ExtendScript-based high-resolution remote sensing image change rapid detection method
Through the automated interaction between ExtendScript and Adobe image processing software, the rapid and efficient processing of high-resolution remote sensing image change detection is achieved, solving the problem of inefficient processing in high-resolution remote sensing image processing in traditional methods, and can be used in personal consumer-grade computers Rapid detection of dynamic changes in forest logging, pests and diseases or fires has released a lot of manpower and time.
Patent Information
- Application Number
- CN202410969566.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-15
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing remote sensing image change detection methods cannot quickly and accurately detect changes without relying on large training data and high computing power support. Especially in high-resolution remote sensing image processing, traditional methods have low processing efficiency and are difficult to adapt to forest resources. Fast monitoring requirements for practical application scenarios such as protection.
Using a fast method of high-resolution remote sensing image change detection based on ExtendScript, communication is established with Adobe image processing software through ExtendScript, efficient memory management and intelligent cache strategy are realized, and the difference method is used to process two-time remote sensing images, and converted to grayscale and mosaic. The image is binarized according to the threshold value to obtain suspected change tiles.
It realizes rapid change detection of 100,000 pixel-level remote sensing images on personal consumer computers, reduces manpower and time consumption, improves the operability and efficiency of detection, and can quickly locate suspected change tiles outdoors and estimate their area. , adapt to the change detection of ultra-high resolution remote sensing images.
Smart Images

Figure CN118968095A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of remote sensing image detection, and in particular relates to a fast method for detecting changes in high-resolution remote sensing images based on ExtendScript. Background Art
[0002] Enshi Prefecture, Hubei Province, located in the hinterland of Wuling Mountain, is known as the "Central China Medicine Storehouse". It is also a major forestry prefecture in Hubei Province with rich forest resources. The forest area is 1.6656 million hectares and the forest coverage rate is 69.2%. It is also known as the "Forest Sea of Western Hubei". Enshi Prefecture has a total of 171 families, 645 genera, and 1,264 species of trees. Among them, there are 60 families, 114 genera, and 249 species of trees; 32 families, 89 genera, and 228 species of shrubs, accounting for about one-seventh of the tree species in the country. There are more than 40 rare tree species under national key protection, including Shuibin, Davidia involucrata, Baldwin, Badong Magnolia, Zhongcalyx, Glossy-leaved Davidia involucrata, Lianxiang Tree, Xiangguo Tree, Eucommia Ginkgo, etc., accounting for about 90% of the tree species listed as national key protection in the province. Enshi Prefecture has been successively rated as a "National Ecological Civilization Construction Demonstration Zone" and a national "Two Mountains" practice and innovation base. The Wuling Mountain area shoulders the responsibility of protecting the ecological environment. With economic development and the increase of human activities, it also faces challenges in protecting the ecological environment. The main reason is that the terrain of the mountainous area is complex. Traditional forest change monitoring mainly uses manual ground survey methods, which have long working cycles, high labor intensity, and poor timeliness. Some remote areas are even more difficult to effectively supervise. Remote sensing imaging technology, as an advanced information collection method, has the characteristics of large amount of information, low cost and fast speed, and has become a very important technical means in forestry ecological monitoring.
[0003] In the past decade, China's high-resolution remote sensing satellites, represented by the Gaofen series, have made significant progress, reaching the world's leading level in spatial resolution, number of bands and revisit period, and playing an increasingly important role in surveying and mapping, environmental protection, resource management and other fields. Specifically, by performing change detection on remote sensing images of the same geographical area at different times, the dynamic changes of forest resources can be monitored, such as the increase or decrease of forest coverage and changes in forest health. For some relatively hidden illegal logging activities and forest degradation caused by human construction expansion, satellite monitoring helps regulatory authorities to take legal and management measures in a timely manner to protect forest resources. In addition, remote sensing images can also quickly detect the occurrence of forest fires and monitor the spread of fires, providing important information for fire fighting and rescue operations. Using remote sensing data for change detection has always been an important research hotspot in the remote sensing community. So far, experts and scholars at home and abroad have developed a large number of remote sensing change detection and recognition algorithms, including image difference, tasseled cap transformation, post-classification comparison, spectral change vector analysis, machine learning and deep learning have also been applied to the field of change detection. With the increasing popularity of new remote sensing images, change detection has also been deeply explored in both hyperspectral images and high-resolution images.
[0004] There are many change detection methods proposed so far, which can be summarized into different types from different perspectives. Specifically, from the perspective of detection data phase, remote sensing image change detection is divided into single-phase classification comparison method, two-phase image change detection and time series change detection; from the perspective of detection analysis granularity, it is mainly divided into pixel / sub-pixel level change detection, object level change detection and scene level change detection; from the perspective of detection prior information, it can be divided into supervised change detection and unsupervised change detection according to whether there is prior data; from the perspective of detection dimension, it can be divided into two-dimensional change detection and three-dimensional change detection; from the perspective of detection time scale, it can be divided into real-time change detection, quasi-real-time change detection and non-real-time change detection. It should be pointed out that the change detection method and the subdivision type are not necessarily one-to-one corresponding, and some methods may involve multiple subdivision types. In the development of change detection technology, the method based on parameter model has the widest application range, among which the most representative is the threshold method based on difference method and ratio method, which compares the obtained difference image with the appropriate change value to obtain the change block.
[0005] Although the academic community has conducted a lot of in-depth research, since remote sensing image change detection is essentially an incomplete information inversion problem, it faces many practical problems to be solved. Scholars in the field generally believe that none of the existing detection methods are suitable for all application scenarios. The corresponding fact is that in the vast majority of existing change detection practical application scenarios, the commonly used methods are still visual interpretation and manual delineation, which require a lot of manpower and time, and have low processing efficiency. The reason for these two current situations is that scientific research pursues the accuracy and high reliability of detection algorithms, while actual needs often pay more attention to the speed and efficiency of the detection process. The two are not contradictory, nor are they always consistent. Taking forest resource protection as an example, in order to quickly grasp the dynamic changes of forest logging, pests and diseases or fires in a short period of time, forestry supervision departments need to quickly locate suspected change blocks on satellite maps, but often cannot provide a large amount of high-quality training data, nor do they have the technical capabilities required to implement intelligent algorithms. In this case, the forest type classification detection algorithm based on deep learning, which is refined to the level of dominant tree species, is difficult to adapt to the needs even if the accuracy is high, so that the universality and efficiency are not satisfactory.
[0006] As remote sensing satellites continue to make breakthroughs and progress in spatial resolution, number of bands, and revisit cycles, at the same spatial scale, higher resolutions bring larger pixel sizes and higher computing power requirements, all of which pose severe challenges to traditional remote sensing image change detection methods based on medium and low resolutions. Taking the remote sensing images used by Enshi Forestry Monitoring as an example, a single high-definition TIFF image that has been preprocessed is as high as 39064*96596 pixels, and saving it in PNG format requires 14GB of storage space. How to achieve rapid or even real-time remote sensing image change detection of forest resources in nature reserves without relying on large training data and high computing power support is a challenge that this invention must address, and it is also a technical problem that needs to be solved. Summary of the invention
[0007] The present invention provides a fast method for high-resolution remote sensing image change detection based on ExtendScript to solve the problem that the current remote sensing image change detection method cannot detect changes quickly and accurately without relying on large training data and high computing power support.
[0008] According to a first aspect of the present invention, there is provided a fast method for detecting changes in high-resolution remote sensing images based on ExtendScript, comprising:
[0009] Step S100, establishing communication between the control unit and Adobe's image processing software based on ExtendScript;
[0010] Step S200: performing environment configuration and parameter setting on the image processing software so that the image processing software has an efficient memory management mechanism and intelligent cache strategy to complete the calculation and storage of large image files;
[0011] Step S300: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to process the two-phase remote sensing images using the difference method to obtain a difference layer; then the difference layer is converted into a grayscale image, and the grayscale image is converted into a mosaic image, and finally the mosaic image is binarized according to a threshold value to obtain a binarized block, and the highlighted part in the binarized block is the suspected change block of the two-phase remote sensing image.
[0012] In an optional implementation, the control unit in step S100 is an ExtendScript script engine, and step S100 specifically includes:
[0013] Step S110, establishing an ExtendScript script, wherein the ExtendScript script is provided with instructions for establishing communication with Adobe's image processing software and instructions for controlling the image processing software to automatically execute corresponding operations in step S300 for the two-phase remote sensing images to be detected;
[0014] Step S120: When the ExtendScript script is run, the script code is loaded into the ExtendScript script engine for interpretation and execution, so that the ExtendScript script engine establishes communication with the script environment of the image processing software.
[0015] In another optional implementation, the ExtendScript script engine communicates with the image processing software at different levels across the DOM and message passing framework; and accesses and operates various functions and objects of the image processing software by calling the API provided by the image processing software to control the image processing software to automatically execute the instructions of the corresponding operation in step S300.
[0016] In another optional implementation, the control unit in step S100 is a Web APP that combines ExtendScript and HTML5 technology to achieve interaction. The Web APP interacts with the local operating system through Node.js, thereby interacting with the image processing software.
[0017] In another optional implementation, in step S200, the environment configuration and parameter settings of the image processing software are performed under the Edit / Preferences menu; the optimization items of the image processing software are as follows:
[0018]
[0019]
[0020] In another optional implementation, when the physical memory reaches its usage limit, the cache strategy and temporary disk setting are the most effective combined optimization strategies; the image processing software uses the PSB layered format to store process image data; the large image file is a large-size image with high resolution and large data volume.
[0021] In another optional implementation, the step S300 specifically includes:
[0022] Step S310: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to execute [Layer / Blending Mode / Difference] in the menu, and uses the difference blending mode to calculate the difference of the two-phase remote sensing images to obtain a difference layer;
[0023] Step S320, the control unit automatically controls the image processing software to execute the [Image / Decolorization] operation in the menu to convert the difference layer into a grayscale image;
[0024] Step S330, the control unit automatically controls the image processing software to execute the [Filter / Mosaic] operation in the menu to convert the grayscale image into a mosaic image;
[0025] Step S340, the control unit automatically controls the image processing software to execute the [Image / Threshold] and [Select / Color Range] operations in the menu to binarize the mosaic image according to the threshold value to obtain a binary block, and the highlighted portion in the binary block is the changed area of the two-phase remote sensing image.
[0026] The beneficial effects of the present invention are:
[0027] 1. The method proposed in the present invention is a simple and effective pixel-level change detection method. In forest resource monitoring, it shows good robustness to atmospheric and light interference. The main advantage of the method proposed in the present invention is that it uses the ExtendScript language to interact with image processing software at the bottom level, thus realizing the complete automation of the remote sensing image change detection process. The method proposed in the present invention does not require training data and technical capabilities for the implementation department or personnel, so that it can quickly grasp the dynamic changes of forest felling, pests or fires in a short period of time, quickly locate suspected change blocks on satellite maps and estimate their areas, and combine advanced computer automatic processing with traditional manual visual interpretation and field review, truly releasing a lot of manpower and time consumed by traditional manual methods. When facing practical application scenarios such as forest resource protection, the method pursues the operability, rapidity and efficiency of remote sensing image change detection, and does not require training data and technical capabilities for the implementation department or personnel.
[0028] The important advantage of the method proposed in the present invention is that it utilizes the efficient memory management mechanism and intelligent caching strategy of the image processing software to realize the calculation and storage of 100,000-pixel ultra-large image files in the remote sensing image change detection process, making it possible to efficiently process 100,000-pixel remote sensing images on personal consumer computers without the need to configure high-performance computers or large graphics workstations. In this way, rapid change detection of large-scale high-resolution remote sensing images can be realized. During outdoor operations, rapid change detection can be performed on 100,000-pixel two-phase remote sensing images, and suspected change blocks can be quickly located and their areas estimated. The present invention does not require segmentation or slicing of large-scale high-resolution remote sensing images. It is based on pixel difference calculation, has no training process, and is simple in method. It can quickly compare sequence images in a quasi-real-time manner to realize dynamic change monitoring.
[0029] 2. The present invention adopts PSB layered format to store process image data, breaking through the limitation of 4G size of TIFF files by most image processing software or GIS software, so that it can adapt to future ultra-high resolution remote sensing image change detection. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 It is a flow chart of an embodiment of a fast method for detecting changes in high-resolution remote sensing images based on ExtendScript of the present invention;
[0031] Figure 2 It is a schematic diagram of communication between ExtendScript and Adobe of the present invention;
[0032] Figure 3 This is a schematic diagram of the underlying elements of the image processing software that ExtendScript directly operates;
[0033] Figure 4It is a composite flow chart of pseudo code for difference calculation based on ExtendScript;
[0034] Figure 5 It is a composite flow chart of pseudo-code for decolorization and mosaic calculation based on ExtendScript;
[0035] Figure 6 It is a pseudo-code composite flow chart based on ExtendScript that highlights blocks by threshold binarization;
[0036] Figure 7 It is a schematic diagram of the core framework of the change detection method of the present invention;
[0037] Figure 8 It is a comparison chart of change detection under atmospheric and light interference;
[0038] Fig. 9 This is a change detection example of the present invention. DETAILED DESCRIPTION
[0039] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention and to make the above-mentioned purposes, features and advantages of the embodiments of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings.
[0040] In the description of the present invention, unless otherwise specified and limited, it should be noted that the term "connection" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal connection between two elements. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.
[0041] See also Figure 1 , is a flow chart of an embodiment of a fast method for high-resolution remote sensing image change detection based on ExtendScript of the present invention. The method may include:
[0042] Step S100, establishing communication between the control unit and Adobe's image processing software based on ExtendScript;
[0043] Step S200: performing environment configuration and parameter setting on the image processing software so that the image processing software has an efficient memory management mechanism and intelligent cache strategy to complete the calculation and storage of large image files;
[0044] Step S300: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to process the two-phase remote sensing images using the difference method to obtain a difference layer; then the difference layer is converted into a grayscale image, and the grayscale image is converted into a mosaic image, and finally the mosaic image is binarized according to a threshold value to obtain a binarized block, and the highlighted part in the binarized block is the suspected change block of the two-phase remote sensing image.
[0045] Among them, ExtendScript is an extended scripting language based on JavaScript, following the ECMA-262 (version 3) and EX4 (EcmaScript for XML) specifications. Since the release of Adobe CC2014, ExtendScript has become an important part of the Adobe CEP (Common Extensibility Platform) framework. As a scripting language, ExtendScript has a simple and clear syntax structure and can be written in any text editor without a complex development environment. The advantages of ExtendScript also include: scripts can be written to implement complex and repetitive automation tasks, greatly improving work efficiency; avoiding errors that may be caused by manual operations, ensuring the consistency and accuracy of processing results; being able to quickly test and verify new ideas and workflows, and achieve cross-software collaboration. ExtendScript is also highly extensible, and you can customize functions as needed, and even create your own plug-ins, extensions, and automation tools to meet specific workflow requirements.
[0046] Compared with the script automation provided by GIMP and ImageMagick, the above-mentioned features of ExtendScript are more advanced and complete in terms of code readability, ease of operation, and richness of functions. In addition, ArcMap has a professional graphical user interface, which allows you to directly detect changes in registered images by selecting the "Image Analysis" window and clicking the "Difference" button. However, manual operations cannot guarantee the consistency and accuracy of the comparison results when automating a large number of complex and repetitive tasks. Matlab and Python+OpenCV technical solutions have a high learning cost and require the configuration of a complex operating environment. In actual application scenarios, staff often do not have the technical capabilities required to implement the solution. It can be seen that high-resolution remote sensing image change detection based on ExtendScript is a new technical solution that is dedicated to the automation of the processing process of remote sensing images by image processing software, and has operability that is difficult to match with similar technical solutions.
[0047] In step S100, the control unit may be an ExtendScript script engine, and step S100 may specifically include:
[0048] Step S110, establishing an ExtendScript script, wherein the ExtendScript script is provided with instructions for establishing communication with Adobe's image processing software and instructions for controlling the image processing software to automatically execute corresponding operations in step S300 for the two-phase remote sensing images to be detected;
[0049] Step S120: When the ExtendScript script is run, the script code is loaded into the ExtendScript script engine for interpretation and execution, so that the ExtendScript script engine establishes communication with the script environment of the image processing software.
[0050] ExtendScript is an extension of JavaScript by Adobe, which provides a common scripting environment for all applications that support Adobe JavaScript, allowing communication between applications through scripts. The ExtendScript scripting engine can communicate with the image processing software at different levels through the cross-DOM and message passing framework, such as Figure 2 As shown; by calling the API (application programming interface) provided by the image processing software to access and operate various functions and objects of the image processing software (i.e., the host program), the image processing software is controlled to automatically execute the instructions of the corresponding operation in step S300. Unlike the "Macro Recording / Macro" of the Microsoft Office series of software and the "Action / Action" of the Adobe series of software, the ExtendScript script can directly access the objects and methods inside the image processing software with instructions that are closer to the bottom layer, and can directly operate the bottom-level elements such as images, layers, colors, and selections inside the image processing software to perform complex image processing tasks and maintain extremely high execution efficiency, such as Figure 3 In step S100, the control unit can be not only an ExtendScript script engine, but also a Web APP that combines ExtendScript and HTML5 technology to achieve interaction. The Web APP can interact with the local operating system through Node.js, thereby interacting with the image processing software.
[0051] In step S200, the environment configuration and parameter settings of the image processing software are performed under the Edit / Preferences menu; the optimization items of the image processing software are as follows:
[0052]
[0053]
[0054] In step S200, when the physical memory reaches the upper limit of usage, the cache strategy and temporary disk setting are the most effective combined optimization strategies; the image processing software uses the PSB layered format to store process image data; the large image file is a large-size image with high resolution and large data volume.
[0055] With the development of aerospace technology, the resolution of remote sensing images has been improved, and the amount of data has also been increasing. The existing public remote sensing image datasets are close to 10 to 100GB in size, and the size of a single remote sensing image is close to 10,000 to 100,000 pixels (for example, for practical application scenarios such as forest resource protection, a single high-definition TIFF satellite image that has been preprocessed is as high as 39064*96596 pixels). The calculation and storage of super-large image files pose a challenge to all image processing processes, and traditional change detection methods developed based on low- and medium-resolution remote sensing images are no exception. The calculation and storage of super-large image files brought about by large spatial scale and high-resolution remote sensing images pose a challenge to all image processing processes, especially change detection methods based on matrix decomposition and texture feature extraction, including deep learning methods. More and more public remote sensing image datasets exceed 100GB in size, and a single remote sensing image exceeds 100,000 pixels in size. The computing power required for change detection is getting higher and higher. For change detection methods based on deep learning, a large amount of high-quality training data is also required. Although segmenting or slicing images with very large pixel sizes can alleviate the pressure of computing and storage to a certain extent, this strategy is not suitable for all change detection algorithms and will also lose some important global properties. Increasing hardware investment to improve computing power may be the most direct solution, but it is often not the most realistic solution. For example, for forestry monitoring departments or personnel, in order to perform rapid change detection and conduct field reviews, they can only carry laptops at most, but not large graphics workstations, due to site and conditions. To this end, Photoshop's efficient memory management mechanism and intelligent caching strategy can be used to achieve the calculation and storage of very large image files, and the best computing and storage environment for high-resolution remote sensing image change detection can be built on the existing hardware configuration of the computer.
[0056] In addition, the step S300 may specifically include:
[0057] Step S310: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to execute [Layer / Blending Mode / Difference] in the menu, and uses the difference blending mode to calculate the difference of the two-phase remote sensing images to obtain a difference layer.
[0058] In this step, the difference blending mode is used as one of the layer blending modes of the image processing software to implement the difference method. The difference blending mode is used to compare the color values of corresponding pixels in two layers and generate a blending result based on the difference. The specific calculation formula of the difference blending mode is: result color = absolute value (mixed color - base color). When using the difference blending mode, if the pixel values of the mixed color and the base color are the same, the result color is black; if the mixed color and the base color are completely opposite, the result color is white; in other cases, the brightness of the result color depends on the degree of difference between the mixed color and the base color. The composite flowchart of the pseudocode for difference calculation based on ExtendScript is as follows: Figure 4 shown.
[0059] Step S320: The control unit automatically controls the image processing software to execute the [Image / Decolorization] operation in the menu to convert the difference layer into a grayscale image.
[0060] Step S330: The control unit automatically controls the image processing software to execute the [Filter / Mosaic] operation in the menu to convert the grayscale image into a mosaic image. The pseudo code composite flow chart of decolorization and mosaic calculation based on ExtendScript is as follows: Figure 5 shown.
[0061] Step S340: The control unit automatically controls the image processing software to execute the [Image / Threshold] and [Select / Color Range] operations in the menu to binarize the mosaic image according to the threshold to obtain a binarized image block. The highlighted portion (e.g., red highlight) in the binarized image block is the suspected changed image block of the two-phase remote sensing image. The pseudo code composite flow chart of binarizing the highlighted image block by threshold based on ExtendScript is as follows: Figure 6 The detection data phase of the present invention is a two-phase image or a sequence image, the detection analysis granularity is pixel level, the inspection prior information is unsupervised, the inspection space dimension is two-dimensional, and the inspection time scale is quasi-real-time change detection.
[0062] The processing framework of the remote sensing image of the present invention is as follows Figure 7As shown in the figure, image registration and field verification are the pre-operation and post-process respectively, which are not discussed here. The main algorithm difference method is a simple and effective pixel-level change detection method. The basic principle is the difference of pixel value subtraction. The main process of the difference method is: for multi-phase images that have been geometrically corrected and radiometrically corrected, the multi-phase images are first compared at the pixel level by the difference operator to obtain the difference image, and then the difference image is detected by the unsupervised detection algorithm to obtain the changed pixels. The main advantage of the difference method is that it is sensitive to pixel changes and has low computational complexity. Compared with the image matrix decomposition algorithm and the texture feature algorithm, it can easily cope with image matrices with ultra-large pixel sizes. This advantage meets the development needs of high-resolution remote sensing satellite images. The main disadvantage of the difference method is that it has high requirements for the image registration accuracy as data preprocessing, and is easily affected by the atmosphere and lighting conditions.
[0063] The present invention first uses the difference method to process the two-phase remote sensing images, then converts the difference layer into a grayscale image, converts the grayscale image into a mosaic image, and finally binarizes the mosaic image according to the threshold value. The highlight part of the binarized image block obtained is the suspected change block. The above image processing method can still maintain good robustness to atmospheric and light interference, thereby ensuring the accuracy of determining the suspected change block. In the data experiment for forest resource change detection, the change detection under atmospheric and light interference is as follows: Figure 8 As shown, from Figure 8 It can be seen that after the above image processing method is used for image processing, the atmospheric and lighting interference will not affect the determination of the changed blocks.
[0064] It should be noted that the above-mentioned image processing software can be Photoshop or Illustrator.
[0065] It can be seen from the above embodiments that the existing remote sensing image change detection methods are becoming more and more complicated, but the operability is getting lower and lower. In many practical application scenarios, they are still based on visual interpretation and manual delineation, which requires a lot of manpower and time, and has low processing efficiency. The method proposed in the present invention is a simple and effective pixel-level change detection method. In forest resource monitoring, it shows good robustness for atmospheric and lighting interference. The main advantage of the method proposed in the present invention is that it uses the ExtendScript language to interact with the image processing software at the bottom level, realizing the complete automation of the remote sensing image change detection process. The method proposed in the present invention has no training data and technical ability requirements for the implementation department or personnel, so that it can quickly grasp the dynamic changes of forest felling, pests or fires in a short period of time, quickly locate suspected change blocks on satellite maps and estimate their areas, and combine advanced computer automatic processing with traditional manual visual interpretation and field review, truly releasing a lot of manpower and time consumed by traditional manual methods; when facing practical application scenarios such as forest resource protection, the operability and fast and efficient remote sensing image change detection are pursued, and there are no training data and technical ability requirements for the implementation department or personnel.
[0066] As remote sensing image data sets become increasingly large, a single remote sensing image is approaching the upper pixel limit of all image processing software. The calculation and storage of ultra-large image files brought by large spatial scale and high-resolution remote sensing images pose a serious challenge to existing remote sensing image change detection methods. The important advantage of the method proposed in the present invention is that it utilizes the efficient memory management mechanism and intelligent caching strategy of image processing software to realize the calculation and storage of 100,000-pixel ultra-large image files in the remote sensing image change detection process, making it possible to efficiently process 100,000-pixel remote sensing images on personal consumer computers, without the need to configure high-performance computers or large graphics workstations, so that rapid change detection of large spatial scale and high-resolution remote sensing images can be realized. In the process of outdoor operations (such as field review of illegal felling points in forestry monitoring), rapid change detection can be performed on 100,000-pixel two-phase remote sensing images, and suspected change blocks can be quickly located and their areas estimated; the present invention does not need to segment or slice large spatial scale and high-resolution remote sensing images, based on pixel difference calculation, there is no training process, the method is simple, and sequence images can be quickly compared in a quasi-real-time manner to realize dynamic change monitoring.
[0067] In particular, the present invention adopts the PSB layered format to store process image data, breaking through the limitation of 4G size of TIFF files by most image processing software or GIS software, so that it can adapt to future ultra-high resolution remote sensing image change detection.
[0068] In one example, in practical application scenarios such as forest resource protection, the present invention uses Photoshop's efficient memory management mechanism and intelligent caching strategy to implement rapid change detection on two pre-processed 39064*96596 pixel high-definition satellite images (PNG format, single file size 14GB) on a consumer-grade laptop computer (configured with Intel Core i7, 64GDDR5, Nvidia T600). It only takes 2 minutes and 36 seconds to quickly locate suspected changed blocks on the satellite map and estimate their area. The entire process is automatically executed and processed by a computer without any intervention, truly freeing up a lot of manpower and time consumed by traditional manual methods. The present invention proposes a method for implementing remote sensing image change detection examples, such as Fig. 9 shown.
[0069] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. The specification and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.
[0070] It will be appreciated that the invention is not limited to the precise construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from its scope. The scope of the invention is governed solely by the appended claims.
Claims
1. A fast method for high-resolution remote sensing image change detection based on ExtendScript, characterized in that: include: Step S100, establishing communication between the control unit and Adobe's image processing software based on ExtendScript; Step S200: performing environment configuration and parameter setting on the image processing software so that the image processing software has an efficient memory management mechanism and intelligent cache strategy to complete the calculation and storage of large image files; Step S300: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to process the two-phase remote sensing images using the difference method to obtain a difference layer; then the difference layer is converted into a grayscale image, and the grayscale image is converted into a mosaic image, and finally the mosaic image is binarized according to a threshold value to obtain a binarized block, and the highlighted part in the binarized block is the suspected change block of the two-phase remote sensing image.
2. The ExtendScript-based fast method for high-resolution remote sensing image change detection according to claim 1, characterized in that: In step S100, the control unit is an ExtendScript script engine, and step S100 specifically includes: Step S110, establishing an ExtendScript script, wherein the ExtendScript script is provided with instructions for establishing communication with Adobe's image processing software and instructions for controlling the image processing software to automatically execute corresponding operations in step S300 for the two-phase remote sensing images to be detected; Step S120: When the ExtendScript script is run, the script code is loaded into the ExtendScript script engine for interpretation and execution, so that the ExtendScript script engine establishes communication with the script environment of the image processing software.
3. The ExtendScript-based fast method for high-resolution remote sensing image change detection according to claim 2 is characterized in that: The ExtendScript script engine communicates with the image processing software at different levels by crossing the DOM and message passing framework; and accesses and operates various functions and objects of the image processing software by calling the API provided by the image processing software to control the image processing software to automatically execute the instructions of the corresponding operation in step S300.
4. The fast method for high-resolution remote sensing image change detection based on ExtendScript according to claim 1, characterized in that: In the step S100, the control unit is a WebAPP that combines ExtendScript and HTML5 technology to achieve interaction. The WebAPP interacts with the local operating system through Node.js, thereby interacting with the image processing software.
5. The fast method for high-resolution remote sensing image change detection based on ExtendScript according to claim 1, characterized in that: In step S200, the environment configuration and parameter settings of the image processing software are performed under the Edit / Preferences menu; the optimization items of the image processing software are as follows:
6. The ExtendScript-based fast method for high-resolution remote sensing image change detection according to claim 5, characterized in that: When the physical memory reaches its upper limit, the cache strategy and temporary disk setting are the most effective combined optimization strategies; the image processing software uses the PSB layered format to store process image data; the large image files are large-size images with high resolution and large data volume.
7. The ExtendScript-based fast method for high-resolution remote sensing image change detection according to claim 1, characterized in that: The step S300 specifically includes: Step S310: for the two-phase remote sensing images to be detected, the control unit automatically controls the image processing software to execute [Layer / Blending Mode / Difference] in the menu, and uses the difference blending mode to calculate the difference of the two-phase remote sensing images to obtain a difference layer; Step S320, the control unit automatically controls the image processing software to execute the [Image / Decolorization] operation in the menu to convert the difference layer into a grayscale image; Step S330, the control unit automatically controls the image processing software to execute the [Filter / Mosaic] operation in the menu to convert the grayscale image into a mosaic image; Step S340, the control unit automatically controls the image processing software to execute the [Image / Threshold] and [Select / Color Range] operations in the menu to binarize the mosaic image according to the threshold value to obtain a binary block, and the highlighted portion in the binary block is the changed area of the two-phase remote sensing image.
Citation Information
Cited By
Surveying and mapping method based on low-altitude unmanned aerial vehicle remote sensing technology
CN119309550A
