Image semantic change intelligent detection system

Abstract

The invention relates to the field of image detection, in particular to an image semantic change intelligent detection system. The consistency parameter of each unit is determined through the texture analysis module, the light and shadow interference characteristics of each unit are analyzed through the light and shadow analysis module, the tendency analysis module distinguishes the tendency of each unit to cause irrelevant change of remote sensing data, and the remote sensing data is adaptively screened through the change detection module. Different analysis modes are adopted for the screened remote sensing data and the unscreened remote sensing data, so that the data analysis quantity of the remote sensing data is reduced, the reliability of change detection is ensured, and the change detection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of image detection, and particularly to an intelligent detection system for image semantic change. Background Art

[0002] Satellite remote sensing technology has been developed for many years and now has the characteristics of multiple platforms, multiple sensors, and multiple angles. The obtained remote sensing images have the characteristics of high spectral resolution, high radiation resolution, high spatial resolution, and high temporal resolution. Using these image data, change detection of surface targets and natural and non-natural phenomena can be realized, so as to better solve the actual problems.

[0003] Chinese Patent Publication No.: CN110414343B, discloses a method for detecting ships in on-orbit satellite remote sensing images. The disclosed method uses superpixels to segment high-resolution remote sensing images and achieve target pre-positioning on the premise of ensuring the integrity of the target, and uses local features and variable threshold nearest neighbor verification to implement the target detection algorithm in the pre-positioning unit. Under the performance constraints and real-time requirements of the spaceborne platform, the operation efficiency of the algorithm of the present invention can meet the real-time detection requirements of the spaceborne platform and can accurately detect large ship targets in high-resolution remote sensing images.

[0004] However, the following problems still exist in the prior art: The remote sensing data faced during current change detection is massive, and analysis requires a large amount of computing power. Moreover, some changes in the remote sensing data are not substantial changes and may be caused by atmospheric interference or light changes, resulting in low change detection efficiency and poor accuracy. Summary of the Invention

[0005] Therefore, the present invention provides an intelligent detection system for image semantic change to overcome the problems in the prior art that analyzing massive remote sensing data requires a large amount of computing power, and some changes in the remote sensing data may be caused by atmospheric interference or light changes, resulting in low change detection efficiency and poor accuracy.

[0006] To solve the above problems, the present invention provides an intelligent detection system for image semantic change, including: A texture analysis module, which is used to extract the color histogram and contour features of historical remote sensing data in each unit, and analyze the visual consistency parameter in each unit according to the difference amount of the color histogram and the distribution difference of the contour features; A light and shadow analysis module, which is used to determine the light and shadow specific time domain segments based on the light and shadow features of different historical time nodes in each unit, and analyze the light and shadow interference parameters in each unit according to the changes of the historical remote sensing data in the light and shadow specific time domain segments; The tendency analysis module, which is respectively connected to the texture analysis module and the light and shadow analysis module, is used to determine the remote sensing change excitation features based on the visual consistency parameter and the light and shadow interference parameter, so as to distinguish the tendency of each unit to cause irrelevant changes in remote sensing data; The change detection module is used to screen the remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data, and analyze the screened remote sensing data, including, extracting the dual-temporal remote sensing data corresponding to the unit, inputting the dual-temporal remote sensing data into a pre-trained change recognition model, and determining whether the dual-temporal remote sensing data has undergone substantial changes based on the change recognition model; wherein, the temporal interval of the dual-temporal remote sensing data is predetermined.

[0007] Further, the texture analysis module is used to determine the difference amount of the color histogram and the distribution difference of the contour features. Among them, the texture analysis module determines the difference amount based on the Euclidean distance between the color histograms of the corresponding historical remote sensing data at each moment within the unit; the texture analysis module determines the average distribution distance of each contour feature based on the corresponding historical remote sensing data at each moment within the unit, and determines the distribution difference of the contour features based on the average distribution distance.

[0008] Further, the texture analysis module is used to calculate the visual consistency parameter, including, determining the ratio of the difference amount to a predetermined difference amount threshold as the color consistency component; determining the ratio of the distribution difference to the distribution difference threshold as the distribution consistency component; weighted summing the color consistency component and the distribution consistency component to obtain the visual consistency parameter.

[0009] Further, the light and shadow analysis module determines the light and shadow specific time domain segment based on the light and shadow features at different historical time nodes within each unit, including: obtaining the light and shadow features of the unit at different historical time nodes, including the solar altitude angle, the solar azimuth angle, and the atmospheric transmittance; identifying the mutation time period according to the change of the light and shadow features in the time domain dimension; merging the continuous mutation time periods to form a continuous time period, which is defined as the light and shadow specific time domain segment.

[0010] Further, the process of the light and shadow analysis module analyzing the light and shadow interference parameters of each unit based on the changes of the historical remote sensing data within the light and shadow specific time domain segment includes: respectively obtaining the historical remote sensing data of the unit at the start time and the end time of the light and shadow specific time domain segment, and determining the extracted chromaticity change amount and the contour area change amount; To determine the light and shadow interference parameters based on the chromaticity change amount and the contour area change amount.

[0011] Further, the tendency analysis module is used to determine the remote sensing change excitation features based on the visual consistency parameter and the light and shadow interference parameter, including, To perform a weighted sum of the visual consistency parameter and the light and shadow interference parameter to obtain the remote sensing change excitation feature.

[0012] Further, the tendency analysis module is used to distinguish the tendency of each unit to cause irrelevant changes in the remote sensing data, where, If the remote sensing change excitation feature corresponding to the unit is greater than or equal to the predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in the remote sensing data is the first tendency; If the remote sensing change excitation feature corresponding to the unit is less than the predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in the remote sensing data is the second tendency.

[0013] Further, the change detection module filters the remote sensing data based on the tendency of each unit to cause irrelevant changes in the remote sensing data, where, Filter out the remote sensing data corresponding to the units with the first tendency.

[0014] Further, the change detection module is also used to analyze the unfiltered remote sensing data, including, To extract the difference features between the unfiltered remote sensing data and the basic remote sensing data, including the difference amount of the color histogram and the distribution difference of the contour features; If the difference condition is satisfied, it is determined that there is a substantial change in the unit corresponding to the remote sensing data; Among them, the difference condition includes that the difference amount is greater than the predetermined sample difference amount threshold and the distribution difference is greater than the predetermined sample distribution difference threshold.

[0015] Further, the change detection module is used to call historical remote sensing data based on the remote sensing data and the time phase interval, and combine the remote sensing data with the historical remote sensing data to form dual-time remote sensing data.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows. The present invention determines the consistency parameter of each unit through the texture analysis module, analyzes the light and shadow interference characteristics of each unit through the light and shadow analysis module, the tendency analysis module distinguishes the tendency of each unit to cause irrelevant changes in the remote sensing data, and adaptively filters the remote sensing data through the change detection module. Different analysis methods are adopted for the filtered remote sensing data and the unfiltered remote sensing data. Furthermore, the amount of data analysis of the remote sensing data is reduced, the reliability of change detection is ensured, and the change detection efficiency is improved.

[0017] In particular, the texture analysis module of the present invention determines the visual consistency parameter. In actual situations, the geographical features in each unit are different, and their tendencies to be affected by the atmosphere and light and shadow features are different. Furthermore, the tendency to generate irrelevant changes in the corresponding remote sensing images is also different. Based on this, the present invention considers the difference in the color histograms of the historical remote sensing data corresponding to the recognition units and the distribution differences in the contour features. In actual situations, the geographical features of some units are simple, such as plains and inland seas. In their historical remote sensing data, the colors are visually single and have little variability. Therefore, the color histogram is used to reflect the visual color variability of the historical remote sensing data corresponding to the unit. Moreover, the contour features of the remote sensing data corresponding to the unit to a certain extent reflect the entity contours or light and shadow contours existing within the unit. The distribution changes in the contour features can to a certain extent reflect the changes in the light and shadow contours generated by the entity after being illuminated by light, and thus reflect the tendency to generate irrelevant changes in the remote sensing data. If the geographical features of the unit are simple, the influence of light and shadow illumination on its remote sensing data is small, and the corresponding distribution differences in its contour features are small. Based on this, the visual consistency parameter is calculated, providing a basis for determining the remote sensing change excitation features subsequently, facilitating the adaptive analysis of remote sensing data, thereby reducing the amount of data analysis of remote sensing data, ensuring the reliability of change detection, and improving the change detection efficiency.

[0018] In particular, the light and shadow features at different historical time nodes within each unit will change rapidly within a certain period of time. By analyzing the remote sensing data before and after the rapid change, it can better reflect the influence of the light and shadow change on the remote sensing data within the unit. The present invention determines the light and shadow specific time domain segments, analyzes the changes in the historical remote sensing data, and then determines the light and shadow interference parameters, which can reflect the tendency of the remote sensing data within the unit to generate irrelevant changes due to light and shadow illumination, and have strong data representativeness, providing data support for calculating the remote sensing change excitation features subsequently, facilitating the adaptive analysis of remote sensing data, thereby reducing the amount of data analysis of remote sensing data, ensuring the reliability of change detection, and improving the change detection efficiency.

[0019] In particular, the present invention considers and calculates the remote sensing change excitation features from two dimensions, considering the tendency of the remote sensing data corresponding to the unit to generate irrelevant changes. For the units with the first tendency, the tendency of the corresponding remote sensing data to generate irrelevant changes is relatively high, and a computationally intensive in-depth analysis method is adopted to ensure capturing the change features and avoiding omission. For the units with the second tendency, the tendency of the corresponding remote sensing data to generate irrelevant changes is relatively low, and the changes generated by them are mostly substantial changes. Therefore, the changes in their relatively basic image features can be directly used to identify whether substantial changes have occurred, thereby reducing the amount of data analysis of remote sensing data, ensuring the reliability of change detection, and improving the change detection efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic structural diagram of the intelligent detection system for image semantic changes in the invention embodiment; Figure 2 A logic block diagram for distinguishing the tendency of each unit in an invention embodiment to cause irrelevant changes in remote sensing data; Figure 3 A logic block diagram for screening remote sensing data based on the tendency of each unit in an invention embodiment to cause irrelevant changes in remote sensing data; Figure 4 A logic block diagram for analyzing the screened remote sensing data and the unscreened remote sensing data in an invention embodiment. Detailed implementation manners

[0021] In order to make the objectives and advantages of the present invention clearer and more understandable, the present invention will be further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0022] The preferred implementation manners of the present invention will be described below with reference to the accompanying drawings. Those skilled in the art should understand that these implementation manners are only used to explain the technical principles of the present invention and do not limit the protection scope of the present invention.

[0023] It should be noted that in the description of the present invention, the terms indicating directions or positional relationships such as "upper", "lower", "left", "right", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings. This is only for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.

[0024] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installation", "connection", and "connection" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0025] Please refer to Figure 1 as shown Figure 1 is a schematic structural diagram of an intelligent detection system for image semantic changes in an invention embodiment. The intelligent detection system for image semantic changes in the embodiment of the present invention includes: A texture analysis module, which is used to extract the color histogram and contour features of the historical remote sensing data in each unit, and analyze the visual consistency parameters in each unit based on the difference amount of the color histogram and the distribution difference of the contour features; A light and shadow analysis module, which is used to determine a light and shadow specific time domain segment based on the light and shadow characteristics at different historical time nodes within each unit, and analyze the light and shadow interference parameters of each unit according to the changes in historical remote sensing data within the light and shadow specific time domain segment; A tendency analysis module, which is respectively connected to the texture analysis module and the light and shadow analysis module, and is used to determine remote sensing change excitation characteristics based on visual consistency parameters and light and shadow interference parameters, so as to distinguish the tendency of each unit to cause irrelevant changes in remote sensing data; A change detection module, which is used to screen remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data, and analyze the screened remote sensing data, including, extracting the dual-temporal remote sensing data corresponding to the unit, inputting the dual-temporal remote sensing data into a pre-trained change recognition model, and determining whether the dual-temporal remote sensing data has undergone substantial changes based on the change recognition model; wherein, the time interval of the dual-temporal remote sensing data is determined in advance.

[0026] Specifically, the specific structures of the texture analysis module, the light and shadow analysis module, the tendency analysis module, and the change detection module are not limited, and can all be composed of logic components or combinations of logic components. The logic components include field programmable processors, computers, or microprocessors in computers.

[0027] Specifically, in implementation, a unit is a geospatial area, and its scope is equal to the image coverage range of the remote sensing image.

[0028] Specifically, in implementation, the remote sensing data is a remote sensing image obtained by a collection device, which will not be elaborated here.

[0029] Specifically, the form of the change recognition model is not limited. Those skilled in the art can select an open-source image processing model in the prior art that can recognize substantial changes in remote sensing data, or can also train a change recognition model that realizes the corresponding function by themselves. For example, a number of remotely sensed data with substantial changes are manually labeled in advance as training samples to train the corresponding change recognition model. The model architecture of the change recognition model is not limited. For example, a neural network architecture can be adopted, which will not be elaborated here.

[0030] Specifically, the texture analysis module is used to determine the difference amount of the color histogram and the distribution difference of the contour features, wherein, the texture analysis module determines the difference amount based on the Euclidean distance between the color histograms of the historical remote sensing data corresponding to each moment within the unit; It can be understood that there are multiple color histograms, and the Euclidean distances between the color histograms need to be calculated separately, and the average value of the Euclidean distances is determined as the difference amount; The texture analysis module determines the average distribution distance of each contour feature based on the corresponding historical remote sensing data at each moment within the unit, and determines the distribution difference of the contour feature based on the average distribution distance.

[0031] In implementation, the average distribution distance is the mean of the distances between the centers of each contour and the centers of the nearest contours; The variance of the average distribution distance corresponding to the historical remote sensing data at each moment is determined as the distribution difference.

[0032] In implementation, the color histogram of the RGB color system is adopted, and the color histogram includes sub-color histograms of three color channels; Therefore, when calculating the Euclidean distance between two color histograms, it is necessary to calculate the sub-Euclidean distances between the sub-color histograms of the corresponding color channels respectively, and take the mean of the sub-Euclidean distances as the Euclidean distance between the two color histograms.

[0033] Specifically, the texture analysis module is used to calculate the visual consistency parameter, including, used to determine the color consistency component as the ratio of the difference amount to the predetermined difference amount threshold; used to determine the distribution consistency component as the ratio of the distribution difference to the distribution difference threshold; used to perform a weighted sum of the color consistency component and the distribution consistency component to obtain the visual consistency parameter.

[0034] Specifically, in implementation, the weight of the color consistency component is 0.55, and the weight of the distribution consistency component is 0.45.

[0035] Specifically, the predetermined difference amount threshold and the distribution difference threshold are determined in advance. Among them, the historical remote sensing data within the corresponding units of deserts, oceans, and gobi are determined by those skilled in the art in advance, the difference amount and the distribution difference corresponding to each unit are determined, the mean value of the difference amount and the mean value of the distribution difference are solved, the difference amount threshold is set as the product of the mean value of the difference amount and the error offset coefficient, the distribution difference threshold is set as the product of the mean value of the distribution difference and the error offset coefficient, and the error offset coefficient is selected within the interval [1.15, 1.25].

[0036] The texture analysis module of the present invention determines the visual consistency parameter. In actual situations, the geographical features in each unit are different, and their tendencies to be affected by the atmosphere and light and shadow features are different. Furthermore, the tendency to generate irrelevant changes in the corresponding remote sensing images is also different. Based on this, the present invention considers the difference in the color histograms of the historical remote sensing data corresponding to the recognition units and the distribution differences of the contour features. In actual situations, the geographical features of some units are simple, such as plains and inland seas. In their historical remote sensing data, the colors are visually single and have little variability. Therefore, the color histogram is used to reflect the visual color variability of the historical remote sensing data corresponding to the unit. Moreover, the contour features of the remote sensing data corresponding to the unit to a certain extent reflect the entity contours or light and shadow contours existing within the unit. The distribution changes of the contour features can to a certain extent reflect the changes in the light and shadow contours generated by the entity after light and shadow irradiation, and thus reflect the tendency to generate irrelevant changes in the remote sensing data. If the geographical features of the unit are simple, the influence of light and shadow irradiation on its remote sensing data is small, and the corresponding distribution differences of its contour features are small. Based on this, the visual consistency parameter is calculated, providing a basis for determining the remote sensing change excitation features in the subsequent stage, facilitating the adaptive analysis of remote sensing data, thereby reducing the amount of data analysis of remote sensing data, ensuring the reliability of change detection, and improving the change detection efficiency.

[0037] Specifically, the light and shadow analysis module determines the light and shadow specific time period based on the light and shadow features at different historical time nodes within each unit, including for obtaining the light and shadow features of the unit at different historical time nodes, including the solar altitude angle, solar azimuth angle, and atmospheric transmittance; for identifying the mutation time period according to the changes in the light and shadow features in the time domain dimension; for merging consecutive mutation time periods to form a continuous time period, defined as the light and shadow specific time period.

[0038] In implementation, the change speed of the solar altitude angle, the change speed of the solar azimuth angle, and the change speed of the atmospheric transmittance are respectively determined within each time period, and the time period is selected within the interval [5 minutes, 10 minutes]; If the change speed of the solar altitude angle within the time period is greater than the predetermined altitude angle change speed threshold, or the change speed of the solar azimuth angle is greater than the predetermined azimuth angle change speed threshold, or the change speed of the atmospheric transmittance is greater than the predetermined transmittance change speed threshold, then the time period is determined as the mutation time period. [[ID=I5]]

[0039] The altitude angle change speed threshold is selected within 1.3 to 1.5 times the average value of the historical solar altitude angle change speed; The azimuth angle change speed threshold is selected within 1.3 to 1.5 times the average value of the historical solar azimuth angle change speed; The transmittance change speed threshold is selected within 1.1 to 1.3 times the average value of the historical atmospheric transmittance change speed.

[0040] Specifically, the solar altitude angle, solar azimuth angle, and atmospheric transmittance can be obtained based on the geographical remote sensing information of the area where the unit is located. For example, it can be directly retrieved from an authorized remote sensing information database, or the remote sensing information of the corresponding area can be collected in real time, which will not be elaborated here.

[0041] The light and shadow characteristics of different historical time nodes within each unit will change rapidly during some periods. By analyzing the remote sensing data before and after the rapid change, it can better reflect the impact of light and shadow changes on the remote sensing data within the unit. The present invention determines the specific time domain segment of light and shadow interference, analyzes the changes in historical remote sensing data, and then determines the light and shadow interference parameters, which can reflect the tendency of the remote sensing data within the unit to undergo irrelevant changes due to light and shadow irradiation, and has strong data representativeness. It provides data support for calculating the excitation characteristics of remote sensing changes, facilitates the adaptive analysis of remote sensing data, thereby reducing the amount of data analysis of remote sensing data, ensuring the reliability of change detection, and improving the efficiency of change detection.

[0042] Specifically, the process by which the light and shadow analysis module analyzes the light and shadow interference parameters of each unit based on the changes in historical remote sensing data within the specific time domain segment of light and shadow interference includes: To respectively obtain the historical remote sensing data of the unit at the start time and end time of the specific time domain segment of light and shadow interference, and determine the extracted chromaticity change amount and contour area change amount; To determine the light and shadow interference parameters based on the chromaticity change amount and the contour area change amount.

[0043] In implementation, calculate the ratio of the chromaticity change amount to the standard chromaticity change amount threshold to obtain the chromaticity light and shadow interference component; Calculate the ratio of the contour area change amount to the standard contour area change amount threshold to obtain the contour area light and shadow interference component; Weightedly sum the chromaticity light and shadow interference component and the contour area light and shadow interference component to obtain the light and shadow interference parameter; In implementation, the weight of the chromaticity light and shadow interference component is 0.35, and the weight of the contour area light and shadow interference component is 0.65.

[0044] The standard chromaticity change amount threshold and the standard contour area change amount threshold are determined in advance. Among them, a number of historical remote sensing data of the units within the specific time domain segment of light and shadow interference are obtained in advance, the chromaticity change amount and the contour area change amount are recorded, the mean value of the chromaticity change amount and the mean value of the contour area change amount for each unit are solved, the standard chromaticity change amount threshold is set as the product of the mean value of the chromaticity change amount and the error coefficient, the standard contour area change amount threshold is set as the product of the mean value of the contour area change amount and the error coefficient, and the error coefficient is selected within the interval [0.78, 0.84].

[0045] Specifically, the tendency analysis module is used to determine the remote sensing change excitation features based on the visual consistency parameter and the light and shadow interference parameter, including, weighted summing the visual consistency parameter and the light and shadow interference parameter to obtain the remote sensing change excitation features.

[0046] In implementation, the weights of both the visual consistency parameter and the light and shadow interference parameter are 0.5.

[0047] Please refer to Figure 2 shown Figure 2 as the logic block diagram for distinguishing the tendency of each unit to cause irrelevant changes in remote sensing data in the invention embodiment; the tendency analysis module is used to distinguish the tendency of each unit to cause irrelevant changes in remote sensing data, where, if the remote sensing change excitation feature corresponding to the unit is greater than or equal to the predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in remote sensing data is the first tendency; if the remote sensing change excitation feature corresponding to the unit is less than the predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in remote sensing data is the second tendency.

[0048] In implementation, the remote sensing change excitation feature threshold is selected within the interval [1.23, 1.36].

[0049] Specifically, please refer to Figure 3 shown Figure 3 as the logic block diagram for screening remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data in the invention embodiment, the change detection module screens the remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data, where, screen out the remote sensing data corresponding to the units with the first tendency.

[0050] In implementation, all the remote sensing data corresponding to the units with the first tendency screened out are summarized into the first data set, and all the remote sensing data corresponding to the units not screened out are summarized into the second data set.

[0051] Specifically, please refer to Figure 4 shown Figure 4 as the logic block diagram for analyzing the screened remote sensing data and the unscreened remote sensing data in the invention embodiment.

[0052] Specifically, the change detection module is further used to analyze the unscreened remote sensing data, including, extracting the difference features between the unscreened remote sensing data and the basic remote sensing data, including the difference amount of the color histogram and the distribution difference of the contour features; If the difference condition is satisfied, it is determined that there is a substantial change in the unit corresponding to the remote sensing data; Among them, the difference condition includes that the difference amount is greater than a predetermined sample difference amount threshold and the distribution difference is greater than a predetermined sample distribution difference threshold.

[0053] Specifically, the basic remote sensing data is the remote sensing data collected in advance within the unit for reference, aiming to consider the changes of subsequent remote sensing images based on this basic remote sensing data.

[0054] Specifically, the sample difference amount threshold and the sample distribution difference threshold are determined in advance. The remote sensing data of several units with substantial changes are calibrated by those skilled in the art in advance, and the difference characteristics of each remote sensing data relative to the basic remote sensing data are determined. Then, the average value of the difference amount of the color histogram and the average value of the distribution difference of the contour features are solved. The sample difference amount threshold is set as the product of the average value of the difference amount of the color histogram and the adjustment coefficient, and the sample distribution difference threshold is set as the product of the average value of the distribution difference of the contour features and the adjustment coefficient. The adjustment coefficient is selected within the interval [0.65, 0.85].

[0055] Specifically, the change detection module is used to call historical remote sensing data based on the remote sensing data and the time phase interval, and combine the remote sensing data with the historical remote sensing data to form dual-time-phase remote sensing data.

[0056] The time phase interval is a time period determined based on actual needs. Those skilled in the art can consider the changes of the remote sensing data relative to the historical remote sensing data at what time to set the time phase interval according to needs, which will not be elaborated here.

[0057] The present invention considers and calculates the remote sensing change excitation characteristics from two dimensions, and considers the tendency of the remote sensing data corresponding to the unit to generate irrelevant changes. For the units with the first tendency, the tendency of the corresponding remote sensing data to generate irrelevant changes is relatively high, and a more computationally intensive in-depth analysis method is adopted to ensure capturing the change characteristics and avoiding omission. For the units with the second tendency, the tendency of the corresponding remote sensing data to generate irrelevant changes is relatively low, and the changes generated by them are mostly substantial changes. Therefore, the changes of its relatively basic image features can be directly utilized to identify whether there are substantial changes, thereby reducing the amount of data analysis of the remote sensing data, ensuring the reliability of change detection, and improving the change detection efficiency.

[0058] So far, the technical solution of the present invention has been described in combination with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or replacements to the relevant technical features, and the technical solutions after these changes or replacements will fall within the protection scope of the present invention.

Claims

1. An intelligent detection system for image semantic changes, characterized in that, Including: A texture analysis module, which is used to extract the color histogram and contour features of historical remote sensing data within each unit, and analyze the visual consistency parameter within each unit based on the difference amount of the color histogram and the distribution difference of the contour features; A light and shadow analysis module, which is used to determine the light and shadow specific time domain segment based on the light and shadow features at different historical time nodes within each unit, and analyze the light and shadow interference parameters of each unit based on the changes in historical remote sensing data within the light and shadow specific time domain segment; A tendency analysis module, which is respectively connected to the texture analysis module and the light and shadow analysis module, and is used to determine the remote sensing change excitation feature based on the visual consistency parameter and the light and shadow interference parameter, so as to distinguish the tendency of each unit to cause irrelevant changes in remote sensing data; A change detection module, which is used to screen the remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data, and analyze the screened remote sensing data, Including, Extract the dual-temporal remote sensing data corresponding to the unit, input the dual-temporal remote sensing data into a pre-trained change recognition model, and determine whether the dual-temporal remote sensing data has a substantial change based on the change recognition model; Wherein, the time interval of the dual-temporal remote sensing data is pre-determined.

2. The intelligent detection system for image semantic change according to claim 1, characterized in that The texture analysis module is used to determine the difference amount of the color histogram and the distribution difference of the contour features, wherein, The texture analysis module determines the difference amount based on the Euclidean distance between the color histograms of the historical remote sensing data corresponding to each moment within the unit; The texture analysis module determines the average distribution distance of each contour feature based on the historical remote sensing data corresponding to each moment within the unit, and determines the distribution difference of the contour features based on the average distribution distance.

3. The intelligent detection system for image semantic change according to claim 2, wherein The texture analysis module is used to calculate the visual consistency parameter, including, Determining the color consistency component by taking the ratio of the difference amount to a predetermined difference amount threshold; Determining the distribution consistency component by taking the ratio of the distribution difference to the distribution difference threshold; Weighted summing the color consistency component and the distribution consistency component to obtain the visual consistency parameter.

4. The intelligent detection system for image semantic change according to claim 1, wherein The light and shadow analysis module determines the light and shadow specific time domain segment based on the light and shadow features at different historical time nodes within each unit, including: Obtaining the light and shadow features of the unit at different historical time nodes, including the solar altitude angle, the solar azimuth angle, and the atmospheric transmittance; Identifying the mutation time period according to the change of the light and shadow features in the time domain dimension; Merging the continuous mutation time periods to form a continuous time period, which is defined as the light and shadow specific time domain segment.

5. The intelligent detection system for image semantic change according to claim 4, characterized in that The process of the light and shadow analysis module analyzing the light and shadow interference parameters of each unit based on the changes in historical remote sensing data within the light and shadow specific time domain segment includes: Respectively obtaining the historical remote sensing data of the unit at the start time and the end time of the light and shadow specific time domain segment, and determining the chromaticity change amount and the contour area change amount; Determining the light and shadow interference parameter based on the chromaticity change amount and the contour area change amount.

6. The intelligent detection system for image semantic change according to claim 1, characterized in that The tendency analysis module is used to determine the remote sensing change excitation feature based on the visual consistency parameter and the light and shadow interference parameter, including, Weighted summing the visual consistency parameter and the light and shadow interference parameter to obtain the remote sensing change excitation feature.

7. The intelligent detection system for image semantic change according to claim 1, characterized in that, The tendency analysis module is used to distinguish the tendency of each unit to cause irrelevant changes in remote sensing data. Among them, if the remote sensing change excitation feature corresponding to the unit is greater than or equal to a predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in remote sensing data is the first tendency; if the remote sensing change excitation feature corresponding to the unit is less than the predetermined remote sensing change excitation feature threshold, it is determined that the tendency of the unit to cause irrelevant changes in remote sensing data is the second tendency.

8. The intelligent detection system for image semantic change according to claim 1, characterized in that The change detection module screens the remote sensing data based on the tendency of each unit to cause irrelevant changes in remote sensing data. Among them, the remote sensing data corresponding to the units with the first tendency is screened out.

9. The intelligent detection system for image semantic change according to claim 8, wherein The change detection module is also used to analyze the unscreened remote sensing data, including, used to extract the difference features between the unscreened remote sensing data and the basic remote sensing data, including the difference amount of the color histogram and the distribution difference of the contour features; if the difference condition is met, it is determined that there is a substantial change in the unit corresponding to the remote sensing data; wherein, the difference condition includes that the difference amount is greater than a predetermined sample difference amount threshold and the distribution difference is greater than a predetermined sample distribution difference threshold.

10. The intelligent detection system for image semantic change according to claim 1, characterized in that, The change detection module is used to call historical remote sensing data based on the remote sensing data and the time phase interval, and combine the remote sensing data with the historical remote sensing data to form dual-time-phase remote sensing data.

Images