CN117218552B - Estimation algorithm optimization method and device based on pixel change detection - Google Patents

Abstract

The present invention discloses an estimation algorithm optimization method and device based on pixel change detection, the method comprising: obtaining training sample images of multiple time periods of a target surface area; based on a pixel change recognition algorithm model, identifying stable pixels and change pixels to be detected from the training sample images of the multiple time periods; based on a pixel change period detection algorithm model, detecting relatively stable pixels and change pixels from the change pixels to be detected; aggregating the band reflectivity information of all the stable pixels and the relatively stable pixels and the corresponding stable time periods to obtain a stable training database; aggregating the band reflectivity information of all the change pixels in the same change time period to obtain a change training database. It can be seen that the present invention can make full use of the change characteristics of different pixels to screen out different training data sets, so as to obtain a model that can be targeted for prediction in subsequent training, thereby improving the accuracy of model estimation.

Description

A method and device for optimizing estimation algorithm based on pixel change detection

Technical Field

The present invention relates to the technical field of remote sensing data processing, and in particular to an estimation algorithm optimization method and device based on pixel change detection.

Background technique

With the large population gathering, urban construction land has expanded rapidly, resulting in significant changes in the land surface. This process has a significant impact on the urban environment. The geographic spatial pattern of urban areas is complex and highly heterogeneous. Traditional remote sensing image monitoring and analysis methods are difficult to meet the environmental monitoring needs of rapidly urbanizing areas.

Impervious surface is a typical land cover type in urban areas and a key indicator of urban ecological environment. It provides an irreplaceable sub-pixel level research perspective for monitoring and analyzing urbanization and its ecological and environmental effects. Accurate and efficient remote sensing image information extraction methods are the basis of research related to impervious surfaces and have received continuous attention from many researchers in recent years. However, in existing multi-time series monitoring studies, the systematic errors and random errors of the sub-pixel method itself accumulate to produce compound errors, which often lead to significant impacts on the monitoring results in terms of temporal consistency and temporal resolution, making it difficult to accurately reflect the true process of surface changes. It can be seen that the existing technology has defects that need to be solved urgently.

Summary of the invention

The technical problem to be solved by the present invention is to provide a method and device for optimizing an estimation algorithm based on pixel change detection, which can make full use of the change characteristics of different pixels to screen out different training data sets, so as to facilitate subsequent training to obtain a model that can be used for targeted prediction, thereby effectively improving the accuracy of model estimation.

In order to solve the above technical problems, the first aspect of the present invention discloses an estimation algorithm optimization method based on pixel change detection, the method comprising:

Obtain training sample images of the target surface area at multiple time periods;

Based on the pixel change recognition algorithm model, identifying stable pixels and change pixels to be detected from the training sample images of the multiple time periods;

Based on the pixel change period detection algorithm model, relatively stable pixels and changed pixels are detected from the changed pixels to be detected;

Aggregating the band reflectance information and the corresponding stable time periods of all the stable pixels and the relatively stable pixels to obtain a stable training database; the stable training database is used to train an estimation algorithm model for estimating the impermeable density parameter of the target surface area in the stable time period;

Aggregate the band reflectance information of all the changed pixels in the same change time period to obtain a change training database; the change training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area in the corresponding change time period.

As an optional implementation, in the first aspect of the present invention, the method of identifying stable pixels and change pixels to be detected from the training sample images of the multiple time periods based on the pixel change recognition algorithm model includes:

Inputting the training sample images of the plurality of time periods into a trained stable pixel recognition machine learning model to obtain recognized stable pixels; the stable pixel recognition machine learning model is trained by a training data set including a plurality of training images and corresponding visual change annotations;

The pixels in the training sample image other than the stable pixels are determined as the change pixels to be detected.

As an optional implementation, in the first aspect of the present invention, the method of detecting relatively stable pixels and changing pixels from the changing pixels to be detected based on the pixel change period detection algorithm model includes:

For any of the changed pixel to be measured, calculating the remote sensing index parameter of the changed pixel to be measured;

Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period;

If not, the pixel to be tested is determined as a relatively stable pixel;

If yes, then determine the number of change periods of the pixel to be tested within the time period;

According to the number of change periods and a preset land cover type change detection method, it is determined whether the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional implementation, in the first aspect of the present invention, the calculating of the remote sensing index parameter of the changed pixel to be measured comprises:

Calculate the NDISI index of the pixel to be tested for the change;

Calculate the green component of the tasseled cap transformation of the pixel to be measured;

The NDISI index and the greenness component of the changed pixel to be measured are normalized to obtain the remote sensing index parameter of the changed pixel to be measured.

As an optional implementation, in the first aspect of the present invention, the determining whether the remote sensing index parameter of the pixel to be detected changes within a preset time period includes:

Classifying the remote sensing index parameters into multiple levels;

It is determined whether the level of the remote sensing index parameter of the pixel to be detected changes within a preset time period.

As an optional implementation, in the first aspect of the present invention, determining the number of change periods of the changed pixel to be detected within the time period includes:

The number of time periods in which the level of the remote sensing index parameter of the changed pixel to be measured changes within the time period cycle is determined to obtain the number of change periods of the changed pixel to be measured.

As an optional implementation, in the first aspect of the present invention, the method of determining, based on the number of change periods and a preset land cover type change detection method, whether the change pixel to be detected is a relatively stable pixel or a change pixel comprises:

Determine whether the number of change periods is 1, and if not, determine the changed pixel to be detected as a changed pixel;

If so, based on a preset land cover type change detection method, it is determined that the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional implementation, in the first aspect of the present invention, the method for detecting a change in land cover type based on a preset method determines whether the changed pixel to be detected is a relatively stable pixel or a changed pixel, including:

Acquire the land cover data of the first and last two periods in the time period cycle corresponding to the target surface area;

Use a 3*3 moving window to detect the types of the land cover data of the first and last periods, and determine whether the type of the pixel to be detected in the land cover data changes during the change period;

If yes, then the changed pixel to be detected is determined to be a changed pixel;

Otherwise, it is determined that the changed pixel to be tested is a relatively stable pixel.

The second aspect of the present invention discloses an estimation algorithm optimization device based on pixel change detection, the device comprising:

An acquisition module, used to acquire training sample images of a target surface area at multiple time periods;

An identification module, used to identify stable pixels and change pixels to be detected from the training sample images of the multiple time periods based on a pixel change identification algorithm model;

A detection module, used to detect relatively stable pixels and changed pixels from the changed pixels to be detected based on a pixel change period detection algorithm model;

A first aggregation module is used to aggregate the band reflectance information and the corresponding stable time period of all the stable pixels and the relatively stable pixels to obtain a stable training database; the stable training database is used to train an estimation algorithm model for estimating the impermeable density parameter of the target surface area in the stable time period;

The second aggregation module is used to aggregate the band reflectance information of all the changed pixels in the same change time period to obtain a change training database; the change training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area in the corresponding change time period.

As an optional implementation, in the second aspect of the present invention, the specific manner in which the recognition module identifies stable pixels and change pixels to be detected from the training sample images of the multiple time periods based on the pixel change recognition algorithm model includes:

Inputting the training sample images of the plurality of time periods into a trained stable pixel recognition machine learning model to obtain recognized stable pixels; the stable pixel recognition machine learning model is trained by a training data set including a plurality of training images and corresponding visual change annotations;

The pixels in the training sample image other than the stable pixels are determined as the change pixels to be detected.

As an optional implementation, in the second aspect of the present invention, the specific manner in which the detection module detects relatively stable pixels and changing pixels from the changing pixels to be detected based on the pixel change period detection algorithm model includes:

For any of the changed pixel to be measured, calculating the remote sensing index parameter of the changed pixel to be measured;

Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period;

If not, the pixel to be tested is determined as a relatively stable pixel;

If yes, then determine the number of change periods of the pixel to be tested within the time period;

According to the number of change periods and a preset land cover type change detection method, it is determined whether the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional implementation, in the second aspect of the present invention, the specific manner in which the detection module calculates the remote sensing index parameter of the changed pixel to be detected includes:

Calculate the NDISI index of the pixel to be tested for the change;

Calculate the green component of the tasseled cap transformation of the pixel to be measured;

The NDISI index and the greenness component of the changed pixel to be measured are normalized to obtain the remote sensing index parameter of the changed pixel to be measured.

As an optional implementation, in the second aspect of the present invention, the specific manner in which the detection module determines whether the remote sensing index parameter of the pixel to be detected changes within a preset time period includes:

Classifying the remote sensing index parameters into multiple levels;

It is determined whether the level of the remote sensing index parameter of the pixel to be detected changes within a preset time period.

As an optional implementation, in the second aspect of the present invention, the specific manner in which the detection module determines the number of change periods of the changed pixel to be detected within the time period includes:

The number of time periods in which the level of the remote sensing index parameter of the changed pixel to be measured changes within the time period cycle is determined to obtain the number of change periods of the changed pixel to be measured.

As an optional implementation, in the second aspect of the present invention, the detection module determines the specific manner in which the changed pixel to be detected is a relatively stable pixel or a changed pixel according to the number of change periods and a preset land cover type change detection method, including:

Determine whether the number of change periods is 1, and if not, determine the changed pixel to be detected as a changed pixel;

If so, based on a preset land cover type change detection method, it is determined that the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional implementation, in the second aspect of the present invention, the detection module determines the specific manner in which the changed pixel to be detected is a relatively stable pixel or a changed pixel based on a preset land cover type change detection method, including:

Acquire the land cover data of the first and last two periods in the time period cycle corresponding to the target surface area;

Use a 3*3 moving window to detect the types of the land cover data of the first and last periods, and determine whether the type of the pixel to be detected in the land cover data changes during the change period;

If yes, then the changed pixel to be detected is determined to be a changed pixel;

Otherwise, it is determined that the changed pixel to be tested is a relatively stable pixel.

The third aspect of the present invention discloses another estimation algorithm optimization device based on pixel change detection, the device comprising:

A memory storing executable program code;

a processor coupled to the memory;

The processor calls the executable program code stored in the memory to execute part or all of the steps in the estimation algorithm optimization method based on pixel change detection disclosed in the first aspect of the present invention.

The fourth aspect of the present invention discloses a portable terminal for customs sorting, including a graphic code scanning device and a data processing device, wherein the data processing device is used to execute part or all of the steps in the estimation algorithm optimization method based on pixel change detection disclosed in the first aspect of the present invention.

Compared with the prior art, the present invention has the following beneficial effects:

It can be seen that the embodiments of the present invention can screen out stable pixels and changing pixels from training sample images based on the pixel change detection algorithm, and aggregate different pixels to train different estimation models, so as to make full use of the change characteristics of different pixels to screen out different training data sets, so as to facilitate subsequent training to obtain a model that can be used for targeted prediction, thereby effectively improving the accuracy of model estimation.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

FIG1 is a flow chart of an estimation algorithm optimization method based on pixel change detection disclosed in an embodiment of the present invention.

FIG. 2 is a schematic diagram of the structure of an estimation algorithm optimization device based on pixel change detection disclosed in an embodiment of the present invention.

FIG3 is a schematic diagram of the structure of another estimation algorithm optimization device based on pixel change detection disclosed in an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

The terms "second", "secondary", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or equipment.

Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

The present invention discloses an estimation algorithm optimization method and device based on pixel change detection, which can screen out stable pixels and changing pixels from training sample images based on the pixel change detection algorithm, and aggregate different pixels to train different estimation models, so as to make full use of the change characteristics of different pixels to screen out different training data sets, so as to obtain a model that can be targeted for prediction in subsequent training, and effectively improve the accuracy of model estimation. The following are detailed descriptions.

Embodiment 1

Please refer to FIG. 1, which is a flow chart of an estimation algorithm optimization method based on pixel change detection disclosed in an embodiment of the present invention. The estimation algorithm optimization method based on pixel change detection described in FIG. 1 is applied to a data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). As shown in FIG. 1, the estimation algorithm optimization method based on pixel change detection may include the following operations:

101. Obtain training sample images of the target surface area at multiple time periods.

102. Based on the pixel change recognition algorithm model, stable pixels and change pixels to be tested are identified from the training sample images of multiple time periods.

103. Based on the pixel change period detection algorithm model, relatively stable pixels and changing pixels are detected from the changing pixels to be tested.

104. Aggregate the band reflectance information of all stable pixels and relatively stable pixels and the corresponding stable time periods to obtain a stable training database.

The stable training database is used to train an estimation algorithm model for estimating the impervious density parameter of a target surface area for a stable time period.

105. Aggregate the band reflectance information of all changed pixels in the same change time period to obtain a change training database.

The variation training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area for the corresponding variation time period.

It can be seen that the above-mentioned embodiments of the invention can screen out stable pixels and changing pixels from the training sample images based on the pixel change detection algorithm, and aggregate different pixels to train different estimation models, so as to make full use of the change characteristics of different pixels to screen out different training data sets, so as to facilitate subsequent training to obtain a model that can be used for targeted prediction, effectively improving the accuracy of model estimation.

As an optional embodiment, in the above steps, based on the pixel change recognition algorithm model, identifying stable pixels and change pixels to be detected from training sample images of multiple time periods includes:

The training sample images of multiple time periods are input into the trained stable pixel recognition machine learning model to obtain the identified stable pixels.

The pixels other than stable pixels in the training sample images are determined as change pixels to be tested.

Specifically, the stable pixel recognition machine learning model is trained using a training data set including multiple training images and corresponding visual change annotations.

Through the above embodiments, a machine learning model can be used to automatically identify obvious stable pixels, thereby improving the efficiency and accuracy of pixel classification.

As an optional embodiment, in the above steps, based on the pixel change period detection algorithm model, detecting relatively stable pixels and changing pixels from the changing pixels to be detected includes:

For any pixel to be measured with changes, the remote sensing index parameter of the pixel to be measured with changes is calculated;

Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period;

If not, the pixel to be tested is determined as a relatively stable pixel;

If yes, then determine the number of change periods of the pixel to be tested within the time period;

According to the number of change periods and the preset land cover type change detection method, it is determined whether the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional embodiment, in the above step, calculating the remote sensing index parameter of the changed pixel to be measured includes:

Calculate the NDISI index of the pixel to be tested for the change;

Calculate the green component of the tasseled cap transformation of the pixel to be measured;

The NDISI index and greenness component of the pixel to be measured are normalized to obtain the remote sensing index parameter of the pixel to be measured.

As an optional embodiment, in the above step, judging whether the remote sensing index parameter of the pixel to be detected changes within a preset time period includes:

The remote sensing index parameters are divided into multiple levels;

It is determined whether the level of the remote sensing index parameter of the pixel to be measured changes within a preset time period.

As an optional embodiment, in the above step, determining the number of change periods of the change pixel to be detected within the time period includes:

The number of time periods in which the level of the remote sensing index parameter of the change pixel to be measured changes within the time period cycle is determined to obtain the number of change periods of the change pixel to be measured.

As an optional embodiment, in the above steps, determining whether the changed pixel to be detected is a relatively stable pixel or a changed pixel according to the number of change periods and a preset land cover type change detection method includes:

Determine whether the number of change periods is 1, if not, determine the change pixel to be detected as a change pixel;

If so, based on a preset land cover type change detection method, it is determined that the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional embodiment, in the above steps, based on a preset land cover type change detection method, determining that the changed pixel to be detected is a relatively stable pixel or a changed pixel includes:

Obtain the land cover data of the first and last periods in the time period cycle corresponding to the target surface area;

Use a 3*3 moving window to detect the types of land cover data in the first and last periods, and determine whether the type of the pixel to be tested in the land cover data has changed during the change period;

If yes, then the changed pixel to be detected is determined to be a changed pixel;

Otherwise, it is determined that the changed pixel to be tested is a relatively stable pixel.

In a specific implementation scheme, the Level-2/Tier-1 Landsat surface reflectance product released by the Earth Resources Observation Science Center (EROS-USGS) of the United States Geological Survey is used. This product has been processed with geometric precision correction (error ≤ 12m), radiometric calibration and atmospheric correction (LEDAPS algorithm based on the 6S transmission model), and can generally be used for multi-time series analysis.

In this scheme, after obtaining the main and auxiliary data sources and performing data preprocessing, the main research is to establish a multi-time series pixel database through pixel change detection. The main implementation steps are as follows: 1. Visually identify stable/changing pixel samples; 2. Establish a stable pixel recognition model based on remote sensing index; 3. Divide the relative stability period of changing pixels based on time series fitting algorithm.

First, random points are generated in the study area, and stable/changing pixel samples are obtained through visual interpretation with reference to field survey data and high/medium resolution remote sensing images. This step only requires judging whether the pixel has changed or not, without identifying its geographical attributes, and urban areas are convenient for field verification. Therefore, sufficient samples can be obtained quickly and accurately, with a maximum of no more than 10,000 image pixels and no more than 5% of the study area.

Due to the heterogeneity of land cover distribution in urban areas, most pixels in medium-resolution images are basically a mixture of impervious surfaces, vegetation, and soil. Therefore, the focus is on changes related to these key features, and the impervious surface index (NDISI) and the tasseled cap change greenness component are used to detect pixel changes related to artificial surfaces and vegetation in remote sensing. Unlike the normalized difference between building and bare soil index (NDBSI) or other indices related to soil or bare land, NDISI significantly excludes the factors of soil changes and focuses on the spectral changes of pixels caused by artificial surfaces. It can be calculated by the following formula:

Where ρ _NIR , ρ _SWIR1 and ρ _TIR are the 5th, 6th and 11th bands of Landsat 8 OLI image, respectively, and MNDWI (modified normalized difference water index) can be calculated by the following formula:

Where ρ _green and ρ _SWIR1 are the 3rd and 6th bands of Landsat 8 OLI image, respectively.

The greenness component obtained by the tasseled cap transform is used to identify pixel changes related to vegetation. The tasseled cap transform is a linear transformation based on multispectral images that enhances the physical characteristics of the image by reducing the correlation between bands. The first few components obtained after the transformation are closely related to the surface landscape, among which the greenness component is highly sensitive to the state and changes of vegetation and is widely used for vegetation monitoring. For Landsat 8OLI images, the greenness component of the tasseled cap transform is calculated using the following formula:

Greenness＝-0.2941ρ _blue -0.243ρ _green -0.5424ρ _red +0.7276ρ _NIR +0.0713ρ _SWIR1 -0.1608ρ _SWI22 ；

Where ρ _blue , ρ _green , ρ _red , ρ _NIR , ρ _SWIR and ρ _SWIR2 are the 2nd to 7th bands of Landsat 8OLI images, respectively.

Next, the above two indicators are normalized and the pixels are divided into ten levels. If the level of the NDISI and greenness component of a pixel remains unchanged throughout the study period, it has no change period and is identified as a stable pixel; the period when the level changes is defined as a change period. A pixel is identified as a change pixel if it has two or more change periods. Then, land cover data is used to identify the pixel category with only one change period. Using the land cover data of the first and last two periods, a 3*3 moving window is used to improve the sensitivity of land cover change detection.

Based on the identification of stable pixels and the division of relatively stable periods of changing pixels, the Landsat series of surface reflectance images can be reconstructed, a multi-time series pixel database can be established, and the changes in pixel spectra can be recorded year by year.

Using the open source Cubis tool, the reflectance of each band of the medium-resolution image in the training sample was used as a variable, and the impervious density was used as an independent variable to establish a classification and regression tree model to estimate the multi-temporal impervious density of the entire study area. Among them, the impervious density of stable pixels remained unchanged during the study period. Therefore, a classification and regression tree model calculation can be established based on the spectral reflectance of the image during the entire study period. For the changing pixels that have been delineated for the changing period and the relatively stable period, their impervious density in the relatively stable period remains unchanged, similar to the stable pixels. Therefore, the spectral reflectance of the image in the corresponding period is used to estimate the impervious density. For the impervious density of the changing pixels in the changing period, a classification and regression tree model is established based on the image reflectance of the corresponding year to model and estimate it separately. Thus, the multi-temporal distribution of artificial surface changes at the sub-pixel scale of the entire study area is obtained.

The problem of the existing technology that this scheme is intended to solve is that there are compound errors in the existing sub-pixel method for estimating impervious surfaces, and abnormal changes that do not match the actual situation often occur when conducting local multi-time series analysis. In addition, the fixed low temporal resolution also affects the effectiveness of multi-time series monitoring. In the vast majority of existing studies, temporary and drastic changes in the urbanization process cannot be reflected, such as the existence or abandonment of construction sites and temporary factories, the demolition and construction of urban villages, etc. For long-term changes, the exact time point when the change occurs is also difficult to determine. The limitations of the sub-pixel method for estimating impervious surfaces in multi-time series monitoring are the key obstacles to its wider application. In this regard, this scheme is achieved through the following key points:

First, analyze multi-time series remote sensing images through change detection algorithms, divide image pixels into changing pixels and stable pixels, analyze the spectral change characteristics of changing pixels, and divide the relatively stable period;

Second, change the traditional remote sensing method of modeling and estimating single-period images, and aggregate all stable pixels and changing pixels in a relatively stable period, and reconstruct a multi-time series pixel database based on the changes in pixels.

Accordingly, the above solution of the present invention has the following advantages:

1. The solution of the present invention can realize the estimation of impervious surface at the sub-pixel level, deeply analyze the composition of ground objects inside the pixel based on the existing spatial resolution, and can more accurately reflect the spatial distribution and changes of the impervious surface;

2. The solution of the present invention can comprehensively utilize the advantages of the classification and regression tree and linear spectral hybrid analysis algorithms, eliminate or reduce the defects and limitations of the two algorithms through error analysis and correction, and improve the accuracy and stability of impervious surface estimation;

3. The solution of the present invention can adapt to remote sensing images of different types and different spectral characteristics, and has strong versatility and adaptability.

The feature of this solution is that it focuses on the key factors that cause errors in multi-time series impervious surface monitoring, starting from accidental errors, systematic errors to compound errors, comprehensively optimizing the technical process of multi-time series impervious surface monitoring and improving the reliability of monitoring results. Specifically, the innovative points of this project are as follows:

First, a coupling optimization method is proposed based on the error source analysis, which will reduce the systematic error of the CART model and further improve the existing single-period impervious surface estimation accuracy;

Second, the data source of traditional remote sensing image data sets is reconstructed, and impervious surface monitoring is carried out based on the multi-time series pixel database. This can not only effectively reduce the noise and abnormal pixels of single-period images and the accidental errors of single-period estimation results, but also avoid the compound errors caused by the superposition of multiple-period results, significantly improve the temporal resolution and monitoring accuracy, and be consistent with the actual situation of impervious surface changes.

The solution of the present invention is based on an objective pixel change detection method, improves the sub-pixel monitoring method of multi-time series impervious surfaces, and has the following advantages over the prior art:

First, the invention uses a pixel change detection method, which can significantly reduce multi-time series composite errors without a large amount of auxiliary data and cumbersome manual operations, and can be used for monitoring urban artificial surfaces when data sources are limited;

Second, the data source of traditional remote sensing image data sets is reconstructed, and impervious surface monitoring is carried out based on the multi-time series pixel database. This can not only effectively reduce the noise and abnormal pixels of single-period images and the accidental errors of single-period estimation results, but also avoid the compound errors caused by the superposition of multi-period results, significantly improve the temporal resolution and monitoring accuracy, and be consistent with the actual situation of impervious surface changes, providing reliable basic data for urban environment-related research.

Embodiment 2

Please refer to FIG. 2, which is a schematic diagram of the structure of an estimation algorithm optimization device based on pixel change detection disclosed in an embodiment of the present invention. The estimation algorithm optimization device based on pixel change detection described in FIG. 2 is applied to a data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). As shown in FIG. 2, the estimation algorithm optimization device based on pixel change detection may include:

An acquisition module 201 is used to acquire training sample images of a target surface area in multiple time periods;

The recognition module 202 is used to recognize stable pixels and change pixels to be detected from the training sample images of multiple time periods based on the pixel change recognition algorithm model;

A detection module 203 is used to detect relatively stable pixels and changed pixels from the changed pixels to be detected based on a pixel change period detection algorithm model;

The first aggregation module 204 is used to aggregate the band reflectance information of all stable pixels and relatively stable pixels and the corresponding stable time periods to obtain a stable training database; the stable training database is used to train an estimation algorithm model for estimating the impermeable density parameter of the target surface area in the stable time period;

The second aggregation module 205 is used to aggregate the band reflectance information of all changed pixels in the same change time period to obtain a change training database; the change training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area in the corresponding change time period.

As an optional embodiment, the specific method of the recognition module identifying stable pixels and change pixels to be detected from the training sample images of multiple time periods based on the pixel change recognition algorithm model includes:

Inputting training sample images of multiple time periods into a trained stable pixel recognition machine learning model to obtain recognized stable pixels; the stable pixel recognition machine learning model is trained by a training data set including multiple training images and corresponding visual change annotations;

The pixels other than stable pixels in the training sample images are determined as change pixels to be tested.

As an optional embodiment, the specific manner in which the detection module 203 detects relatively stable pixels and changing pixels from the changing pixels to be detected based on the pixel change period detection algorithm model includes:

For any pixel to be measured with changes, the remote sensing index parameter of the pixel to be measured with changes is calculated;

Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period;

If not, the pixel to be tested is determined as a relatively stable pixel;

If yes, then determine the number of change periods of the pixel to be tested within the time period;

According to the number of change periods and the preset land cover type change detection method, it is determined whether the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional embodiment, the specific method in which the detection module 203 calculates the remote sensing index parameter of the changed pixel to be detected includes:

Calculate the NDISI index of the pixel to be tested for the change;

Calculate the green component of the tasseled cap transformation of the pixel to be measured;

The NDISI index and greenness component of the pixel to be measured are normalized to obtain the remote sensing index parameter of the pixel to be measured.

As an optional embodiment, the specific manner in which the detection module 203 determines whether the remote sensing index parameter of the pixel to be detected changes within a preset time period includes:

The remote sensing index parameters are divided into multiple levels;

It is determined whether the level of the remote sensing index parameter of the pixel to be measured changes within a preset time period.

As an optional embodiment, the specific manner in which the detection module 203 determines the number of change periods of the change-to-be-detected pixel within the time period includes:

The number of time periods in which the level of the remote sensing index parameter of the change pixel to be measured changes within the time period cycle is determined to obtain the number of change periods of the change pixel to be measured.

As an optional embodiment, the detection module 203 determines the specific manner of determining whether the changed pixel to be detected is a relatively stable pixel or a changed pixel according to the number of change periods and a preset land cover type change detection method, including:

Determine whether the number of change periods is 1, if not, determine the change pixel to be tested as a change pixel;

If so, based on a preset land cover type change detection method, it is determined that the change pixel to be detected is a relatively stable pixel or a change pixel.

As an optional embodiment, the specific manner in which the detection module 203 determines whether the changed pixel to be detected is a relatively stable pixel or a changed pixel based on a preset land cover type change detection method includes:

Obtain the land cover data of the first and last periods in the time period cycle corresponding to the target surface area;

Use a 3*3 moving window to detect the types of land cover data in the first and last periods, and determine whether the type of the pixel to be tested in the land cover data has changed during the change period;

If yes, then the changed pixel to be detected is determined to be a changed pixel;

Otherwise, it is determined that the changed pixel to be tested is a relatively stable pixel.

Embodiment 3

Please refer to FIG. 3, which is another estimation algorithm optimization device based on pixel change detection disclosed in an embodiment of the present invention. The estimation algorithm optimization device based on pixel change detection described in FIG. 3 is applied to a data processing chip, a processing terminal or a processing server (wherein the processing server may be a local server or a cloud server). As shown in FIG. 3, the estimation algorithm optimization device based on pixel change detection may include:

A memory 301 storing executable program codes;

a processor 302 coupled to the memory 301;

The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the estimation algorithm optimization method based on pixel change detection described in the first embodiment.

Embodiment 4

An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the estimation algorithm optimization method based on pixel change detection described in the first embodiment.

Embodiment 5

An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps of the estimation algorithm optimization method based on pixel change detection described in Example 1.

The above describes specific embodiments of the present specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily have to be performed in the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer-readable storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

The apparatus, equipment, non-volatile computer-readable storage medium and method provided in the embodiments of this specification correspond to each other, and therefore, the apparatus, equipment, and non-volatile computer storage medium also have beneficial technical effects similar to those of the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the corresponding apparatus, equipment, and non-volatile computer storage medium will not be repeated here.

In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and/or hardware.

Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification may be in the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the embodiments of this specification may be in the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

This specification is described with reference to the flowchart and/or block diagram of the method, device (system), and computer program product according to the embodiment of this specification. It should be understood that each process and/or box in the flowchart and/or block diagram, as well as the combination of the process and/or box in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

In a typical configuration, a computing device includes one or more processors (CPU), input/output interfaces, network interfaces, and memory.

The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and/or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

Finally, it should be noted that the estimation algorithm optimization method and device based on pixel change detection disclosed in the embodiment of the present invention discloses only the preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (8)

1. A method for optimizing an estimation algorithm based on pixel change detection, characterized in that the method comprises: Obtain training sample images of the target surface area at multiple time periods; Based on the pixel change recognition algorithm model, identifying stable pixels and change pixels to be detected from the training sample images of the multiple time periods; Based on the pixel change period detection algorithm model, relatively stable pixels and changed pixels are detected from the changed pixels to be detected; based on the pixel change recognition algorithm model, stable pixels and changed pixels to be detected are identified from the training sample images of the multiple time periods, including: Inputting the training sample images of the plurality of time periods into a trained stable pixel recognition machine learning model to obtain recognized stable pixels; the stable pixel recognition machine learning model is trained by a training data set including a plurality of training images and corresponding visual change annotations; Determining the pixels other than the stable pixels in the training sample image as the change pixels to be detected; detecting relatively stable pixels and change pixels from the change pixels to be detected based on the pixel change period detection algorithm model, including: For any of the changed pixel to be measured, calculating the remote sensing index parameter of the changed pixel to be measured; Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period; If not, the pixel to be tested is determined as a relatively stable pixel; If yes, then determine the number of change periods of the pixel to be tested within the time period; According to the number of change periods and a preset land cover type change detection method, determining whether the change pixel to be detected is a relatively stable pixel or a change pixel; Aggregating the band reflectance information and the corresponding stable time periods of all the stable pixels and the relatively stable pixels to obtain a stable training database; the stable training database is used to train an estimation algorithm model for estimating the impermeable density parameter of the target surface area in the stable time period; Aggregate the band reflectance information of all the changed pixels in the same change time period to obtain a change training database; the change training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area in the corresponding change time period. 2. The estimation algorithm optimization method based on pixel change detection according to claim 1, characterized in that the step of calculating the remote sensing index parameter of the pixel to be detected includes: Calculate the NDISI index of the pixel to be tested for the change; Calculate the green component of the tasseled cap transformation of the pixel to be measured; The NDISI index and the greenness component of the changed pixel to be measured are normalized to obtain the remote sensing index parameter of the changed pixel to be measured. 3. The estimation algorithm optimization method based on pixel change detection according to claim 2 is characterized in that the step of judging whether the remote sensing index parameter of the pixel to be detected changes within a preset time period comprises: Classifying the remote sensing index parameters into multiple levels; It is determined whether the level of the remote sensing index parameter of the pixel to be detected changes within a preset time period. 4. The estimation algorithm optimization method based on pixel change detection according to claim 3 is characterized in that the step of determining the number of change periods of the pixel to be detected within the time period comprises: The number of time periods in which the level of the remote sensing index parameter of the changed pixel to be measured changes within the time period cycle is determined to obtain the number of change periods of the changed pixel to be measured. 5. The estimation algorithm optimization method based on pixel change detection according to claim 4 is characterized in that the step of determining whether the pixel to be detected is a relatively stable pixel or a changing pixel based on the number of change periods and a preset land cover type change detection method comprises: Determine whether the number of change periods is 1, and if not, determine the changed pixel to be detected as a changed pixel; If so, based on a preset land cover type change detection method, it is determined that the change pixel to be detected is a relatively stable pixel or a change pixel. 6. The estimation algorithm optimization method based on pixel change detection according to claim 5 is characterized in that the method for detecting land cover type change based on a preset method for detecting land cover type change determines whether the pixel to be detected is a relatively stable pixel or a changing pixel, comprising: Acquire the land cover data of the first and last two periods in the time period cycle corresponding to the target surface area; Use a 3*3 moving window to detect the types of the land cover data of the first and last periods, and determine whether the type of the pixel to be detected in the land cover data changes during the change period; If yes, then the changed pixel to be detected is determined to be a changed pixel; Otherwise, it is determined that the changed pixel to be tested is a relatively stable pixel. 7. A device for optimizing an estimation algorithm based on pixel change detection, characterized in that the device comprises: An acquisition module, used to acquire training sample images of a target surface area at multiple time periods; The recognition module is used to recognize stable pixels and change pixels to be detected from the training sample images of the multiple time periods based on the pixel change recognition algorithm model; the specific manner in which the recognition module recognizes stable pixels and change pixels to be detected from the training sample images of the multiple time periods based on the pixel change recognition algorithm model includes: Inputting the training sample images of the plurality of time periods into a trained stable pixel recognition machine learning model to obtain recognized stable pixels; the stable pixel recognition machine learning model is trained by a training data set including a plurality of training images and corresponding visual change annotations; Determining the pixels in the training sample image other than the stable pixels as the change pixels to be tested; A detection module is used to detect relatively stable pixels and changing pixels from the changing pixels to be detected based on a pixel change period detection algorithm model; the specific method of the detection module detecting relatively stable pixels and changing pixels from the changing pixels to be detected based on the pixel change period detection algorithm model includes: For any of the changed pixel to be measured, calculating the remote sensing index parameter of the changed pixel to be measured; Determine whether the remote sensing index parameter of the pixel to be measured changes within a preset time period; If not, the pixel to be tested is determined as a relatively stable pixel; If yes, then determine the number of change periods of the pixel to be tested within the time period; According to the number of change periods and a preset land cover type change detection method, determining whether the change pixel to be detected is a relatively stable pixel or a change pixel; A first aggregation module is used to aggregate the band reflectance information and the corresponding stable time period of all the stable pixels and the relatively stable pixels to obtain a stable training database; the stable training database is used to train an estimation algorithm model for estimating the impermeable density parameter of the target surface area in the stable time period; The second aggregation module is used to aggregate the band reflectance information of all the changed pixels in the same change time period to obtain a change training database; the change training database is used to train an estimation algorithm model for estimating the impermeable density parameters of the target surface area in the corresponding change time period. 8. A device for optimizing an estimation algorithm based on pixel change detection, characterized in that the device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the estimation algorithm optimization method based on pixel change detection as described in any one of claims 1-6.