currently rely mainly on methods such as field surveys, improve the efficiency and accuracy of archaeological work, and allow archaeologists to devote more energy to more meaningful archaeological research.

[0052] like Figure 1 As shown, in a preferred embodiment of the present invention, the method for assisting archaeological surveys by registering historical remote sensing images includes the following steps S1 to S3. The specific implementation process of each step will be described in detail below.

[0053] S1. Obtain historical remote sensing images used for archaeological surveys as images to be registered, obtain reference images corresponding to the range of images to be registered, and preprocess the images to be registered with reference to the reference images to obtain preprocessed images to be registered.

[0054] The preprocessing includes stitching and merging the images to be registered, correcting the direction, and optimizing the color. The stitching and merging must simultaneously ensure that the positional order of the images to be registered in the partition is correct and that all ground features in the overlapping part of the boundary line of the images to be registered in adjacent partitions are completely overlapped. The color optimization must ensure that the color parameters of the images to be registered are consistent with those of the reference image. The direction correction must ensure that the direction of the images to be registered is consistent with that of the reference image.

[0055] It should be noted that in step S1 of the present invention, the reference image is a remote sensing image that is closest to the preprocessed image to be registered in terms of both age and resolution.

[0056] It should be noted that in step S1 of this invention, the preprocessing includes image stitching and merging, orientation correction, and color optimization, and must comply with the following basic principles of image preprocessing:

[0057] 1) The splicing and merging process should ensure that the position sequence of the partitioned images is correct;

[0058] 2) The splicing and merging process needs to be precisely aligned at the boundary intersection line of adjacent partitioned images in an enlarged state, to ensure that all ground features in the overlapping part are completely overlapped;

[0059] 3) The optimization toning process saves the toning parameter settings of the first image in the image sequence, and the toning parameters of the remaining images are consistent with those of the first image, and need to refer to the toning parameters of the reference image;

[0060] 4) The optimization toning process adjusts the brightness and contrast of the image to ensure that the color contrast of the characteristic ground features with clear boundaries is clear and the control points are prominent.

[0061] The specific process of step S1 of the present application is described in detail below in combination with the drawings.

[0062] It should be noted that in the process of splicing and merging the same image, the currently available historical remote sensing images are too large in file size when scanned, and the images in a certain area A at the time of shooting are cut into N images (A1-A N) according to certain rules. Therefore, for such cut images, the image processing software (such as PS) is used to splice the A1-A N images in a certain area A to form a spliced image in a certain area A before registration. For a small amount of original historical remote sensing images that have not been cut, splicing is not required.

[0063] Specifically, the present embodiment uses CORONA satellite images, and the DS1105-1086DA series has a total of 30 images, DS1105-1086DA130-DS1105-1086DA160, each of which has four local images, for example, DS1105-1086DA130_a, DS1105-1086DA130_b, DS1105-1086DA130_c, and DS1105-1086DA130_d in the DS1105-1086DA130 image. The four local images in the DS1105-1086DA130 image are numbered as shown in Figure 2 The process of splicing the historical remote sensing images is to adjust the size of the images using PS software, and then merging the spliced images, which must meet the above-mentioned basic image preprocessing criteria to ensure accuracy. Then the remaining two local images are merged in turn to obtain a complete spliced image as shown in Figure 3 .

[0064] It should be noted that in the process of correcting the image direction, as shown in Figure 1As shown in the image preprocessing part, the actual coverage range and the tilt angle of the image to be registered are determined by comparing the ground objects in the image to be registered and the reference image. Since the images before registration are displayed in the north-south direction, they are usually inconsistent with the direction of the actual coverage range, which usually causes a large registration error. Therefore, the image to be registered after the splicing and merging in the previous step needs to be rotated accordingly to ensure its consistency with the direction of the reference image. In addition, for the N images cut during the scanning process, the rotation process should not be omitted after splicing, otherwise it will affect the accuracy of the registration stage and cause irregular deformation even if the splicing point is very dense.

[0065] In this embodiment, the position of the DS1105-1086DA130 image in the 91 map is determined and identified by comparing the ground objects (the geographical range of the image can also be directly downloaded from USGS). At this time, it is found that the image to be registered needs to be rotated by 180 degrees, and the north angle of the image is observed to determine that the angle needs to be rotated by 6 degrees counterclockwise. For such a large range of images, if the appropriate angle is not rotated, the splicing point of the registration will be irregularly deformed even if the splicing point is very dense, as shown in Figure 4 .

[0066] It should be noted that in the process of optimizing the color of the image, the brightness and contrast of the rotated image to be registered need to be adjusted to meet the color contrast requirements of highlighting characteristic ground objects and target ranges. It should be noted that in the color optimization step of the present application, only the first image is color adjusted, and when color adjusting other images in the sequence, only the adjustment parameters of the first image need to be referred to. In addition, it should be noted that by adjusting the brightness and contrast, the requirement of "as many characteristic ground objects and landforms as possible in the target range with clear color contrast and prominent control points" should be met.

[0067] In this embodiment, the process of adjusting the image color needs to meet the above-mentioned image preprocessing basic criteria, and the brightness and contrast of the image need to be adjusted for easy identification of traces. Taking DS1105-1086DA130_b as an example, after the color adjustment operation (brightness reduction of 92, contrast increase of 57), it can be found that the mountains, rivers, roads, and villages after adjustment are more obviously prominent, as shown in Figure 5 .

[0068] S2. When the magnification states of the pre-processed image to be registered and the reference image are the same, the positions of the same ground objects in the reference image and the pre-processed image to be registered are selected as homonymous control points, the selected homonymous control points are distributed according to the distribution density judgment standard, after distribution, the homonymous control points are preliminarily registered according to the similarity or geometric relationship between the ground objects, and the preliminary registration result is obtained, and the ground objects corresponding to the homonymous control points need to meet the fixed geographical coordinates, shape and size.

[0069] It should be noted that in step S2 of the present application, the preliminary registration result can be derived from image registration software.

[0070] It should be noted that in step S2 of the present application, the same name control point selects the color transition or line turning point.

[0071] It should be noted that in step S2 of the present application, the same name control point selects the road intersection, bridge midpoint, river intersection, building corner, and dam edge corner.

[0072] It should be noted that in step S2 of the present application, when selecting the same name control point, the following basic principles for extracting the control point need to be met:

[0073] 1) In the same magnification state of the pre-processed to-be-registered image and the reference image.

[0074] 2) Follow the principles of obvious ground features, fixed geographical location, no seasonal changes over time, and accurate positioning to ensure that the selected same name control point corresponds to the ground object that meets the fixed geographical coordinates, shape, and size.

[0075] 3) According to the actual geographical area size covered by the pre-processed to-be-registered image, the points are distributed to meet the point distribution density determination standard.

[0076] It should be noted that in step S2 of the present application, the point distribution density determination standard is as follows: the first distance range, the second distance range, and the third distance range of the preset numerical range are sequentially decreased, and it is pre-judged whether the ground object corresponding to the same name control point is a mountain or a plain: if it is a mountain, one selected point is set in the first distance range; if it is a plain, one selected point is set in the second distance range, and the plain is divided into several plain areas, each of which is preset with a respective density threshold, and the size of the total number of dense lines in each plain area and the density threshold is compared in turn: if it is greater than or equal to, one selected point is set in the third distance range, otherwise no processing is needed, until all plain areas are traversed, and the dense line is a river, a road, and an edge splicing of the to-be-registered image.

[0077] In this embodiment, for large area mountains, follow the principle of relative uniformity, directly set one selected point in the preset first distance range; for gentle terrain, follow the principle of local first-level encryption, first set one selected point in the preset second distance range, and further, for the dense distribution of rivers and roads in the gentle terrain, and the historical remote sensing image edge splicing, follow the principle of local second-level encryption, set one selected point in the preset third distance range.

[0078] The following describes in detail, with reference to the accompanying drawings, the specific process of step S2 of the present invention, the basic principles for extracting control points, and the criteria for determining the density of control points.

[0079] In this invention, the initial registration process requires reference imagery and preprocessed imagery to be registered. Therefore, it is first necessary to acquire reference imagery covering a range corresponding to the imagery to be registered. The principle for acquiring reference imagery is that its age and resolution are closest to the imagery to be registered. Reference imagery can be acquired through imagery platforms with offset-free coordinates, such as Google Maps, 91 Satellite Imagery, or Tianditu. Note that the range of the downloaded reference imagery should be slightly larger than the marked range to avoid insufficient base map area during registration, which could prevent the selection of control points in certain areas. In this embodiment, as... Figure 1 As shown in the image registration section, preprocessed historical satellite images and downloaded reference satellite remote sensing images are imported, and subsequent registration processing is performed.

[0080] During the registration process, the selection of corresponding control points must comply with the basic principles of control point extraction:

[0081] 1) When selecting control points with the same name, the selection should be performed under the same magnified state as the preprocessed image to be registered and the reference image to ensure accuracy.

[0082] Due to the large coverage area and complex land cover types involved, an appropriate small scale should be selected to find the same coverage area and the approximate location of the same land cover on the reference image and the preprocessed image to be registered. Based on this, the two images are simultaneously enlarged to accurately determine the location of the same control points at a large scale. However, the scale should not be too large at this time, with the standard being that there is no mosaic of pixel units.

[0083] In this embodiment, using a 1.8-meter resolution CORON image as an example, it is advisable to first locate the same coverage area and feature locations on the reference image and the preprocessed image to be registered at a scale of 1:200. Then, simultaneously zoom in on both images at a scale of 1:20 to accurately locate the control points to be selected. In particular, care should be taken not to zoom in too much. For example, when the scale is zoomed in to 1:10, the image pixels will be displayed as a mosaic raster, which will make the image blurry and the location difficult to determine.

[0084] 2) When selecting control points with the same name, the principles of "obvious ground features, fixed geographical location that does not change with time and season, and accurate positioning" should be followed.

[0085] The same name control point should be selected at the color transition or line turning point which can be accurately positioned. Preferably, the position with relatively fixed and color prominent position such as road intersection, bridge midpoint, river intersection, building corner, dike edge corner and the like should be selected in priority, and the position such as river with large curvature (easy to change course and swing with time and season), mountain (characteristic is fuzzy and difficult to accurately position), farmland (position and size are easy to change) and the like should be avoided.

[0086] 3) When selecting the same name control point, the points are arranged according to the area size of the pre-processed image to be registered, and the arrangement density meets the arrangement density determination standard to ensure the relative uniformity and reasonable local encryption.

[0087] The arrangement density determination standard of the same name control point should adopt the following suggestions:

[0088] 1) Relative uniformity: for large area mountain, one point is arranged in the preset first distance range. Specifically, there is one point for every 4.4-6.2 square kilometers, which is suitable for the arrangement of large area mountain.

[0089] 2) Local first encryption: for flat terrain, one point is arranged in the preset second distance range. Specifically, there is one point for every 3.2-4.4 square kilometers, which is suitable for the arrangement of flat terrain such as mountain and plain junction, plain, farmland and the like.

[0090] Preferably, when the point within 1 km from the boundary of the spliced boundary is selected, in addition to the local first encryption of the arrangement density, the control point close to the boundary edge should be selected in priority to ensure the accuracy.

[0091] 3) Local second encryption: for the edge splicing of historical remote sensing image and the dense distribution of river and road, one point is arranged in the preset third distance range. Specifically, there is one point for every 1.5-3.2 square kilometers, which is suitable for the arrangement of the edge splicing of map and the dense distribution of river and road.

[0092] Preferably, when the point within 0.2 km from the boundary of the spliced boundary is selected, in addition to the local second encryption of the arrangement density, the intersection of each river, each road and the boundary should be selected as the control point in priority.

[0093] In this embodiment, during the initial registration process, images near the study area downloaded from 91 Satellite Image Assistant are directly imported into image registration software (e.g., Global Mapper) at level 15 accuracy as reference images. After importing the historical remote sensing images to be registered, the initial registration begins. Note that the registration process in this embodiment must strictly follow the basic principles of control point extraction described above. Topographical features with minimal variations, such as road intersections, bridges, and building corners, should be selected, avoiding locations that cannot be precisely located, such as river bends or mountain peaks. If a bridge is encountered, the midpoint of the bridge's connecting line should be selected because river levels vary with the seasons, making it impossible to accurately locate the riverbank, while the midpoint of the bridge is fixed. Figure 6 As shown, selecting a point in the middle of the bridge and then adding it to the list creates a control point with the same name. When selecting a point, the image to be registered and the reference image need to be enlarged to the same size for easier selection. Additionally, selecting road intersections is also a good option, such as... Figure 7 As shown. After successfully selecting a point, if you need to fine-tune its position, you can double-click the point you want to move. The point will turn yellow, indicating that you can now modify it. After modification, add the point to the list. If you find the position might be incorrect after successfully selecting a point, you can double-click the point; it will turn yellow and you can then delete it.

[0094] Distribution density diagram as shown Figure 8 As shown. Based on the above basic principles for extracting control points, the selection of control points needs to ensure relative uniformity and reasonable local densification, and the distribution of points should be reasonable according to the area size of the registered image. For example, since most of the image consists of large mountain areas, a relatively uniform control point density should be selected (with an average of one point every 4.4 to 6.2 square kilometers), resulting in 686 points. The perimeter is 499 kilometers, the area is 4224 square kilometers, and the average distribution is one point every 6.15 square kilometers. In particular, when selecting points at the stitching points, they should be placed closer to the image edge and slightly denser than in other areas. In this embodiment, as... Figure 9 As shown, when placing points within 1 km of the stitched boundary, the following density should be adopted: ② Local Level 1 Densification (one selected point per 3.2–4.4 square kilometers on average); when placing points within 0.2 km of the stitched boundary, the density should meet ③ Local Level 2 Densification (one selected point per 1.5–3.2 square kilometers on average). Mountains in the image exhibit elevation differences, resulting in significantly greater deformation than plains. To improve registration accuracy, additional points should be added near valleys and mountain foothills during registration, conforming to ② Local Level 1 Densification (one selected point per 3.2–4.4 square kilometers on average).

[0095] S3. Verify the preliminary registration results according to the image verification standards: If the verification is successful, the preliminary registration results will be used as the final registration results and output. The archaeological site points will be identified by interpretation using the final registration results after verification. Otherwise, the preliminary registration results will be optimized by local readjustment of the corresponding control points until the verification is successful.

[0096] The specific process of the local readjustment method is as follows: For the preliminary registration results, first check whether the selected corresponding control points are correct: if correct, retain the selected corresponding control points; if incorrect, delete the incorrect corresponding control points and reselect, and add new corresponding control points in the local areas where ground feature deviation occurs, until all selected corresponding control points are correct.

[0097] In step S3 of this invention, the image calibration standard is to check whether the preliminary registration result completely overlaps with the reference image, and whether there are blurred or ghosted parts in the preliminary registration result.

[0098] The following describes in detail the verification process and local readjustment method of step S3 of the present invention with reference to the accompanying drawings.

[0099] It should be noted that, in this invention, as Figure 1 As shown in the image verification section, after exporting the preliminary registration results, they need to be reviewed according to image verification standards to determine whether local readjustment is required. If the preliminary registration results completely overlap with the reference image and there are no blurred or ghosting areas, the verification is considered successful, and the preliminary registration results are output as the final registration results. Conversely, if the preliminary registration results do not completely overlap with the reference image and there are blurred or ghosting areas, the verification is considered unsuccessful, and optimization is required according to local readjustment methods (including reselection and densification of control points). In this embodiment, the transparency of the historical remote sensing image is first adjusted to 50% on the Global Mapper page, and any areas requiring adjustment are checked according to the image verification standards. Figure 10 As shown, the preliminary registration results are clear, with roads, rivers, villages, etc., all corresponding to each other, indicating that the registration results are accurate. Figure 11 As shown, the presence of ghosting and significant road offset indicates inaccurate registration. This is likely due to incorrect nearby point selection or insufficient point density, necessitating recalibration. In addition to adjusting transparency, a roll-up method can be used to check the initial registration results by comparing the rolled-up image with the reference image to check for terrain shifts.

[0100] When optimizing according to the local readjustment method, specifically, first, check whether the selected control points are correct. For example, sometimes a road or a hilltop is selected by mistake due to similar terrain. If this occurs, the incorrect control points need to be deleted and then reselected. When there is only a slight deviation, the existing control points only need to be slightly adjusted to ensure that each type of ground object is correctly matched. After the checking process is completed, the control points in the local area are further encrypted. For example, when some ground objects have a significant deviation, but the nearby selected control points are accurate, it means that the image deformation in the local area is irregular, and there may be local distortion or compression, stretching, etc. Therefore, the control points in the area need to be encrypted, that is, additional control points are added to the ground objects with significant deviations. Then, a new preliminary registration result is exported, and the rechecking is performed again. The above rechecking and judgment steps are repeated until the preliminary registration result passes the rechecking. At this time, the preliminary registration result that passes the rechecking is taken as the final registration result and is output. In the archaeological survey, the preliminary registration result that passes the rechecking can be analyzed and identified by the visual interpretation method or computer interpretation by the archaeologists, and is used for subsequent archaeological survey and analysis.

[0101] The above-described embodiments are only a preferred scheme of the present application, and are not intended to limit the present application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, any technical scheme obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present application.

Claims

1. A method for assisting archaeological surveys using historical remote sensing image registration, characterized in that, Includes the following steps: S1. Obtain historical remote sensing images used for archaeological surveys as images to be registered, obtain reference images corresponding to the range of images to be registered, and preprocess the images to be registered with reference to the reference images to obtain preprocessed images to be registered. The preprocessing includes stitching and merging the images to be registered, correcting the direction, and optimizing the color. The stitching and merging must simultaneously ensure that the positional order of the images to be registered in the partition is correct and that all ground features in the overlapping part of the boundary line of the images to be registered in adjacent partitions are completely overlapped. The color optimization must ensure that the color parameters of the images to be registered are consistent with the reference image. The direction correction must ensure that the direction of the images to be registered is consistent with the direction of the reference image. S2. When the preprocessed image to be registered and the reference image are in the same magnification state, select the locations of the same ground features from the reference image and the preprocessed image to be registered as corresponding control points. Place the selected corresponding control points according to the point density judgment standard. After placement, perform preliminary registration based on the similarity or geometric relationship between ground features to obtain the preliminary registration result. The ground features corresponding to the corresponding control points must satisfy that the geographic coordinates, shape and size are fixed. S3. Verify the preliminary registration results according to the image verification standards: If the verification is successful, the preliminary registration results will be used as the final registration results and output. The archaeological site points will be identified by interpretation using the final registration results after verification. Otherwise, the preliminary registration results will be optimized by local readjustment of the corresponding control points until the verification is successful. The specific process of the local readjustment method is as follows: For the preliminary registration results, first check whether the selected corresponding control points are correct: if correct, retain the selected corresponding control points; if incorrect, delete the incorrect corresponding control points and reselect, and add new corresponding control points in the local areas where ground feature deviation occurs, until all selected corresponding control points are correct.

2. The method for archaeological surveys using historical remote sensing image registration as described in claim 1, characterized in that, In step S1, the reference image is a remote sensing image that is closest to the image to be registered in terms of both age and resolution.

3. The method for archaeological surveys using historical remote sensing image registration as described in claim 1, characterized in that, In step S2, the criteria for determining the density of control points are as follows: a first distance range, a second distance range, and a third distance range with successively decreasing preset numerical ranges are used, and it is pre-determined whether the land feature corresponding to the same control point is a mountain or a plain: if it is a mountain, a selected point is set in the first distance range; if it is a plain, a selected point is set in the second distance range, and the plain is divided into several plain areas. A density threshold is preset for each plain area, and the total number of dense lines in each plain area is compared with the density threshold: if it is greater than or equal to the threshold, a selected point is set in the third distance range; otherwise, no processing is required, until all plain areas have been traversed. The dense lines are rivers, roads, and the edge stitching of the image to be registered.

4. The method for archaeological surveys using historical remote sensing image registration as described in claim 3, characterized in that, The first distance range is 4.4 to 6.2 square kilometers, the second distance range is 3.2 to 4.4 square kilometers, and the third distance range is 1.5 to 3.2 square kilometers.

5. The method for archaeological surveys using historical remote sensing image registration as described in claim 1, characterized in that, In step S2, the preliminary registration results are exported by the image registration software.

6. The method for archaeological surveys using historical remote sensing image registration as described in claim 5, characterized in that, The image registration software used is Global Mapper.

7. The method for archaeological surveys using historical remote sensing image registration as described in claim 1, characterized in that, In step S3, the image calibration standard is to check whether the preliminary registration result completely overlaps with the reference image, and whether there are blurred or ghosted parts in the preliminary registration result.

8. The method for archaeological surveys using historical remote sensing image registration as described in claim 1, characterized in that, In step S2, select the control point with the same name at the color change point or the line bend point.

9. A method for archaeological surveys using historical remote sensing image registration as described in claim 8, characterized in that, In step S2, the corresponding control points are selected as road intersections, bridge midpoints, river intersections, building corners, and dam edge corners.

Images