A remote sensing image self-adaptive enhancement method based on piecewise hybrid mapping

Abstract

The present application relates to the technical field of remote sensing image adaptive enhancement, and particularly relates to a remote sensing image adaptive enhancement method based on segmented mixed mapping. The method comprises the following steps: obtaining original remote sensing image data to be enhanced, performing image feature statistics based on the original remote sensing image data to obtain global image feature data; calculating a global nonlinear enhancement mapping curve based on the global image feature data to obtain global nonlinear enhancement curve data; adaptively calculating image segmentation threshold values based on the global image feature data to obtain adaptive segmentation threshold value data; calculating local linear enhancement mapping curves for different segments based on the adaptive segmentation threshold value data to obtain local linear enhancement curve data; and calculating mixed weights based on the original remote sensing image data and the adaptive segmentation threshold value data; and the present application constructs a segmented mixed enhancement mapping mechanism to realize the synergistic enhancement of global contrast improvement and local detail optimization.

Description

Technical Field

[0001] This invention relates to the field of adaptive enhancement technology for remote sensing images, and in particular to an adaptive enhancement method for remote sensing images based on segmented hybrid mapping. Background Technology

[0002] Remote sensing imagery is widely used in resource surveys, environmental monitoring, and target recognition. However, due to factors such as imaging conditions, sensor performance, and atmospheric interference, acquired remote sensing images often suffer from insufficient contrast, concentrated grayscale distribution, and unclear local details, directly affecting the accuracy of subsequent information extraction and analysis. Existing technologies often employ histogram equalization, linear stretching, or single nonlinear enhancement methods to process images. While these methods can improve overall brightness and contrast to some extent, they lack the ability to differentiate between different grayscale ranges, easily leading to over-enhancement or loss of detail. This is especially true in complex terrain scenes, where it is difficult to simultaneously achieve global contrast enhancement and local detail preservation. Furthermore, some enhancement methods based on segmented processing, while introducing interval division, are prone to discontinuities at segment boundaries, resulting in visual artifacts. Therefore, how to balance global enhancement and local detail optimization within a unified framework and achieve a smooth transition in the mapping process has become a pressing technical problem to be solved in the field of remote sensing image enhancement. Summary of the Invention

[0003] Therefore, it is necessary to provide a remote sensing image adaptive enhancement method based on segmented hybrid mapping to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a remote sensing image adaptive enhancement method based on piecewise hybrid mapping includes the following steps: Step S1: Obtain the original remote sensing image data to be enhanced, and perform image feature statistics based on the original remote sensing image data to obtain global image feature data; Step S2: Based on the global image feature data, calculate the global nonlinear enhancement mapping curve to obtain the global nonlinear enhancement curve data; Step S3: Based on global image feature data, adaptively calculate the image segmentation threshold to obtain adaptive segmentation threshold data; Step S4: Based on the adaptive segmented threshold data, calculate the local linear enhancement mapping curve for different segments to obtain the local linear enhancement curve data; Step S5: Based on the original remote sensing image data and the adaptive segmented threshold data, calculate the mixing weight to obtain pixel-level mixing weight data; Step S6: Based on the global nonlinear enhancement curve data, local linear enhancement curve data, and pixel-level mixed weight data, perform weighted fusion calculation to obtain the final enhancement mapping curve data; Step S7: Perform pixel value mapping transformation on the original remote sensing image data based on the final enhanced mapping curve data, and output the enhanced remote sensing image data.

[0005] This scheme constructs a hybrid enhancement mechanism combining a global nonlinear enhancement mapping curve and a piecewise local linear enhancement mapping curve, and introduces adaptive mixing weights based on pixel segment positions to achieve hierarchical enhancement processing of remote sensing images. Specifically, step S2 establishes a global nonlinear enhancement curve, which is beneficial for overall dynamic range stretching and contrast improvement. Steps S3 and S4 construct adaptive piecewise and local linear enhancement curves, which is beneficial for targeted enhancement of detail information in different grayscale ranges. Steps S5 and S6 introduce pixel-level mixing weights and perform weighted fusion, which effectively avoids the abrupt segmentation problem in traditional piecewise enhancement and achieves a continuous and smooth transition of the mapping curve. Therefore, the enhanced remote sensing image can maintain the continuity and stability of local detail structure while improving overall brightness and contrast, thereby improving the image visual quality and the reliability of subsequent information extraction. Attached Figure Description

[0006] Figure 1 This is a flowchart illustrating the steps of an adaptive enhancement method for remote sensing images based on segmented hybrid mapping. Figure 2 Image enhancement mapping relationship and segmentation strategy diagram; Figure 3 This is an aerial view of a remotely sensed image. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0007] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0008] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0009] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0010] To achieve the above objectives, please refer to Figures 1 to 3 An adaptive enhancement method for remote sensing images based on segmented hybrid mapping includes the following steps: All specific values ​​involved in this embodiment are exemplary parameters used to clearly illustrate the technical operation process and are not the only limitation of the present invention.

[0011] Step S1: Obtain the original remote sensing image data to be enhanced, and perform image feature statistics based on the original remote sensing image data to obtain global image feature data; Step S2: Based on the global image feature data, calculate the global nonlinear enhancement mapping curve to obtain the global nonlinear enhancement curve data; Step S3: Based on global image feature data, adaptively calculate the image segmentation threshold to obtain adaptive segmentation threshold data; Step S4: Based on the adaptive segmented threshold data, calculate the local linear enhancement mapping curve for different segments to obtain the local linear enhancement curve data; Step S5: Based on the original remote sensing image data and the adaptive segmented threshold data, calculate the mixing weight to obtain pixel-level mixing weight data; Step S6: Based on the global nonlinear enhancement curve data, local linear enhancement curve data, and pixel-level mixed weight data, perform weighted fusion calculation to obtain the final enhancement mapping curve data; Step S7: Perform pixel value mapping transformation on the original remote sensing image data based on the final enhanced mapping curve data, and output the enhanced remote sensing image data.

[0012] In this embodiment, the original remote sensing image data is acquired, and statistical processing is performed on all pixel values ​​in the image to calculate the minimum and maximum pixel values. Based on this, the global dynamic range of the image is determined according to the difference between the minimum and maximum pixel values. Furthermore, frequency statistics are performed on all pixel values ​​to generate corresponding pixel value distribution histograms. Through the above processing, global image feature data, including global dynamic range data and global histogram distribution data, is obtained as the input basis for subsequent global enhancement modeling and segmented calculation.

[0013] Based on the global dynamic range data and global histogram distribution data obtained in step S1, the overall grayscale distribution characteristics of the image are analyzed to determine the parameter form of the nonlinear mapping function. For example, a global nonlinear enhancement mapping function can be constructed using a logarithmic function or a Gamma function. Subsequently, the input pixel value range is mapped and calculated according to the mapping function to generate the corresponding global nonlinear enhancement curve. This global nonlinear enhancement curve data is used to describe the adjustment relationship of the overall contrast of the image and serves as a basic mapping branch in subsequent fusion calculations.

[0014] Based on the global histogram distribution data obtained in step S1, the cumulative distribution function is calculated on the pixel value distribution to obtain the cumulative probability distribution of the pixel value; according to the preset segmentation probability threshold, multiple initial segmentation points are determined on the cumulative probability distribution curve; further, combined with the spatial distribution characteristics of the original remote sensing image data, the initial segmentation points are subjected to spatial coherence constraints and optimization processing to eliminate the segmentation instability problem caused by local abnormal fluctuations, and finally adaptive segmentation threshold data is obtained.

[0015] The adaptive segmentation threshold data is used to uniformly divide pixel value ranges and provide a segmentation basis for local enhancement and weight calculation.

[0016] Based on the adaptive segmented threshold data obtained in step S3, the overall pixel value range is divided into multiple continuous sub-intervals. For each sub-interval, a corresponding linear mapping function is constructed to achieve local enhancement of pixel values ​​within that interval. At the boundaries between adjacent sub-intervals, a smooth transition process is performed on the corresponding linear mapping function to ensure the continuity of the mapping curve at the segmentation points. Through the above processing, local linear enhancement curve data covering the entire pixel value range is obtained, which is used to characterize the detail enhancement relationship within different grayscale intervals.

[0017] For each pixel in the original remote sensing image data, the segment interval to which the pixel value belongs is determined based on the adaptive segmentation threshold data obtained in step S3; the positional relationship of the pixel value relative to the boundary of the segment interval is further calculated and converted into a normalized distance; based on the normalized distance, the corresponding mixed weight value is calculated through a preset weight function.

[0018] This yields pixel-level hybrid weight data for each pixel, used to characterize the weight distribution relationship between global nonlinear enhancement and local linear enhancement for that pixel.

[0019] For each pixel value in the original remote sensing image, the corresponding global mapping value is obtained based on the global nonlinear enhancement curve data obtained in step S2, and the corresponding local mapping value is obtained based on the local linear enhancement curve data obtained in step S4. Then, the global mapping value and the local mapping value are weighted and summed together using the pixel-level hybrid weight data obtained in step S5 to obtain the final mapping result for that pixel. By performing the above fusion calculation on all pixels, complete final enhancement mapping curve data is formed, which is used to uniformly describe the enhancement transformation relationship of pixel values.

[0020] Each pixel value in the original remote sensing image data is input into the final enhancement mapping curve data obtained in step S6, and the corresponding enhanced pixel value is obtained by lookup or function calculation. Then, all enhanced pixel values ​​are reorganized according to their original spatial locations to generate an enhanced remote sensing image matrix. Finally, the enhanced remote sensing image data is stored or output to obtain the image enhancement result.

[0021] Please refer to [link / reference needed] for further information. Figure 2 The histogram corresponds to the global histogram distribution data obtained in the image. The P2, P60, and P98 quantiles marked in the image represent the adaptive segmented threshold data calculated based on the probability threshold, used to divide the pixel value range into intervals. This corresponds to the global nonlinear enhancement principle and is used to process low pixel value areas (such as near P2); linear stretching (highlights) is consistent with the principle of constructing local linear enhancement curves for high pixel value intervals (such as near P98).

[0022] Please refer to [link / reference needed] for further information. Figure 3 This study showcases a typical complex remote sensing scene comprising urban buildings, vegetation, and bare land (brightly colored areas). These features generate a global histogram distribution with multi-peak characteristics. Adaptive segmentation thresholds (e.g., P2, P60, P98) intelligently divide pixel value ranges into different regions corresponding to shadows, mid-gray buildings, and bright bare land. Local linear enhancement curves are calculated for each region (e.g., shadows and highlights) to optimize shadow details and suppress highlight overexposure. By weighted fusion of global and local mapping curves and mapping transformation of all pixels in the image, the contrast and detail of various features are adaptively enhanced while maintaining a natural overall transition, resulting in a visually superior enhanced image.

[0023] Preferably, step S1 includes the following steps: Step S11: Acquire the original remote sensing image data to be enhanced; Step S12: Perform pixel value statistics on the original remote sensing image data to obtain the minimum and maximum pixel values; Step S13: Calculate the global dynamic range based on the minimum and maximum pixel values ​​to obtain global dynamic range data; Step S14: Calculate the global pixel value histogram based on the original remote sensing image data and global dynamic range data to obtain the global histogram distribution data.

[0024] In this embodiment, the original remote sensing image data to be enhanced is read from the remote sensing data source and represented as a two-dimensional pixel matrix, where each element in the matrix corresponds to a pixel value. This original remote sensing image data serves as the basic input data for subsequent pixel statistics and feature calculations. Based on the original remote sensing image data obtained in step S11, all pixel values ​​in the image are statistically analyzed, and the minimum and maximum values ​​among all pixel values ​​are recorded as the minimum and maximum pixel values ​​of the current image, respectively. The minimum and maximum pixel values ​​are used to characterize the grayscale range of the image and serve as input for subsequent dynamic range calculations. Based on the minimum and maximum pixel values ​​obtained in step S12, a difference calculation is performed on the two to obtain the corresponding dynamic range difference; this dynamic range difference is determined as the global dynamic range data of the current image. The global dynamic range data is used to characterize the span of the overall grayscale distribution of the image and serves as the constraint range in the subsequent histogram construction process.

[0025] Based on the raw remote sensing image data obtained in step S11, the frequency of occurrence of each pixel value is statistically analyzed to obtain the pixel value frequency distribution. Building upon this, and combining the global dynamic range data obtained in step S13, the pixel values ​​are uniformly segmented or normalized to construct corresponding pixel value histograms. This forms global histogram distribution data covering the entire dynamic range. The global histogram distribution data is used to characterize the probability distribution of pixel values ​​and serves as the input basis for subsequent global nonlinear enhancement curve calculation and segmentation threshold determination.

[0026] Preferably, step S13 includes: Step S131: Calculate the difference between the minimum pixel value and the maximum pixel value to obtain the dynamic range difference data; Step S132: Determine the dynamic range difference data as the global dynamic range data.

[0027] In this embodiment, the minimum pixel value obtained in step S12 is denoted as... The maximum pixel value is denoted as Perform a difference calculation on the two, that is, by - The corresponding dynamic range difference data is obtained. This dynamic range difference is used to reflect the span of pixel value distribution in the current remote sensing image.

[0028] The dynamic range difference data obtained in step S131 is directly output as global dynamic range data and used to characterize the overall grayscale variation range of the original remote sensing image data. This global dynamic range data is used as the basic parameter for constructing pixel value histograms and calculating enhancement mapping functions in subsequent steps.

[0029] Preferably, step S14 includes the following steps: Step S141: Based on the original remote sensing image data, count the frequency of each pixel value to obtain pixel value frequency distribution data; Step S142: Combine global dynamic range data and pixel value frequency distribution data to generate a global pixel value histogram and obtain global histogram distribution data.

[0030] In this embodiment, all pixels in the original remote sensing image data obtained in step S11 are traversed and counted, and the frequency of occurrence of each pixel value is accumulated according to the pixel value level to obtain the frequency of each pixel value; all pixel values ​​and their corresponding frequencies are combined to form pixel value frequency distribution data, which is used to reflect the occurrence of each gray level in the image.

[0031] Based on the pixel value frequency distribution data obtained in step S141, and combined with the global dynamic range data obtained in step S13, the pixel value range is uniformly constrained, and the pixel values ​​are binned according to a preset segmentation method. The frequencies within each bin are summarized to construct the corresponding pixel value histogram, thereby obtaining global histogram distribution data covering the entire dynamic range. This global histogram distribution data is used to describe the overall distribution characteristics of pixel values ​​and serves as the input basis for subsequent global nonlinear enhancement curve calculation and segmented threshold determination.

[0032] Preferably, step S2 includes the following steps: Step S21: Calculate the parameters of the global nonlinear enhancement curve based on the global dynamic range data and the global histogram distribution data; Step S22: Construct a global nonlinear enhancement mapping function based on the global nonlinear enhancement curve parameters; Step S23: Generate global nonlinear enhancement curve data based on the global nonlinear enhancement mapping function.

[0033] In this embodiment, the range of pixel values ​​is determined using the global dynamic range data obtained in step S13, and the distribution characteristics of pixel values ​​are analyzed in conjunction with the global histogram distribution data obtained in step S14. Based on the gray-level concentration and distribution offset in the histogram, the parameters of the nonlinear mapping function are determined, such as the Gamma coefficient or logarithmic transformation coefficient, so that the denser pixel distribution range is stretched and the sparser distribution range remains stable, thereby obtaining the global nonlinear enhancement curve parameters.

[0034] Based on the global nonlinear enhancement curve parameters obtained in step S21, a corresponding nonlinear function form (such as the Gamma function or the logarithmic function) is selected to construct a mapping relationship between pixel value input and output, forming a global nonlinear enhancement mapping function. This mapping function is used to describe the overall enhancement trend of pixel values ​​throughout the entire dynamic range.

[0035] The global nonlinear enhancement mapping function constructed in step S22 is applied to the entire pixel value range, and the mapping result is calculated for each pixel value to obtain the functional relationship data between the corresponding input pixel value and the output pixel value, thereby forming the global nonlinear enhancement curve data. This curve data serves as the input for the global enhancement branch in subsequent fusion calculations.

[0036] Preferably, step S3 includes the following steps: Step S31: Calculate the cumulative distribution function based on the global histogram distribution data to obtain the pixel cumulative probability distribution data; Step S32: Based on the pixel cumulative probability distribution data, calculate the initial segmentation points according to the preset segmentation probability threshold; Step S33: Based on the initial segmentation points and the original remote sensing image data, perform spatial coherence optimization to obtain adaptive segmentation threshold data.

[0037] In this embodiment, based on the global histogram distribution data obtained in step S14, the frequency corresponding to each pixel value is normalized to obtain the probability distribution of each pixel value. On this basis, the probabilities are progressively accumulated according to the pixel value size pattern to calculate the corresponding cumulative distribution function, thereby forming pixel cumulative probability distribution data. This data is used to reflect the cumulative distribution of pixel values ​​from low to high.

[0038] Based on the pixel cumulative probability distribution data obtained in step S31, several segmentation probability thresholds (e.g., 0.25, 0.5, 0.75, etc.) are pre-set, and the pixel values ​​corresponding to the threshold positions are found in the cumulative distribution function. These pixel values ​​are determined as initial segmentation points. The initial segmentation points are used to divide the pixel value range into multiple intervals, providing an initial segmentation basis for subsequent local enhancement modeling.

[0039] For the initial segmentation points obtained in step S32, combined with the original remote sensing image data acquired in step S11, the spatial distribution of pixels corresponding to each segmentation point is analyzed. For segmentation points that exhibit discrete or discontinuous spatial distribution, their positions are adjusted or merged to ensure the relative spatial continuity of pixels within the same segment. Through the above optimization process, adaptive segmentation threshold data is obtained. This data is used to uniformly divide pixel value intervals and serves as the basis for subsequent local linear enhancement curve construction and mixed weight calculation.

[0040] Preferably, step S4 includes the following steps: Step S41: Divide the pixel value range into multiple sub-intervals based on the adaptive segmented threshold data to obtain multiple segmented interval range data; Step S42: Construct a linear mapping function for each segmented interval range of data to obtain multiple initial local linear enhancement curves; Step S43: Perform smooth transition processing on adjacent initial local linear enhancement curves at the segmentation points to obtain local linear enhancement curve data.

[0041] In this embodiment, based on the adaptive segmented threshold data obtained in step S33, the overall pixel value range is divided according to each segmented threshold, forming multiple continuous and non-overlapping sub-intervals; each sub-interval is defined by two adjacent segmented thresholds, thereby obtaining multiple segmented interval range data. The segmented interval range data is used to limit the effective range of the subsequent local enhancement function.

[0042] Based on the segmented interval range data obtained in step S41, a corresponding linear mapping function is constructed for each sub-interval, so that the pixel values ​​in the interval are enhanced according to a linear relationship. By establishing linear mapping relationships for all segmented intervals, multiple initial local linear enhancement curves are obtained to characterize the local enhancement characteristics in different grayscale intervals.

[0043] For the multiple initial local linear enhancement curves obtained in step S42, the corresponding linear mapping functions are smoothed at the boundary points of each adjacent segment interval, ensuring that the mapping relationship changes continuously at the segment points and avoiding abrupt changes or discontinuities. Through the above smoothing transition process, local linear enhancement curve data covering the entire pixel value range is obtained. This data is used for subsequent fusion calculations with the global nonlinear enhancement curves.

[0044] Preferably, step S5 includes the following steps: Step S51: For each pixel in the original remote sensing image data, determine the segment interval to which its pixel value belongs, and obtain the pixel segment assignment data; Step S52: Based on the pixel segmentation data, calculate the normalized distance from the pixel value to the boundary of its segment interval; Step S53: Calculate the blending weights based on the normalized distance using a preset weighting function to obtain pixel-level blending weight data.

[0045] In this embodiment, for each pixel value in the original remote sensing image data, it is compared with the adaptive segmentation threshold data obtained in step S33 to determine the range within which the pixel value falls, thereby determining its segmentation interval; the segmentation interval number corresponding to each pixel is recorded to form pixel segmentation attribution data. This data is used to identify the positional relationship of each pixel in the segmentation system.

[0046] Based on the pixel segmentation data obtained in step S51, for each pixel, the upper and lower boundary values ​​of its segment interval are obtained; the distances from the pixel value to the lower and upper boundaries of the interval are calculated, and normalized according to the interval width to obtain the normalized position distance of the pixel within the current segment interval. The normalized distance is used to characterize the relative position of the pixel within the segment interval.

[0047] Based on the normalized distance obtained in step S52, it is input into a preset weighting function for calculation, such as a linear function or a smoothing function, to convert the normalized distance into a corresponding weight value. This weight value is used to represent the contribution ratio of the current pixel between global nonlinear enhancement and local linear enhancement. The above calculation is performed on all pixels to obtain the corresponding pixel-level fusion weight data. This data serves as the weight input in subsequent fusion calculations.

[0048] Preferably, step S6 includes the following steps: Step S61: Based on pixel-level hybrid weight data, obtain the global nonlinear enhancement curve mapping value corresponding to the current pixel to obtain global mapping value data; Step S62: Based on pixel-level hybrid weight data, obtain the local linear enhancement curve mapping value corresponding to the current pixel to obtain local mapping value data; Step S63: Perform a weighted summation based on the global mapping value data, local mapping value data, and corresponding pixel-level hybrid weight data to obtain the final enhanced mapping curve data.

[0049] In this embodiment, for each pixel value in the original remote sensing image data, the pixel value is input into the global nonlinear enhancement curve data obtained in step S23. The global nonlinear mapping result corresponding to the pixel value is obtained through lookup or function calculation, and this result serves as the global mapping value data for that pixel. This global mapping value reflects the enhancement result after adjusting the overall grayscale distribution.

[0050] Based on step S61, for the same pixel value, according to the pixel segmentation data obtained in step S51, its segment interval is determined; the pixel value is then input into the local linear enhancement curve of the corresponding segment obtained in step S43 to obtain its local linear mapping result, which is used as the local mapping value data for that pixel. This local mapping value is used to reflect the detail enhancement effect within the local interval.

[0051] For each pixel, the global mapping value data obtained in step S61 and the local mapping value data obtained in step S62 are combined with the pixel-level fusion weight data obtained in step S53 for weighted calculation. That is, the two types of mapping values ​​are weighted and summed according to a preset weight relationship to obtain the final mapping result corresponding to the pixel. The above fusion calculation is performed on all pixels to form complete final enhanced mapping curve data. This data is used to uniformly describe the final enhanced mapping relationship of pixel values.

[0052] Preferably, step S7 includes the following steps: Step S71: For each pixel value in the original remote sensing image data, find or calculate the corresponding enhanced pixel value based on the final enhancement mapping curve data; Step S72: Combine all enhanced pixel values ​​to form an enhanced image matrix to obtain enhanced remote sensing image data; Step S73: Store or transmit the enhanced remote sensing image data and output the final result.

[0053] In this embodiment, for each pixel value in the original remote sensing image data obtained in step S11, it is input into the final enhancement mapping curve data obtained in step S63. The corresponding mapping output value is obtained by looking up a table or calculating a function, and this output value is determined as the enhanced pixel value of that pixel. This process is performed on all pixels one by one, thereby completing the pixel-level enhancement calculation. Based on the enhanced pixel values ​​of all pixels obtained in step S71, the enhanced pixel values ​​are rearranged and combined according to the spatial positional relationship of the original remote sensing image data to construct a corresponding two-dimensional image matrix, thereby forming the enhanced remote sensing image data. This enhanced image maintains the same spatial structure as the original image.

[0054] The enhanced remote sensing image data obtained in step S72 is stored according to a preset data format, or transmitted via a communication interface to output the enhancement result. This enhanced remote sensing image data can be used for subsequent image analysis, target recognition, or display applications.

[0055] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A remote sensing image adaptive enhancement method based on piecewise hybrid mapping, characterized in that, Includes the following steps: Step S1: Obtain the original remote sensing image data to be enhanced, and perform image feature statistics based on the original remote sensing image data to obtain global image feature data; Step S2: Based on the global image feature data, calculate the global nonlinear enhancement mapping curve to obtain the global nonlinear enhancement curve data; Step S3: Based on global image feature data, adaptively calculate the image segmentation threshold to obtain adaptive segmentation threshold data; Step S4: Based on the adaptive segmented threshold data, calculate the local linear enhancement mapping curve for different segments to obtain the local linear enhancement curve data; Step S5: Based on the original remote sensing image data and the adaptive segmented threshold data, calculate the mixing weight to obtain pixel-level mixing weight data; Step S6: Based on the global nonlinear enhancement curve data, local linear enhancement curve data, and pixel-level mixed weight data, perform weighted fusion calculation to obtain the final enhancement mapping curve data; Step S7: Perform pixel value mapping transformation on the original remote sensing image data based on the final enhanced mapping curve data, and output the enhanced remote sensing image data.

2. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Acquire the original remote sensing image data to be enhanced; Step S12: Perform pixel value statistics on the original remote sensing image data to obtain the minimum and maximum pixel values; Step S13: Calculate the global dynamic range based on the minimum and maximum pixel values ​​to obtain global dynamic range data; Step S14: Calculate the global pixel value histogram based on the original remote sensing image data and global dynamic range data to obtain the global histogram distribution data.

3. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 2, characterized in that, Step S13 includes the following steps: Step S131: Calculate the difference between the minimum pixel value and the maximum pixel value to obtain the dynamic range difference data; Step S132: Determine the dynamic range difference data as the global dynamic range data.

4. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 3, characterized in that, Step S14: The following steps are followed: Step S141: Based on the original remote sensing image data, count the frequency of each pixel value to obtain pixel value frequency distribution data; Step S142: Combine global dynamic range data and pixel value frequency distribution data to generate a global pixel value histogram and obtain global histogram distribution data.

5. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Calculate the parameters of the global nonlinear enhancement curve based on the global dynamic range data and the global histogram distribution data; Step S22: Construct a global nonlinear enhancement mapping function based on the global nonlinear enhancement curve parameters; Step S23: Generate global nonlinear enhancement curve data based on the global nonlinear enhancement mapping function.

6. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S3 includes the following steps: Step S31: Calculate the cumulative distribution function based on the global histogram distribution data to obtain the pixel cumulative probability distribution data; Step S32: Based on the pixel cumulative probability distribution data, calculate the initial segmentation points according to the preset segmentation probability threshold; Step S33: Based on the initial segmentation points and the original remote sensing image data, perform spatial coherence optimization to obtain adaptive segmentation threshold data.

7. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Divide the pixel value range into multiple sub-intervals based on the adaptive segmented threshold data to obtain multiple segmented interval range data; Step S42: Construct a linear mapping function for each segmented interval range of data to obtain multiple initial local linear enhancement curves; Step S43: Perform smooth transition processing on adjacent initial local linear enhancement curves at the segmentation points to obtain local linear enhancement curve data.

8. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: For each pixel in the original remote sensing image data, determine the segment interval to which its pixel value belongs, and obtain the pixel segment assignment data; Step S52: Based on the pixel segmentation data, calculate the normalized distance from the pixel value to the boundary of its segment interval; Step S53: Calculate the blending weights based on the normalized distance using a preset weighting function to obtain pixel-level blending weight data.

9. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S6 includes the following steps: Step S61: Based on pixel-level hybrid weight data, obtain the global nonlinear enhancement curve mapping value corresponding to the current pixel to obtain global mapping value data; Step S62: Based on pixel-level hybrid weight data, obtain the local linear enhancement curve mapping value corresponding to the current pixel to obtain local mapping value data; Step S63: Perform a weighted summation based on the global mapping value data, local mapping value data, and corresponding pixel-level hybrid weight data to obtain the final enhanced mapping curve data.

10. The remote sensing image adaptive enhancement method based on segmented hybrid mapping according to claim 1, characterized in that, Step S7 includes the following steps: Step S71: For each pixel value in the original remote sensing image data, find or calculate the corresponding enhanced pixel value based on the final enhancement mapping curve data; Step S72: Combine all enhanced pixel values ​​to form an enhanced image matrix to obtain enhanced remote sensing image data; Step S73: Store or transmit the enhanced remote sensing image data and output the final result.

Images