1. A rural garbage downscaling space-time distribution inversion method based on multi-temporal remote sensing analysis is characterized by comprising the following specific processes:

acquiring remote sensing data and space geographic data related to a research area in the boundary of a research village and town area, and uniformly acquiring space-time scale uniform related data from the remote sensing data and the space geographic data in a spatial scale manner;

the remote sensing data and the space geographic data related to the research area in the boundary of the research village and town area comprise: researching map base map and administrative boundary data, NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, WorldPop population distribution and settlement point data of the United nations, water body contour data of the whole country and OSM traffic network data;

cutting the space-time scale unified related data obtained in the step one to obtain space element division of a minimum space unit grid L x L m for garbage generation inversion refinement modeling;

preprocessing the spatio-temporal scale unified related data in each grid divided in the step two to obtain a characteristic value of each data in the spatio-temporal scale unified related data;

step four, integrating all the characteristic values obtained in the step three, uniformly summing all the characteristic values in each grid contained in the researched towns, using the sum as a group of independent variables in the garbage inversion prediction process, using the garbage yield of the towns as a dependent variable, obtaining an analysis data set in the inversion analysis process formed by the independent variables and the dependent variable, then compiling an LSTM model, and training the LSTM model by using the analysis data set to obtain the trained LSTM model, namely obtaining the correlation between the domestic garbage yield and each space geographic characteristic;

and step five, bringing the correlation relationship between the domestic garbage yield obtained in the step four and each space geographic characteristic into each grid divided in the step two, then inputting the remote sensing data and the space geographic data related to the research area of the last year into the trained LSTM deep learning model to obtain the annual garbage yield of each L × L m grid space range, wherein the annual garbage yields of all grids form the space-time distribution downscaling inversion result of the domestic garbage.

2. The town rubbish scale-reduction space-time distribution inversion method based on the multi-temporal remote sensing analysis as claimed in claim 1, is characterized in that: in the first step, the spatial scale of the remote sensing data and the spatial geographic data is uniformly acquired to obtain the uniform relevant data of the space-time scale, and the method comprises the following steps:

cutting the acquired NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, Union world population distribution and settlement point data, national water body profile data and OSM traffic network data by using the village and town level administrative boundary profile of the research area to obtain related data with uniform space-time scale in the boundary of the research area;

and cutting the acquired data by utilizing the village and town level administrative boundary contour of the research area by using an ArcGIS software cutting function.

3. The town rubbish scale-reduction space-time distribution inversion method based on the multi-temporal remote sensing analysis as claimed in claim 2, wherein: in the second step, the space-time scale unified relevant data obtained in the first step is cut to obtain the space element division of L L m in the research villages and towns, and the method comprises the following steps:

and cutting the relevant data after unified space-time scale by using the boundary of the village-town level administrative region of the research region to form a minimum space unit grid L L m for inversion and fine modeling of garbage generation, and then planing off the water body by using the water body contour data.

4. The town rubbish scale-reduction space-time distribution inversion method based on multi-temporal remote sensing analysis as claimed in claim 3, wherein: in the third step, preprocessing the spatio-temporal scale unified related data in each grid divided in the second step to obtain the characteristic value of each data in the spatio-temporal scale unified related data, and the method comprises the following steps:

thirdly, carrying out radiometric calibration, atmospheric correction, orthorectification and coordinate correction pretreatment on NPP/VIIRS night light data, and then carrying out data resampling, noise reduction and normalization on the pretreated night light data to obtain night light data analysis characteristics;

step two, cutting Landset8 land utilization data by using the spatial grids divided in the step two to obtain gridded satellite data conditions, then calculating the occupation ratios of water bodies, vegetations, buildings and bare lands in each grid, and performing normalization operation on all the occupation ratios to obtain land utilization data analysis characteristics;

the Landset8 land utilization data represent the land use condition in a village and town scene;

thirdly, acquiring the average elevation, the maximum elevation and the minimum elevation of the Aster surface elevation data in the space grid divided in the second step, then performing reclassification operation on slope data in the elevation data, and then performing normalization processing on the reclassified data to acquire surface elevation data analysis characteristics;

step four, respectively adding population and population total in the WorldPop population distribution and settlement point data of the United nations in the space grid range divided in the step two, and carrying out data normalization operation to obtain population distribution and settlement point data analysis characteristics;

step three, breaking the OSM traffic network by using the grid lines of the spatial grids obtained in the step two, obtaining the length weighted total value in each spatial grid, and normalizing the weighted sum of all the lengths to obtain the traffic network data analysis characteristics;

the length weighted total value in the spatial grid is used for representing road transport capacity.

5. The town rubbish scale-reduction space-time distribution inversion method based on multi-temporal remote sensing analysis as claimed in claim 4, wherein: in the third step, data resampling, noise reduction and normalization operations are performed on the preprocessed night light data to obtain analysis characteristics of the night light data, and the method comprises the following steps:

and performing data resampling operation on the preprocessed night light data in the space grids divided in the step two, positioning the data precision to 500m, performing data noise reduction on the areas with the negative night light, calculating the night light intensity value in each grid of all the grids subjected to noise reduction, and performing normalization operation on all the night light intensity values to obtain the analysis characteristics of the subsequent garbage yield space distribution inversion process.

6. The town rubbish scale-reduction space-time distribution inversion method based on the multi-temporal remote sensing analysis as claimed in claim 5, wherein: ENVI5.3.1 software is adopted for carrying out radiometric calibration, atmospheric correction, orthotropic correction and coordinate correction pretreatment on NPP/VIIRS night light data in the third step.

7. The town rubbish scale-reduction space-time distribution inversion method based on the multi-temporal remote sensing analysis as claimed in claim 6, wherein: in the third step, normalization processing is performed on the heavily classified data to obtain the analysis characteristics of the earth surface elevation data, and the method comprises the following steps:

calculating the average gradient of the gradient data after reclassification by using a GIS surface analysis function, and then performing normalization operation on the average gradient to be used as an analysis characteristic of the surface elevation data;

the gradient data after reclassification operation is performed on the gradient data in the elevation data is as follows: the sloping direction to the north is 1; northeast and northwest of 3; east, west is 5; southeast, southwest and south are 7.

8. The rural garbage downscaling space-time distribution inversion method based on multi-temporal remote sensing analysis according to claim 7, is characterized in that: and in the third step, the population distribution of the WorldPop in the United nations and the population sum of the settlement points in the data of the settlement points in the space grid range divided in the second step are respectively added, and a GIS space counting method is adopted.

9. The rural garbage downscaling space-time distribution inversion method based on multi-temporal remote sensing analysis according to claim 8, characterized in that: and the length weighted total value in each space grid in the step III and V is obtained by multiplying the number of the road lanes by the length of the road.

10. The rural garbage downscaling space-time distribution inversion method based on multi-temporal remote sensing analysis according to claim 9, is characterized in that: and in the second step, L is 500.

Background

The planning management of the domestic garbage is developed at home and abroad to become an important new industry, and the wide attention is attracted. In order to meet the requirement of fine management construction of a sanitary environment in a village and town scene, the magnitude of the domestic garbage related data needs to be refined in time and space scales so as to meet the management requirement of per capita daily domestic garbage yield and the coverage range of a transfer system in the non-waste city construction index of China and ensure the urban and rural integrated global coverage of the service and management functions of the domestic garbage, so that the data analysis and yield prediction of the domestic garbage become the research focus in the field at present.

At present, comprehensive analysis of remote sensing data is widely applied to the processes of atmospheric quality inversion and river water flow inversion, but application research aiming at environmental sanitation macro management is of little concern; the method is characterized in that the distribution characteristics of domestic garbage space generation are not clear in the high-speed development scene of cities and towns in China, basic management data are lacked, the practical situation and the accurate management of data management broken links are realized, the management scale is more in villages, the analysis scale resolution ratio of garbage space-time change is low, and the current situation that the current rural areas in China develop at high speed and the population flows greatly is not met. That is to say, the existing data analysis and yield prediction of the garbage are difficult due to the small garbage generation amount, the scattered garbage distribution and the complex components in rural areas in China, so that the problem that the advanced layout of the household garbage collection and transportation setting device in the villages and the small towns cannot be realized is caused.

Disclosure of Invention

The invention aims to solve the problem that the data analysis and the yield prediction of garbage are difficult at present and further the advanced layout of a garbage collection and transportation setting device of a village and a town cannot be realized due to the fact that no inversion method for the garbage size space-time distribution of the village and the town exists at present, and provides a garbage size-reduction space-time distribution inversion method based on multi-temporal remote sensing analysis.

The specific process of the town rubbish downscaling space-time distribution inversion method based on the multi-temporal remote sensing analysis comprises the following steps:

the method comprises the steps of firstly, obtaining remote sensing data and space geographic data related to a research area in the boundary of a village and town area, and uniformly obtaining space-time scale uniform related data by uniformly obtaining the space scale of the remote sensing data and the space geographic data;

the remote sensing data and the spatial geographic data related to the research area within the research area boundary comprise: researching map base map and administrative boundary data, NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, WorldPop population distribution and settlement point data of the United nations, water body contour data of the whole country and OSM traffic network data;

step two, cutting the space-time scale unified relevant data obtained in the step one to obtain space element division of L L m in the research villages and towns;

preprocessing the spatio-temporal scale unified related data in each grid divided in the step two to obtain a characteristic value of each data in the spatio-temporal scale unified related data;

step four, synthesizing all the characteristic values obtained in the step three, uniformly summing all the characteristic values obtained in the step three in all grids contained in the research villages and towns, using the sum as a group of independent variables in a garbage inversion prediction process, using the garbage yield of the villages and towns as a dependent variable, obtaining an analysis data set in an inversion analysis process formed by the independent variables and the dependent variable, then compiling a model LSTM model, and obtaining a trained LSTM model, namely obtaining the correlation between the domestic garbage yield and each spatial geographic characteristic;

and step five, bringing the correlation between the domestic garbage yield obtained in the step four and each space geographic characteristic into each grid divided in the step two, inputting the remote sensing data and the space geographic data related to the research area of the last year into the trained LSTM deep learning model to obtain the annual garbage yield of each L × L m grid space range after data calculation, wherein the annual garbage yields of all grids form the space-time distribution downscaling inversion result of the domestic garbage.

The invention has the beneficial effects that:

according to the method, based on the condition of data base shortage in a village and town scene, Python language and ArcGIS software are combined, spatial geographic data such as land utilization, night light, water, elevation and a traffic network which are easy to obtain are utilized, multi-source heterogeneous data fusion is carried out, the spatial distribution characteristics of the domestic garbage in the village and town are objectively and scientifically inverted by carrying out characteristic delineation on relevant influence elements of the domestic garbage in a research area, then the incidence relation between the spatial remote sensing data and the domestic garbage yield is established by using a deep learning algorithm, the management scale of the domestic garbage is reduced to a space grid of 500m square from a village and town or a transfer station as a unit, and the inverse of the reduced scale spatial distribution of the domestic garbage in the village and town based on multi-temporal remote sensing analysis is obtained; based on the refinement of the domestic garbage output data management on the space-time scale, the extensive environmental sanitation operation data statistical mode taking villages and towns or transit points as management units is changed, and the comprehensiveness, the accuracy and the timeliness of the management are improved. According to the method, the distribution characteristics of the domestic garbage in the research area, which synthesizes various space influence factors, can be obtained through the operation result of the method, so that the method is used for carrying out data-driven analysis and explanation.

Drawings

FIG. 1 is a schematic flow chart of the present invention.

Detailed Description

The first embodiment is as follows: the method for inverting the scale-reduction space-time distribution of the garbage in the villages and the towns based on the multi-temporal remote sensing analysis comprises the following specific processes:

acquiring remote sensing data and space geographic data related to a research area in the boundary of a research village and town area, and uniformly acquiring space-time scale uniform related data from the remote sensing data and the space geographic data in a spatial scale manner;

the remote sensing data and the spatial geographic data related to the research area comprise: researching map base map and administrative boundary data, NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, WorldPop population distribution and settlement point data of the United nations, water body contour data of the whole country and OSM traffic network data;

NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, WorldPop population distribution and settlement point data of the United nations, water body contour data of the whole country and OSM traffic network data which are obtained by utilizing the village and town level administrative boundary contour of the research area are cut to obtain related data with uniform space-time scale in the boundary of the research area;

the acquired data is cut by utilizing the village and town level administrative boundary contour of the research area, and the data is cut by utilizing an ArcGIS software cutting function;

step two, cutting the relevant data after unifying the space-time scale by using a fishing net generating method for data management in an Arcgis10.7 software toolbox, and obtaining the space element division of 500 × 500m in the research area:

cutting relevant data after unified space-time scale by using the boundary of a village-town level administrative region of a research region to form a minimum space unit grid (500 x 500 m) for inversion and fine modeling of garbage generation, and then removing the water body by using water body contour data, namely domestic garbage cannot be generated on the water body;

step three, preprocessing the spatio-temporal scale unified relevant data in each grid divided in the step two to obtain each data characteristic value obtained in the step one, and the method comprises the following steps:

step three, preprocessing NPP/VIIRS night light data to obtain the analysis characteristics of the night light data:

the NPP/VIIRS night light data represent human mouth activity intensity and distribution conditions in a town scene;

firstly, ENVI5.3.1 software is used for carrying out a series of preprocessing such as radiometric calibration, atmospheric correction, orthotropic correction, coordinate correction and the like on NPP/VIIRS night light data to ensure the accuracy of the data.

And secondly, performing data resampling operation on the preprocessed night light data by utilizing the spatial grids divided in the second step, positioning the data precision to 500m, performing data noise reduction on the area with the negative night light, calculating the night light intensity value in each grid of all the grids after noise reduction, and performing normalization operation on all the night light intensity values to obtain the analysis characteristic of the subsequent garbage yield spatial distribution inversion process.

Step two, preprocessing Landset8 land use data to obtain land use data analysis characteristics:

the Landset8 land utilization data represent the land use condition in a village and town scene;

and secondly, calculating the occupation ratio of water, vegetation, buildings and bare land in each grid, and normalizing all occupation ratios to obtain the analysis characteristics of the subsequent garbage yield space distribution inversion process.

Thirdly, preprocessing the Aster surface elevation data to obtain surface elevation data analysis characteristics:

firstly, acquiring the average elevation, the maximum elevation and the minimum elevation of the Aster surface elevation data in the space grid divided in the step two, and performing reclassification operation on slope data in the elevation data because different slope directions of the surface can influence the residential characteristics of residents, wherein the north direction of the slope direction is 1; northeast and northwest of 3; east, west is 5; 7, calculating the average gradient of the gradient data after reclassification by using a GIS surface analysis function, and then performing normalization operation on the average gradient to serve as analysis characteristics;

step four, preprocessing the population distribution and the settlement point data of WorldPop in the United nations to obtain the analysis characteristics of the population distribution and the settlement point data:

respectively adding the population sum and the settlement point sum in the space grid range divided in the step two by using a GIS space statistical method, and performing data normalization operation to serve as an inversion analysis characteristic;

step three, preprocessing the OSM traffic network data to obtain traffic network data analysis characteristics:

breaking the traffic network by using the grid lines of the spatial grids obtained in the step two, obtaining a length weighted total value in each spatial grid, and normalizing the weighted sum of all the lengths to be used as an inversion analysis characteristic;

the total length weighted value in the space grid is obtained by multiplying the number of the lanes of the road by the length of the road;

the length weighted total value in the spatial grid is used for representing road transport capacity.

Step four, integrating all characteristic values obtained in the step three, uniformly adding all characteristic values in all grids contained in the researched towns and using the sum as a group of independent variables in the garbage inversion prediction process, using the garbage yield of the towns and the towns as a dependent variable, obtaining an analysis data set in the inversion analysis process formed by the independent variables and the dependent variable, then compiling a model by using Python language, training an LSTM model by using the data set, and then obtaining the correlation between the domestic garbage yield and each space geographic characteristic through the trained LSTM deep learning model;

and step five, bringing the correlation relationship between the domestic garbage yield obtained in the step four and each space geographic characteristic into each grid, inputting the remote sensing data and the space geographic data related to the research area of the last year into a trained LSTM deep learning model to obtain the annual garbage yield of each grid of 500 x 500m after data calculation, wherein the annual garbage yields of all the grids form a space-time distribution downscaling inversion result of the domestic garbage, and further can be used for assisting the resource allocation detail decision of the domestic garbage management process.

Example (b):

the method is applied to the space-time distribution inversion situation of domestic garbage in special administrative districts of hong Kong and surrounding towns of the special administrative districts in 2021, the current domestic garbage classification statistical types of hong Kong are divided into urban solid wastes (including household wastes and industrial and commercial wastes) and whole building wastes, the research is carried out on the severe situation that the landfill site of hong Kong is about to be fully loaded, four areas such as hong Kong island, Jiulong, New Yongquan, and Tao island and 18 small areas are taken as statistical units, the inversion research of garbage classification generation is carried out on the demand of exploring the space-time distribution of local garbage, the space-time distribution inversion situation of the domestic garbage in 2020 year is obtained, and decision support is provided for the layout of a future domestic garbage collection and storage system, and transportation system, and the specific process is as follows:

firstly, an official website is adopted in environment protection administration of special administrative districts of hong Kong, and the data of the garbage yield in 2012 and 2019 are downloaded, collected and sorted. A total of 704 data records of 4 types of domestic garbage in 8 years in 18 areas are obtained, and one of the data records is added into a map surface element field of a division boundary of a special administrative district of hong Kong to obtain vector map data with garbage yield.

Then, NPP/VIIRS night light data, Landset8 land utilization data, Aster surface elevation data, Union world population distribution and settlement point data, national water body contour data, OSM traffic network data and the like in 2012 and 2019 are collected in network channels such as a satellite website official network, a geospatial data cloud, a Union official network and the like, cutting is carried out by utilizing the administrative boundary contour of the hong Kong special administrative region, relevant data with unified space-time scale in the boundary of the research area is obtained, a geospatial database is established, and unified management is carried out by utilizing Arcgis10.7 software;

and (4) executing steps two to five in the first embodiment by using the data, and finally obtaining the space-time inversion result of the domestic garbage in the hong Kong special administrative district and the surrounding towns in 2020.