CN115601596B

CN115601596B - Drawing method and device for recognizing forest felling distribution after fire

Info

Publication number: CN115601596B
Application number: CN202211296896.3A
Authority: CN
Inventors: 许尔琪; 李科为
Original assignee: Institute of Geographic Sciences and Natural Resources of CAS
Current assignee: Institute of Geographic Sciences and Natural Resources of CAS
Priority date: 2022-10-21
Filing date: 2022-10-21
Publication date: 2023-04-07
Anticipated expiration: 2042-10-21
Also published as: CN115601596A

Abstract

The invention discloses a mapping method and a mapping device for identifying forest deforestation distribution after fire.A long-time sequence annual cloud-free optimal image is obtained by carrying out image synthesis on remote sensing image data of a target area; carrying out pixel screening on the long-time sequence annual cloud-free optimal image to obtain a mature forest pixel, and meanwhile, carrying out standardized processing on the spike-cap transformation brightness index and the spike-cap transformation humidity index of the mature forest pixel to calculate a forest disturbance index, thereby effectively solving the difficulty that the judgment basis of forest disturbance characteristics is insufficient; by setting the threshold value of the index of the forest disturbance caused by cutting, classifying each pixel in the forest region passing fire year by year according to the threshold value of the index of the forest disturbance caused by cutting and the index of the forest disturbance caused by cutting, the forest cutting behavior after fire can be effectively identified, the spatial distribution image of the forest cutting after fire is generated, and more accurate dynamic information of the forest after fire can be provided for a forest manager.

Description

Mapping method and device for recognition of post-fire forest felling distribution

Technical Field

The invention relates to the technical field of remote sensing mapping, in particular to a mapping method and a device for recognizing forest felling distribution after fire.

Background

The forest is one of the most important land ecosystems on the earth, is the main body of land carbon sink, stores about 45 percent of carbon in the land ecosystems, and provides important ecosystem services of climate regulation, biological diversity protection, water source conservation and the like. Due to the influence of climate change and human activities, the severity, frequency and disturbance range of natural and artificial disturbance of the forest are obviously increased, which may cause irreversible degradation of the forest ecosystem, and seriously affect the capability and sustainable development of the forest for providing ecosystem service. As one of the inherent attributes of the forest ecosystem, fire and deforestation constitute the main types of disturbances of the forest ecosystem, changing its composition, structure and function in a relatively discrete manner. The forest disturbance mapping based on remote sensing becomes an important means for monitoring the forest dynamics of a long-time sequence, accurately identifies the forest disturbance dynamics, and has important significance for knowing the current situation and the long-term development trend of a forest ecological system.

The current common forest felling identification method mainly comprises a comparison method after classification, an index difference method and a time sequence analysis method. However, the existing research can not distinguish the felling from other disturbance types well, when a plurality of forest disturbances occur continuously, the initial fire can reduce the reflectivity of the forest canopy obviously, so that the subsequent forest felling is difficult to distinguish, and especially when the two disturbances occur continuously and quickly, the possibility of spectrum confusion is higher, and the method has great difficulty in remote sensing identification.

In view of the lack of an effective method for detecting the forest felling after fire at present, the forest felling after fire can further destroy biological heritage, change the initial state of forest recovery, and further influence the direction and speed of forest succession, thereby causing deviation on the forest vegetation recovery effect evaluation and the subsequent forest management. Therefore, it is urgently needed to provide a mapping method for detecting forest deforestation after a fire.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: the mapping method and the mapping device for the post-fire forest cutting distribution recognition are provided, the influence of forest disturbance factors such as fire and cutting on remote sensing recognition is reduced, and the accuracy of the post-fire forest cutting behavior recognition is improved.

In order to solve the technical problem, the invention provides a drawing method and a drawing device for identifying the post-fire forest felling distribution, wherein the drawing method comprises the following steps:

acquiring and calculating a burning index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to remote sensing image data based on the remote sensing image data of a target area;

setting an initial date range of image synthesis, acquiring all target area remote sensing images within the initial date range in a plurality of years, and carrying out image synthesis on all the target area remote sensing images to obtain a long-time sequence annual cloud-free optimal image;

extracting an unfired forest region and a fired forest region from the long-time sequence annual cloud-free optimal image based on a preset burning index threshold and a differential burning index threshold, and performing pixel screening on the unfired forest region to obtain a mature forest pixel in the long-time sequence annual cloud-free optimal image;

acquiring and calculating a standard spike-cap conversion brightness index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of the forest with excessive fire in different years year by year based on the mean value and the standard deviation corresponding to the spike-cap conversion brightness index and the spike-cap conversion humidity index of the mature forest pixels, and obtaining a forest disturbance index corresponding to each pixel in the forest area of the forest with excessive fire according to the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index;

obtaining an un-felled sample area after fire and a felled sample area after fire, carrying out layered random sampling treatment on the un-felled sample area after fire and the felled sample area after fire to obtain sample pixels of the un-felled sample area after fire and sample pixels of the felled sample area after fire, calculating and setting a felled forest disturbance index threshold value based on sample forest disturbance indexes corresponding to the sample pixels of the un-felled sample area after fire and the sample pixels of the felled sample area after fire;

classifying each pixel in the forest region with excessive fire year by year according to the forest disturbance index and the forest disturbance index threshold value, obtaining the forest region with initial fire cutting corresponding to different years, and generating a spatial distribution image of the forest with initial fire cutting.

In a possible implementation manner, after generating the space-time distribution image of the initial post-fire forest felling, the method further comprises the following steps:

calculating differential forest disturbance indexes of all sample pixels of the non-felled sample area after fire and the sample pixels of the felled sample area after fire in adjacent years based on an image differential method, and setting a differential disturbance index threshold according to the differential forest disturbance indexes;

calculating a difference forest disturbance index corresponding to each pixel in the forest area with excessive fire in adjacent years based on the forest disturbance index corresponding to each pixel in the forest area with excessive fire;

classifying each pixel in the forest region of the forest which is over fire year according to the forest disturbance index threshold value and the difference forest disturbance index, obtaining a primary forest cutting region of the forest which is over fire and a secondary forest cutting region of the forest which is over fire, and correspondingly generating a primary forest cutting space-time distribution image and a secondary forest cutting space-time distribution image of the forest which is over fire.

In a possible implementation manner, after obtaining the forest regions after the initial fire corresponding to different years is cut, the method further includes:

setting each pixel in the area of the forest cut after the initial fire as a forest cut after the fire, acquiring first forest disturbance indexes corresponding to all forest cut after the fire in the current year, and acquiring second forest disturbance indexes corresponding to all forest cut after the fire in the preset year;

judging whether the disturbance difference value of the first forest disturbance index and the second forest disturbance index is within a preset difference value range, if not, determining that the pixels of the forest which is cut down after fire are classified wrongly, correcting the pixels of the forest which is cut down after fire into pixels of the forest which is not cut down after fire, and if so, determining that the pixels of the forest which is cut down after fire are classified correctly.

In a possible implementation manner, based on a preset burning index threshold and a differential burning index threshold, extracting an unfired forest region and a fired forest region from the long-time sequence annual cloud-free optimal image, specifically including:

acquiring a combustion index corresponding to each pixel in the long-time sequence annual cloud-free optimal image, and comparing the combustion index with the combustion index threshold;

if the burning index is larger than the burning index threshold value, setting the pixels as forest pixels, integrating all forest pixels to obtain a forest area, and if the burning index is smaller than or equal to the burning index threshold value, setting the pixels as non-forest pixels, and integrating all non-forest pixels to obtain a non-forest area;

calculating a differential combustion index corresponding to each pixel of the forest region in adjacent years, and comparing the differential combustion index with the differential combustion index threshold;

if the differential burning index is smaller than or equal to the differential burning index threshold value, setting the forest pixels as non-flaming forest pixels, integrating all the non-flaming forest pixels to obtain a non-flaming forest region, and if the burning index is larger than or equal to the burning index threshold value, setting the forest pixels as flaming forest pixels, integrating all the flaming forest pixels to obtain a flaming forest region.

In a possible implementation manner, pixel screening is performed on the forest region without fire to obtain mature forest pixels in the long-time sequence annual cloud-free optimal image, and the method specifically includes:

calculating the mean value and standard deviation of the combustion index of each forest pixel without fire in the forest area without fire in a plurality of years, and calculating the corresponding variation coefficient of each forest pixel without fire according to the mean value and standard deviation of the combustion index;

if the mean value of the combustion index of the forest pixels without fire passing is larger than a preset threshold value of the mean value of the combustion index and the coefficient of variation is smaller than a preset threshold value of the coefficient of variation, setting the forest pixels without fire passing as mature forest pixels;

and acquiring mature forest pixels in all target area remote sensing images for synthesizing the long-time sequence annual cloud-free optimal image scene by scene to obtain all mature forest pixels in the long-time sequence annual cloud-free optimal image.

In a possible implementation manner, the method includes the steps of acquiring year by year and calculating a standard spike-cap conversion brightness index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of the forest which has passed fire in different years based on a mean value and a standard deviation corresponding to the spike-cap conversion brightness index and the spike-cap conversion humidity index of the mature forest pixels, and specifically includes the following steps:

acquiring all target area remote sensing images corresponding to the current year in the synthesized long-time sequence annual cloud-free optimal images year by year, calculating the spike-cap transformation brightness index and the mean value and the standard deviation corresponding to the spike-cap transformation humidity index of all mature forest pixels in each target area remote sensing image, and obtaining the spike-cap transformation brightness index mean value, the spike-cap transformation humidity index mean value, the spike-cap transformation brightness index standard deviation and the spike-cap transformation humidity index standard deviation corresponding to each target area remote sensing image;

acquiring each fire forest pixel of the fire forest region in each target region remote sensing image, and acquiring a fire spike cap transformation brightness index and a fire spike cap transformation humidity index corresponding to each fire forest pixel;

inputting the over-fire spike-cap conversion brightness index, the spike-cap conversion brightness index mean value and the spike-cap conversion brightness index standard deviation into a preset standard spike-cap conversion brightness index calculation formula to obtain a standard spike-cap conversion brightness index;

and inputting the over-fire spike-cap transformation humidity index, the spike-cap transformation humidity index mean value and the spike-cap transformation humidity index standard deviation into a preset standard spike-cap transformation humidity index calculation formula to obtain a standard spike-cap transformation humidity index.

In a possible implementation manner, obtaining a forest disturbance index corresponding to each pixel in the forest area according to the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index specifically includes:

inputting the standard spike cap conversion brightness index and the standard spike cap conversion humidity index into a forest disturbance index calculation formula to obtain a forest disturbance index corresponding to each fire forest pixel in the fire forest area, wherein the forest disturbance index calculation formula is as follows:

wherein mDI is the forest disturbance index, B _n Conversion of the luminance index, W, for the standard spike-cap _n Standard spike cap shift humidity index.

In a possible implementation manner, setting a threshold of the index of occurrence of deforestation based on each sample pixel of the sample area not deforested after fire and a sample forest disturbance index corresponding to each sample pixel of the sample area deforested after fire specifically includes:

acquiring sample forest disturbance indexes corresponding to sample pixels in a sample area which is not felled after fire and sample pixels in a sample area which is felled after fire, sequencing all the sample forest disturbance indexes from low to high, and establishing a first numerical value histogram of the sample forest disturbance indexes;

selecting a forest disturbance index value of more than 90% of the sample forest disturbance indexes in the first numerical value histogram, and setting the forest disturbance index value as a deforestation index threshold value.

In one possible implementation, after generating the initial post-fire forest felling space-time distribution image, and the secondary post-fire forest felling space-time distribution image, the method further includes:

performing spatial filtering processing on the initial post-fire forest cutting space-time distribution image to obtain an initial post-fire forest cutting space-time distribution filtering image;

and performing spatial filtering processing on the space-time distribution image of the primary felling of the forest after the fire and the space-time distribution image of the secondary felling of the forest after the fire to obtain a space-time distribution filtering image of the primary felling of the forest after the fire and a space-time distribution filtering image of the secondary felling of the forest after the fire.

In a possible implementation manner, the method for calculating the combustion index, the spike-cap transformation luminance index and the spike-cap transformation humidity index based on the remote sensing image data of the target area comprises the following steps:

acquiring remote sensing image data of a target area in multiple years, wherein the remote sensing image data comprise remote sensing images with cloud amount less than 50% provided by a Landsat5 satellite TM sensor and a Landsat7 satellite ETM + sensor;

performing mask processing on low-quality pixels in the remote sensing image data based on a QA wave band, wherein the low-quality pixels comprise cloud pixels and cloud shadow pixels;

calculating a combustion index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to each pixel in each remote sensing image based on remote sensing image multi-band data in the remote sensing image data after mask processing, wherein the remote sensing image multi-band data are Landsat 5/7 wave band from 1 st wave band to 5 th wave band and 7 th wave band.

In a possible implementation manner, a calculation formula for calculating a combustion index, a spike-cap conversion brightness index and a spike-cap conversion humidity index corresponding to each pixel in each remote sensing image is specifically as follows:

formula for calculating the burning index:

NBR＝(NIR–SWIR2)/(NIR+SWIR2)；

the calculation formula of the spike-cap transformation brightness index is as follows:

TCB＝0.2043×Blue+0.4158×Green+0.5524×Red+0.5741×NIR+0.3124×SWIR1+0.2303×SWIR2；

the calculation formula of the spike cap transformation humidity index is as follows:

TCW＝0.0315×Blue+0.2021×Green+0.3102×Red+0.1594×NIR-0.6806×SWIR1+0.6109×SWIR2；

in the formula, blue, green, red, NIR, SWIR1 and SWIR2 are respectively wave band 1-Blue, wave band 2-Green, wave band 3-Red, wave band 4-near infrared, wave band 5-short wave infrared 1 and wave band 7-short wave infrared 2 of Landsat5TM/7ETM + images.

The invention also provides a mapping device for identifying the post-fire forest felling distribution, which comprises the following components: the system comprises a remote sensing image data acquisition module, a long-time sequence annual cloud-free optimal image synthesis module, a mature forest pixel screening module, a forest disturbance index calculation module, a deforestation index threshold value setting module and an initial fire-post forest deforestation space-time distribution image generation module;

the remote sensing image data acquisition module is used for acquiring and calculating a combustion index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to remote sensing image data based on remote sensing image data of a target area;

the long-time sequence annual cloud-free optimal image synthesis module is used for setting an initial date range of image synthesis, acquiring all target area remote sensing images in the initial date range in a plurality of years, and carrying out image synthesis on all the target area remote sensing images to obtain a long-time sequence annual cloud-free optimal image;

the mature forest pixel screening module is used for extracting an unfired forest region and an unfired forest region from the long-time sequence annual cloud-free optimal image based on a preset burning index threshold and a difference burning index threshold, and performing pixel screening on the unfired forest region to obtain a mature forest pixel in the long-time sequence annual cloud-free optimal image;

the forest disturbance index calculation module is used for acquiring year by year, calculating a standard spike-cap conversion brightness index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of the forest with excessive fire in different years based on a mean value and a standard deviation corresponding to the spike-cap conversion brightness index and the spike-cap conversion humidity index of the mature forest pixels, and obtaining a forest disturbance index corresponding to each pixel in the forest with excessive fire according to the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index;

the threshold value setting module for the index of the disturbance of the deforested forest is used for acquiring an un-deforested sample area after fire and a deforested sample area after fire, carrying out layering random sampling treatment on the un-deforested sample area after fire and the deforested sample area after fire to obtain sample pixels of the un-deforested sample area after fire and sample pixels of the deforested sample area after fire, calculating and setting the threshold value of the index of the disturbance of the deforested forest based on the sample pixels of the un-deforested sample area after fire and the sample pixels corresponding to the sample pixels of the deforested sample area after fire;

and the initial post-fire forest cutting space-time distribution image generation module is used for classifying each pixel in the over-fire forest area year by year according to the forest disturbance index and the generated cutting forest disturbance index threshold value to obtain initial post-fire cutting forest areas corresponding to different years and generate initial post-fire forest cutting space-time distribution images.

In a possible implementation manner, the mapping apparatus for identifying the post-fire forest felling distribution provided by the invention further comprises: the device comprises a module for generating a space-time distribution image for the primary felling of the forest after the fire and a space-time distribution image for the secondary felling of the forest after the fire;

the device comprises a post-fire forest primary cutting space-time distribution image generation module, a post-fire forest secondary cutting space-time distribution image generation module and a post-fire forest secondary cutting space-time distribution image generation module, wherein the post-fire forest primary cutting space-time distribution image generation module is used for calculating differential forest disturbance indexes of all post-fire uncut sample area sample pixels and all post-fire cut sample area sample pixels in adjacent years based on an image differential method, and setting a differential disturbance index threshold value according to the differential forest disturbance indexes;

the post-fire forest primary cutting space-time distribution image and the post-fire forest secondary cutting space-time distribution image generating module are used for calculating a difference forest disturbance index corresponding to each pixel in the post-fire forest area in adjacent years based on the forest disturbance index corresponding to each pixel in the post-fire forest area;

the module for generating the space-time distribution image of the primary cutting of the forest after the fire and the space-time distribution image of the secondary cutting of the forest after the fire is used for classifying each pixel in the forest passing area year by year according to the threshold value of the disturbance index of the cutting forest and the disturbance index of the difference forest, obtaining the area of the primary cutting of the forest after the fire and the area of the secondary cutting of the forest after the fire, and correspondingly generating the space-time distribution image of the primary cutting of the forest after the fire and the space-time distribution image of the secondary cutting of the forest after the fire.

In a possible implementation manner, the initial post-fire forest felling space-time distribution image generation module is further configured to, after obtaining initial post-fire forest regions corresponding to different years:

setting each pixel in the area of the forest which is cut down after the initial fire as a pixel of the forest which is cut down after the fire, acquiring first forest disturbance indexes corresponding to all pixels of the forest which are cut down after the fire in the current year, and acquiring second forest disturbance indexes corresponding to all pixels of the forest which are cut down after the fire in the preset year;

In a possible implementation manner, the mature forest pixel screening module is configured to extract an unfired forest region and a fired forest region from the long-time sequence annual cloud-free optimal image based on a preset combustion index threshold and a differential combustion index threshold, and specifically includes:

calculating a differential combustion index corresponding to each pixel of the forest region in adjacent years, and comparing the differential combustion index with a differential combustion index threshold value;

if the differential burning index is smaller than or equal to the differential burning index threshold value, setting the forest pixels as non-burning forest pixels, integrating all the non-burning forest pixels to obtain a non-burning forest region, and if the burning index is larger than or equal to the burning index threshold value, setting the forest pixels as burning forest pixels, integrating all burning forest pixels to obtain a burning forest region.

In a possible implementation manner, the mature forest pixel screening module is configured to perform pixel screening on the forest region without fire to obtain a mature forest pixel in the long-time sequence annual cloud-free optimal image, and specifically includes:

and acquiring mature forest pixels in all target area remote sensing images for synthesizing the long-time sequence annual cloud-free optimal images scene by scene to obtain all mature forest pixels in the long-time sequence annual cloud-free optimal images.

In a possible implementation manner, the forest disturbance index calculation module is configured to obtain and calculate a standard spike-cap conversion luminance index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of the forest on fire in different years based on a mean value and a standard deviation corresponding to the spike-cap conversion luminance index and the spike-cap conversion humidity index of the mature forest pixel year by year, and specifically includes:

In a possible implementation manner, the forest disturbance index calculation module is configured to obtain, according to the standard spike-cap transformation luminance index and the standard spike-cap transformation humidity index, a forest disturbance index corresponding to each pixel in the forest area over fire, and specifically includes:

wherein mDI is the forest disturbance index, B _n Conversion of the luminance index, W, for the standard spike-cap _n Standard spike cap shift humidity index. In a possible implementation manner, the module for setting the threshold of the index of occurrence of deforestation is configured to set the threshold of the index of occurrence of deforestation based on each sample pixel of the sample area that is not deforested after fire and the sample forest disturbance index corresponding to each sample pixel of the sample area that is deforested after fire, and specifically includes:

In a possible implementation manner, the mapping apparatus for identifying the post-fire forest felling distribution provided by the present invention further includes: a filtering module;

the filtering module is used for carrying out spatial filtering processing on the initial post-fire forest cutting space-time distribution image to obtain an initial post-fire forest cutting space-time distribution filtering image;

the filtering module is used for carrying out spatial filtering processing on the post-fire forest primary cutting space-time distribution image and the post-fire forest secondary cutting space-time distribution image to obtain the post-fire forest primary cutting space-time distribution filtering image and the post-fire forest secondary cutting space-time distribution filtering image.

In a possible implementation manner, the remote sensing image data obtaining module is configured to obtain and calculate a combustion index, a spike-cap transformation luminance index, and a spike-cap transformation humidity index based on remote sensing image data of a target area, and specifically includes:

remote sensing image data of a target area in multiple years are obtained and are based, wherein the remote sensing image data comprise remote sensing images with the cloud amount less than 50% provided by a Landsat5 satellite TM sensor and a Landsat7 satellite ETM + sensor;

carrying out mask processing on low-quality pixels in the remote sensing image data based on a QA wave band, wherein the low-quality pixels comprise cloud pixels and cloud shadow pixels;

In a possible implementation manner, the remote sensing image data acquisition module is configured to calculate a calculation formula of a combustion index, a spike-cap transformation luminance index, and a spike-cap transformation humidity index corresponding to each pixel in each remote sensing image, and specifically includes the following steps:

formula for calculating the burning index:

NBR＝(NIR–SWIR2)/(NIR+SWIR2)；

The invention also provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to realize the mapping method for the identification of the post-fire forest felling distribution.

The invention also provides a computer readable storage medium, which comprises a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located is controlled to execute the mapping method for the identification of the post-fire forest felling distribution.

Compared with the prior art, the drawing method and the drawing device for identifying the felling distribution of the forest after fire have the following beneficial effects that:

carrying out image synthesis on the obtained remote sensing image data of the target area to obtain a long-time sequence annual cloud-free optimal image; carrying out pixel screening on the long-time sequence annual cloud-free optimal image to obtain a mature forest pixel, meanwhile, carrying out standardized processing on the spike-cap transformation brightness index and the spike-cap transformation humidity index of the mature forest pixel, and calculating a forest disturbance index based on the standard spike-cap transformation brightness index and the standard spike-cap transformation humidity index, so that the difficulty that the judgment basis of forest disturbance characteristics is insufficient can be effectively solved; meanwhile, a stable and reliable threshold value of the index of the disturbance of the felled forest is established by amplifying the difference between the brightness of the spike-cap conversion and the humidity index of the spike-cap conversion between the area of the unblanked forest after fire and the area of the felled forest after fire and the area of the mature forest after fire, each pixel in the area of the forest after fire is classified year by year, the forest felling behavior after fire can be effectively identified, the area of the forest after initial fire corresponding to different years is obtained, the distribution image of the forest after initial fire when felling is generated, more accurate dynamic information of the forest after fire can be provided for a forest manager, and the problems that only single disturbance factor space-time characteristics can be extracted by the current forest disturbance map, the spectrum of multiple disturbance factors is easy to be confused, and the range of the forest after fire is difficult to extract are solved.

Drawings

FIG. 1 is a schematic flow chart diagram of an embodiment of a mapping method for post-fire forest felling distribution identification provided by the present invention;

FIG. 2 is a schematic structural diagram of an embodiment of a mapping device for recognition of post-fire forest felling distribution provided by the invention;

FIG. 3 is a schematic structural diagram of a mapping device for identifying the post-fire forest felling distribution according to another embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Example 1

Referring to fig. 1, fig. 1 is a schematic flow chart of an embodiment of a mapping method for identifying a post-fire forest felling distribution according to the present invention, as shown in fig. 1, the method includes steps 101 to 106, which are specifically as follows:

step 101: and acquiring and calculating a burning index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to the remote sensing image data based on the remote sensing image data of the target area.

In one embodiment, remote sensing image data of a target area in multiple years is obtained and is based on, wherein the remote sensing image data comprises remote sensing images with the cloud amount less than 50% provided by a Landsat5 satellite TM sensor and a Landsat7 satellite ETM + sensor; carrying out mask processing on low-quality pixels in the remote sensing image data based on a QA wave band, wherein the low-quality pixels comprise cloud pixels and cloud shadow pixels; calculating a combustion index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to each pixel in each remote sensing image based on remote sensing image multi-band data in the remote sensing image data after mask processing, wherein the remote sensing image multi-band data are Landsat 5/7 wave band from 1 st wave band to 5 th wave band and 7 th wave band.

Preferably, the remote sensing image data of the target area is acquired from a Google Earth engine platform, and the acquired remote sensing image data of the target area is a remote sensing image in a forest growing season.

As an illustration in this embodiment: the great-Xing AnLing forest area is taken as a target area, a great fire disaster occurs in 6.5.1987 in the area, and multiple times of artificial cutting occur in the local area after the fire, so that different forest communities are recovered, and therefore the great-Xing AnLing forest area is taken as the target area in the embodiment. The remote sensing image in the growing season of a target area is obtained based on a Google earth engine platform, and the remote sensing image with the cloud amount less than 50% is provided by a Landsat5 satellite TM sensor and a Landsat7 satellite ETM + sensor in 6-8 months in 1986-2011; according to poor-quality pixels such as cloud and cloud shadow in QA wave band mask remote sensing images, the images after Landsat-7ETM + airborne scanning line corrector failure (SLC-off) are preferably used, because the images can still provide effective information of forest dynamics and improve time resolution.

In one embodiment, according to 6 wavebands, such as the 1 st waveband, the 5 th waveband, the 7 th waveband and the like of Landsat 5/7 after mask processing, a burning index, a spike-cap conversion brightness index and a spike-cap conversion humidity index corresponding to each pixel in each remote sensing image are calculated, and a specific calculation formula is as follows:

formula for calculating the burning index:

NBR＝(NIR–SWIR2)/(NIR+SWIR2)；

the calculation formula of the spike-cap conversion brightness index is as follows:

in the formula, blue, green, red, NIR, SWIR1 and SWIR2 are respectively the wave band 1-Blue, the wave band 2-Green, the wave band 3-Red, the wave band 4-near infrared, the wave band 5-short wave infrared 1 and the wave band 7-short wave infrared 2 of the Landsat5TM/7ETM + image.

Step 102: setting an initial date range of image synthesis, acquiring all target area remote sensing images within the initial date range in a plurality of years, and carrying out image synthesis on all target area remote sensing images to obtain a long-time sequence annual cloud-free optimal image.

In one embodiment, the starting date range of image synthesis is determined based on the phenological information of the target area vegetation, wherein the phenological information of the target area is the date when the growth peak of the vegetation of the target area is reached.

In one embodiment, a start date range of image composition is determined according to the condition of the target area, wherein the start date range includes a target date and a time range. Preferably, the target date and time range of image synthesis is 7 months and 15 days ± 33 days in the embodiment, and it should be noted that the target date set in the image synthesis in the year of fire should be as close as possible to the fire occurrence date to avoid the influence of grass growth under the forest after fire, so as to more truly depict the forest state after fire.

Because the vegetation growth state information can be influenced by the difference of the acquisition dates of the images in different years, the target date and time range is set to be 7 months and 15 days +/-33 days so as to acquire the complete image covering the target area to the maximum extent, and simultaneously, the interference of the phenological difference in different years is reduced.

In one embodiment, all target area remote sensing images within the starting date range in the current year are obtained based on a Google Earth engine platform, the obtaining date of the target area remote sensing images and the number of days from the target date are calculated based on the Google Earth engine platform, an ee.

Step 103: extracting an unfired forest region and an unfired forest region from the long-time sequence annual cloud-free optimal image based on a preset burning index threshold and a difference burning index threshold, and performing pixel screening on the unfired forest region to obtain mature forest pixels in the long-time sequence annual cloud-free optimal image.

In one embodiment, a burning index corresponding to each pixel in the long-time sequence annual cloud-free optimal image is obtained, and the burning index is compared with a burning index threshold; if the burning index is larger than the burning index threshold value, setting the pixels as forest pixels, integrating all forest pixels to obtain a forest area, and if the burning index is smaller than or equal to the burning index threshold value, setting the pixels as non-forest pixels, and integrating all non-forest pixels to obtain a non-forest area.

Preferably, through repeated experiments, the burning index threshold is set to 0.5, based on the burning index > 0.5, to sufficiently extract the forest range.

In one embodiment, a differential combustion index corresponding to each pixel of the forest region in adjacent years is calculated, and the differential combustion index is compared with a differential combustion index threshold value; if the differential burning index is smaller than or equal to the differential burning index threshold value, setting the forest pixels as non-burning forest pixels, integrating all the non-burning forest pixels to obtain a non-burning forest region, and if the burning index is larger than or equal to the burning index threshold value, setting the forest pixels as burning forest pixels, integrating all burning forest pixels to obtain a burning forest region.

In one embodiment, the setting of the differential burn index threshold is based on an area of fire in the forest region estimated by aerial survey to determine the differential burn index threshold, and preferably, the differential burn index threshold is set to 0.23.

In one embodiment, the differential burn index is calculated as follows:

dNBR＝NBR _y-1 -NBR _y ；

wherein dNBR is a differential flammability index, NBR _y Is the burning index of the current year, NBR _y-1 Is the burn index of the previous year in the current year.

In one embodiment, an Arcgis reclassification module is used for acquiring a mask of the non-fire forest region, and a combustion index mean value and a combustion index standard deviation of each non-fire forest pixel in the non-fire forest region in a plurality of years are calculated based on ee.Reducer.mean () and ee.Reducer.stdDev () functions in a Google Earth engine platform.

In one embodiment, the coefficient of variation corresponding to each forest pixel without fire is calculated according to the mean value of the combustion index and the standard deviation of the combustion index, wherein the calculation formula of the coefficient of variation is as follows:

CV＝μ _NBR /σ _NBR ；

in the formula, mu _NBR Representing the mean value, σ, of the burning index NBR of a single non-fired forest pixel over a plurality of years _NBR Represents the standard deviation of the burning index NBR of a single non-fired forest pixel in a plurality of years.

In one embodiment, if the average value of the combustion index of the forest pixels without fire is greater than a preset threshold value of the average value of the combustion index, and the coefficient of variation is smaller than a preset threshold value of the coefficient of variation, the forest pixels without fire are set as mature forest pixels.

Preferably, the preset combustion index mean threshold is 90%, and the preset coefficient of variation threshold is 0.1. When mu is _NBR >Mature forest pixels can be fully extracted by 90 percent; the coefficient of variation is less than 0.1, which indicates weak variation, and by combining the coefficient of variation, the greater the possibility of the mature forest pixel being stable.

In one embodiment, mature forest pixels in all target area remote sensing images for synthesizing the long-time sequence annual cloud-free optimal image are selected scene by scene through a Google Earth engine platform by using a combustion index mean threshold and a variation coefficient threshold, and mature forest pixels in an overlapping area of two scenes of images are collected to obtain all mature forest pixels in the long-time sequence annual cloud-free optimal image.

Step 104: and calculating a standard spike-cap conversion brightness index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of the forest with excessive fire in different years according to the mean value and the standard difference corresponding to the spike-cap conversion brightness index and the spike-cap conversion humidity index of the mature forest pixels year by year, and obtaining a forest disturbance index corresponding to each pixel in the forest with excessive fire according to the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index.

In one embodiment, all target area remote sensing images corresponding to the current year in the synthesized long-time sequence year cloud-free optimal images are acquired year by year, and the spike-cap conversion brightness index, the mean value of the spike-cap conversion humidity index, the mean value of the spike-cap conversion brightness index and the standard difference of the spike-cap conversion humidity index corresponding to each target area remote sensing image are obtained on the basis of ee.reducer.mean () and ee.reducer.stdev () functions in a Google Earth engine platform.

In one embodiment, each fire forest pixel of a fire forest area in each target area remote sensing image is obtained, and a fire spike cap conversion brightness index and a fire spike cap conversion humidity index corresponding to each fire forest pixel are obtained; inputting the over-fire spike-cap conversion brightness index, the spike-cap conversion brightness index mean value and the spike-cap conversion brightness index standard deviation into a preset standard spike-cap conversion brightness index calculation formula to obtain a standard spike-cap conversion brightness index; and inputting the over-fire spike-cap transformation humidity index, the spike-cap transformation humidity index mean value and the spike-cap transformation humidity index standard deviation into a preset standard spike-cap transformation humidity index calculation formula to obtain a standard spike-cap transformation humidity index.

The preset standardized spike-cap conversion brightness index calculation formula is as follows:

B _n ＝(B-B _μ )/B _σ ；

the preset standardized spike-cap transformation humidity index calculation formula is as follows:

W _n ＝(W-W _μ )/W _σ ；

in the formula, B _μ And W _μ Mean value of variation brightness and humidity index of ear-cap of mature forest pixel in remote sensing image of each target area, B _σ And W _σ Standard deviation corresponding to the fringe-cap change brightness and humidity index of mature forest pixels in each target area remote sensing image, B _n As a standard spike-cap transform luminance index, W _n Is standard spike cap transformation humidity index.

In one embodiment, the spectral signals of the earth surface soil and the vegetation structure of the forest region which is cut down after fire and the forest region which is not cut down after fire are significantly different, wherein the spectral difference between the forest region which is cut down after fire and the forest region which is not cut down after fire is a numerical difference of the spike-cap transformation luminance index TCB under disturbance, the absorption of the earth surface burnt after fire and the vegetation on solar short-wave radiation is increased, the spike-cap transformation luminance index TCB value is kept at a lower level until the vegetation after fire gradually starts to recover, the vegetation coverage is reduced due to the cutting down of the forest after fire, the spectral feature of bare soil is enhanced, and the spike-cap transformation luminance index TCB value is suddenly increased; the TCW component of the spike-cap transformation humidity index comprises a SWIR wave band, can reflect the difference of forest felling and fire disaster on forest canopy moisture and structural damage, and compared with the situation that the land vegetation coverage is completely eliminated by the fire disaster, the absolute value of the TCW component of the spike-cap transformation humidity index can be larger, and meanwhile, the TCW component is insensitive to the early-stage under-forest grass irrigation growth of vegetation recovery after the fire.

Based on the spectrum difference, a forest disturbance index is provided through the standardized brightness and humidity index to respectively reflect the spectral distance between the pixels of the forest cut after fire and the forest not cut after fire and the mature forest, wherein the standard spike cap conversion brightness index is regular, the forest is cut, and the forest is not cut after fire if the standard spike cap conversion brightness index is negative.

In one embodiment, the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index are input into a forest disturbance index calculation formula to obtain a forest disturbance index corresponding to each fire forest pixel in the fire forest area, wherein the forest disturbance index calculation formula is as follows:

In one embodiment, the larger the positive value of the forest disturbance index mDI, the higher the probability that a forest pixel undergoes deforestation; the smaller negative value of the forest disturbance index mDI shows that the possibility that the forest pixels experience fire is higher; the forest disturbance index mDI value of the stable forest pixel is close to 0.

Step 105: the method comprises the steps of obtaining an un-felled sample area after fire and a felled sample area after fire, carrying out layered random sampling processing on the un-felled sample area after fire and the felled sample area after fire to obtain sample pixels of the un-felled sample area after fire and sample pixels of the felled sample area after fire, calculating and setting a threshold value of the index of the felled forest disturbance based on sample forest disturbance corresponding to the sample pixels of the un-felled sample area after fire and the sample pixels of the felled sample area after fire.

In one embodiment, the adopted land cover classifications comprise four categories of non-forest, forest without fire and forest with fire, and typical and pure forest regions without fire and forest regions with fire in the first 3 years of forest mask regions are selected according to research needs when high-resolution images and Landsat series images in Google earth are used for visual interpretation, wherein the high-resolution images of the Google earth are from Digital Global and CNES/Airbus.

In one embodiment, the collection years of the sample pixels of the sample area which is not cut down after fire and the sample pixels of the sample area which is cut down after fire are the first 3 years after the fire occurs in the target area, so that the spectral interference caused by grass irrigation and growth under the forest after fire is avoided, meanwhile, the spectral difference of fire and cut down after fire is reflected to the maximum degree, the number of the samples of the sample area which is not cut down after fire and the samples of the sample area which is cut down after fire which are acquired every year are respectively set to be 1000, and the uniformity of spatial distribution of the samples is randomly sampled in a layering manner and considered.

In one embodiment, the method comprises the steps of obtaining sample forest disturbance indexes corresponding to sample pixels in an unharvested sample area after fire and sample pixels in the unharvested sample area after fire, sequencing all the sample forest disturbance indexes in a sequence from low to high, and establishing a first numerical value histogram of the sample forest disturbance indexes.

Specifically, an Extract Multi Values to Points module in Arcgis software is used for extracting forest disturbance indexes corresponding to each sample pixel in the area where samples are not cut down after fire and each sample pixel in the area where samples are cut down after fire, ggplot2 in an R language is used for sequencing from low to high, and a first numerical histogram of the sample forest disturbance indexes of the sample pixels in the area where samples are not cut down after fire and the sample pixels in the area where samples are cut down after fire is built.

In one embodiment, the forest disturbance index values of more than 90% of the samples in the first histogram of values are selected and set as the deforestation occurrence index threshold.

Step 106: classifying each pixel in the forest region with excessive fire year by year according to the forest disturbance index and the forest disturbance index threshold value, obtaining the forest region with initial fire cutting corresponding to different years, and generating a spatial distribution image of the forest with initial fire cutting.

In one embodiment, a forest disturbance index corresponding to each forest pixel of the forest area with excessive fire is obtained, if the forest disturbance index is larger than a threshold value of the forest disturbance index with felling, the forest pixels with excessive fire are classified into forest pixels with excessive fire, all forest pixels with excessive fire are integrated, and the forest area with initial fire felling is obtained. Preferably, the deforestation index threshold is set to 5.13.

In one embodiment, the area of the forest cut after the initial fire comprises the first forest cut after the fire, the second forest cut after the fire and the non-recovered forest after the fire, wherein the second forest cut refers to the situation that the target area is cut in the first year and the burnt wood is not completely cleaned due to the intermediate cutting mode after the fire occurs, and the second forest cut is performed again in the second year; the non-recovered forest after felling means that the felled forest will maintain forest disturbance index greater than 5.13 until the vegetation is recovered to certain level.

In one embodiment, the difference forest disturbance indexes of all the sample pixels in the non-felled sample area after fire and the sample pixels in the felled sample area after fire in adjacent years are calculated based on an image difference method.

The calculation formula of the differential forest disturbance index is as follows:

ΔmDI＝mDI _t+1 －mDI _t ；

in the formula, mDI _t And mDI _t+1 And respectively representing the forest disturbance indexes of the samples of t year and t +1 year.

In one embodiment, a differential disturbance index threshold value is set according to the differential forest disturbance index.

Specifically, ggplot2 in the R language is used for sorting all the differential forest disturbance indexes of the sample pixels in the sample area which is not felled after fire and the sample pixels in the sample area which is felled after fire according to the sequence from low to high, a second numerical value histogram of the differential forest disturbance indexes of the sample pixels in the sample area which is not felled after fire and the sample pixels in the sample area which is felled after fire is established, and a differential forest disturbance index threshold value is set to be a differential forest disturbance index range corresponding to the second numerical value histogram of more than 90% of samples.

In one embodiment, based on the forest disturbance index corresponding to each pixel in the forest fire passing area, calculating a differential forest disturbance index corresponding to each pixel in the forest fire passing area in adjacent years; classifying each pixel in the forest region of the forest which is over fire year according to the forest disturbance index threshold value and the difference forest disturbance index, obtaining a primary forest cutting region of the forest which is over fire and a secondary forest cutting region of the forest which is over fire, and correspondingly generating a primary forest cutting space-time distribution image and a secondary forest cutting space-time distribution image of the forest which is over fire.

Specifically, a differential forest disturbance index corresponding to each pixel in the forest region with excessive fire is obtained, and if the differential forest disturbance index is larger than a differential forest disturbance index threshold value, an initial forest cutting region with excessive fire and a secondary forest cutting region with excessive fire are extracted from the range of the forest with initial fire cutting.

In one embodiment, due to abnormal mutation of forest disturbance index values caused by environmental noises such as cloud and cloud shadow, the influence of the environmental noises is reduced by setting rules according to the characteristic of the growth speed of vegetation in a target area; the specific rule is as follows: and if the classified pixels of the forest which is cut down after fire are restored to the state before cutting down within two years, the pixels of the forest which is cut down after fire are considered to be wrongly classified due to environmental noise and difference of physical and weather, and the pixels of the forest which is not cut down after fire are corrected, otherwise, the pixels of the forest which is cut down after fire are considered to be correctly classified. And correcting the recognized annual forest felling result, and reclassifying the post-fire felled forest pixels which are wrongly classified due to the environmental noise into uncut forest pixels by using an Arcgis heavy classification module.

Specifically, after obtaining the area of felling the forest after the initial fire, still include: acquiring first forest disturbance indexes corresponding to all pixels of the forest which is cut down after fire in the current year, and acquiring second forest disturbance indexes corresponding to all pixels of the forest which is cut down after fire in the preset year; judging whether the disturbance difference value of the first forest disturbance index and the second forest disturbance index is within a preset difference value range, if not, determining that the pixels of the forest cut down after the fire are classified wrongly, correcting the pixels of the forest cut down after the fire into pixels of the forest which are not cut down after the fire, and if so, determining that the pixels of the forest cut down after the fire are classified correctly. The first forest disturbance index is a forest disturbance index corresponding to a forest pixel cut down after a fire in the current year, and the second forest disturbance index is a forest disturbance index corresponding to a forest pixel cut down after a fire in a preset year;

in one embodiment, the number of pixels of the forest cut after fire in the forest area cut after initial fire is obtained, and if the number of the pixels of the forest cut after fire is less than the preset pixel data, the forest area cut after initial fire is reclassified. Specifically, the areas of forest cut after initial fire of less than 9 pixels are reclassified as areas of forest not cut after fire using the Sieve module in the ENVI software.

In one embodiment, because the generated result of the temporal-spatial distribution image of the felling of the fire forest may contain mixed pixels and salt and pepper noise, the spatial filtering processing is also performed on the initial temporal-spatial distribution image of the felling of the fire forest to obtain a temporal-spatial distribution filtered image of the felling of the fire forest; and performing spatial filtering processing on the space-time distribution image of the primary felling of the forest after the fire and the space-time distribution image of the secondary felling of the forest after the fire to obtain a space-time distribution filtering image of the primary felling of the forest after the fire and a space-time distribution filtering image of the secondary felling of the forest after the fire.

In conclusion, the mapping method and device for identifying the cutting distribution of the forest after fire provided by the invention provide a forest disturbance index, can effectively judge the difference of the influence of disturbance such as fire, cutting and the like on spectral characteristics, and can realize the identification of the cutting space-time characteristics of the forest after fire by establishing a mature forest pixel screening method and applying mass data of a Google earth engine big data cloud computing platform by combining an index threshold method and an image difference method, solve the difficulty of remote sensing spectrum confusion of continuous multi-disturbance events of the forest, solve the problems of difficult identification and easy neglect of cutting of the forest after fire, and have the characteristics of simplicity and easiness in use, high automation degree, high classification precision and the like.

Example 2

Referring to fig. 2, fig. 2 is a schematic structural diagram of an embodiment of a mapping apparatus for identifying post-fire forest felling distribution provided by the present invention, as shown in fig. 2, the apparatus includes a remote sensing image data acquisition module 201, a long time sequence annual cloud-free optimal image synthesis module 202, a mature forest pixel screening module 203, a forest disturbance index calculation module 204, an index threshold setting module 205 for occurrence of felling forest disturbance, and an initial post-fire forest felling space-time distribution image generation module 206, which are as follows:

the remote sensing image data acquisition module 201 is configured to acquire and calculate a combustion index, a spike-cap transformation luminance index, and a spike-cap transformation humidity index corresponding to remote sensing image data based on remote sensing image data of a target area.

The long-time sequence annual cloud-free optimal image synthesis module 202 is configured to set an initial date range of image synthesis, acquire all target area remote sensing images within the initial date range in multiple years, and perform image synthesis on all target area remote sensing images to obtain a long-time sequence annual cloud-free optimal image.

The mature forest pixel screening module 203 is configured to extract an unfired forest region and an unfired forest region from the long-time sequence annual cloud-free optimal image based on a preset burning index threshold and a difference burning index threshold, and perform pixel screening on the unfired forest region to obtain a mature forest pixel in the long-time sequence annual cloud-free optimal image.

The forest disturbance index calculation module 204 is configured to obtain the spike-cap conversion brightness index and the spike-cap conversion humidity index of the mature forest pixel year by year, and calculating a standard spike-cap conversion brightness index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area in different years, and obtaining a forest disturbance index corresponding to each pixel in the forest area according to the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index.

The threshold value setting module 205 for the index of occurrence of disturbance of felled forest is used for obtaining an area of samples which are not felled after fire and an area of samples which are felled after fire, performing layered random sampling processing on the area of samples which are not felled after fire and the area of samples which are felled after fire to obtain sample pixels of the area of samples which are not felled after fire and sample pixels of the area of samples which are felled after fire, calculating and setting the threshold value of the index of occurrence of disturbance of felled forest based on the sample pixels of the area of samples which are not felled after fire and the sample pixels of the area of samples which are felled after fire.

The initial post-fire forest cutting space-time distribution image generation module 206 is configured to classify each pixel in the over-fire forest area year by year according to the forest disturbance index and the threshold of the occurrence of cutting forest disturbance index, obtain initial post-fire cutting forest areas corresponding to different years, and generate an initial post-fire forest cutting space-time distribution image.

In an embodiment, the mapping apparatus for identifying the cutting distribution of the fire forest provided in this embodiment further includes a time-space distribution image for the first cutting of the fire forest and a time-space distribution image for the second cutting of the fire forest generating module 207, as shown in fig. 3, where fig. 3 is a schematic structural diagram of another embodiment of the mapping apparatus for identifying the cutting distribution of the fire forest provided in the present invention.

In an embodiment, the block 207 is configured to calculate, based on an image difference method, difference forest disturbance indexes of all sample pixels in a sample area that is not felled after fire and sample pixels in a sample area that is felled after fire in adjacent years, and set a difference disturbance index threshold according to the difference forest disturbance indexes; calculating a difference forest disturbance index corresponding to each pixel in the forest area with excessive fire in adjacent years based on the forest disturbance index corresponding to each pixel in the forest area with excessive fire; classifying each pixel in the forest region of the forest which has passed fire year by year according to the threshold value of the forest disturbance index which has fallen and the difference forest disturbance index to obtain a region in which the forest is firstly felled after fire and a region in which the forest is secondly felled after fire, and correspondingly generating a space-time distribution image in which the forest is firstly felled after fire and a space-time distribution image in which the forest is secondly felled after fire.

In one embodiment, the initial post-fire forest felling space-time distribution image generation module 206 is further configured to, after obtaining the initial post-fire forest regions corresponding to different years: setting each pixel in the area of the forest cut after the initial fire as a forest cut after the fire, acquiring first forest disturbance indexes corresponding to all forest cut after the fire in the current year, and acquiring second forest disturbance indexes corresponding to all forest cut after the fire in the preset year; judging whether the disturbance difference value of the first forest disturbance index and the second forest disturbance index is within a preset difference value range, if not, determining that the pixels of the forest cut down after the fire are classified wrongly, correcting the pixels of the forest cut down after the fire into pixels of the forest which are not cut down after the fire, and if so, determining that the pixels of the forest cut down after the fire are classified correctly.

In an embodiment, the mature forest pixel screening module 203 is configured to extract an unfired forest region and a fired forest region from the long-time sequence annual cloud-free optimal image based on a preset burning index threshold and a differential burning index threshold, and specifically includes: acquiring a combustion index corresponding to each pixel in the long-time sequence annual cloud-free optimal image, and comparing the combustion index with the combustion index threshold; if the burning index is larger than the burning index threshold value, setting the pixels as forest pixels, integrating all forest pixels to obtain a forest area, and if the burning index is smaller than or equal to the burning index threshold value, setting the pixels as non-forest pixels, and integrating all non-forest pixels to obtain a non-forest area; calculating a differential combustion index corresponding to each pixel of the forest region in adjacent years, and comparing the differential combustion index with a differential combustion index threshold value; if the differential burning index is smaller than or equal to the differential burning index threshold value, setting the forest pixels as non-burning forest pixels, integrating all the non-burning forest pixels to obtain a non-burning forest region, and if the burning index is larger than or equal to the burning index threshold value, setting the forest pixels as burning forest pixels, integrating all burning forest pixels to obtain a burning forest region.

In an embodiment, the mature forest pixel screening module 203 is configured to perform pixel screening on the forest region without fire to obtain a mature forest pixel in the long-time sequence annual cloud-free optimal image, and specifically includes: calculating the mean value and standard deviation of the combustion index of each forest pixel without fire in the forest area without fire in a plurality of years, and calculating the corresponding variation coefficient of each forest pixel without fire according to the mean value and standard deviation of the combustion index; if the mean value of the combustion index of the forest pixels without fire passing is larger than a preset threshold value of the mean value of the combustion index and the coefficient of variation is smaller than a preset threshold value of the coefficient of variation, setting the forest pixels without fire passing as mature forest pixels; and acquiring mature forest pixels in all target area remote sensing images for synthesizing the long-time sequence annual cloud-free optimal image scene by scene to obtain all mature forest pixels in the long-time sequence annual cloud-free optimal image.

In an embodiment, the forest disturbance index calculation module 204 is configured to obtain and calculate a standard spike-cap conversion luminance index and a standard spike-cap conversion humidity index corresponding to each pixel in the forest area of fire forest in different years based on a mean value and a standard deviation corresponding to the spike-cap conversion luminance index and the spike-cap conversion humidity index of the mature forest pixels year by year, and specifically includes: acquiring all target area remote sensing images corresponding to the current year in the synthesized long-time sequence annual cloud-free optimal images year by year, calculating the spike-cap transformation brightness index and the mean value and the standard deviation corresponding to the spike-cap transformation humidity index of all mature forest pixels in each target area remote sensing image, and obtaining the spike-cap transformation brightness index mean value, the spike-cap transformation humidity index mean value, the spike-cap transformation brightness index standard deviation and the spike-cap transformation humidity index standard deviation corresponding to each target area remote sensing image; acquiring each fire forest pixel of the fire forest region in each target region remote sensing image, and acquiring a fire spike cap transformation brightness index and a fire spike cap transformation humidity index corresponding to each fire forest pixel; inputting the over-fire spike-cap conversion brightness index, the spike-cap conversion brightness index mean value and the spike-cap conversion brightness index standard deviation into a preset standard spike-cap conversion brightness index calculation formula to obtain a standard spike-cap conversion brightness index; and inputting the over-fire spike-cap transformation humidity index, the spike-cap transformation humidity index mean value and the spike-cap transformation humidity index standard deviation into a preset standard spike-cap transformation humidity index calculation formula to obtain a standard spike-cap transformation humidity index.

In an embodiment, the occurrence deforestation disturbance index threshold setting module 205 is configured to set the occurrence deforestation disturbance index threshold based on each sample pixel in the non-deforested sample area after fire and the sample forest disturbance index corresponding to each sample pixel in the deforested sample area after fire, and specifically includes: acquiring sample forest disturbance indexes corresponding to sample pixels in a sample area which is not felled after fire and sample pixels in a sample area which is felled after fire, sequencing all the sample forest disturbance indexes from low to high, and establishing a first numerical value histogram of the sample forest disturbance indexes; selecting a forest disturbance index value of more than 90% of the sample forest disturbance indexes in the first numerical value histogram, and setting the forest disturbance index value as a deforestation index threshold value.

In an embodiment, the forest disturbance index calculation module 204 is configured to obtain, according to the standard spike-cap transformation luminance index and the standard spike-cap transformation humidity index, a forest disturbance index corresponding to each pixel in the forest area over fire, and specifically includes: inputting the standard spike-cap conversion brightness index and the standard spike-cap conversion humidity index into a forest disturbance index calculation formula to obtain a forest disturbance index corresponding to each fire forest pixel in the fire forest region, wherein the forest disturbance index calculation formula is as follows:

wherein mDI is the forest disturbance index, B _n As a standard spike-cap transform luminance index, W _n Standard spike cap shift humidity index. In an embodiment, the drawing device for identifying the deforestation distribution after fire provided by this embodiment further includes: and a filtering module.

and the filtering module is used for carrying out spatial filtering processing on the space-time distribution image obtained by primary felling of the forest after the fire and the space-time distribution image obtained by secondary felling of the forest after the fire to obtain a space-time distribution filtering image obtained by primary felling of the forest after the fire and a space-time distribution filtering image obtained by secondary felling of the forest after the fire.

In an embodiment, the remote sensing image data obtaining module 201 is configured to obtain and calculate a combustion index, a spike-cap transformation luminance index, and a spike-cap transformation humidity index based on remote sensing image data of a target area, and specifically includes: acquiring remote sensing image data of a target area in multiple years, wherein the remote sensing image data comprise remote sensing images with cloud amount less than 50% provided by a Landsat5 satellite TM sensor and a Landsat7 satellite ETM + sensor; carrying out mask processing on low-quality pixels in the remote sensing image data based on a QA wave band, wherein the low-quality pixels comprise cloud pixels and cloud shadow pixels; calculating a combustion index, a spike-cap transformation brightness index and a spike-cap transformation humidity index corresponding to each pixel in each remote sensing image based on remote sensing image multi-band data in the remote sensing image data after mask processing, wherein the remote sensing image multi-band data are Landsat 5/7 wave band from 1 st wave band to 5 th wave band and 7 th wave band.

In an embodiment, the remote sensing image data obtaining module 201 is configured to calculate a calculation formula of a combustion index, a spike-cap transformation luminance index, and a spike-cap transformation humidity index corresponding to each pixel in each remote sensing image, and specifically includes:

formula for calculating the burning index:

NBR＝(NIR–SWIR2)/(NIR+SWIR2)；

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working process of the apparatus described above may refer to the corresponding process in the foregoing method embodiment, and is not described herein again.

It should be noted that the above-mentioned embodiments of the mapping apparatus for identification of the post-fire forest felling distribution are merely schematic, wherein the modules described as separate parts may or may not be physically separate, and the parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.

On the basis of the embodiment of the mapping method for the identification of the cutting distribution of the forest after fire, another embodiment of the present invention provides a mapping terminal device for the identification of the cutting distribution of the forest after fire, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the mapping method for the identification of the cutting distribution of the forest after fire according to any embodiment of the present invention is implemented.

Illustratively, the computer program may be partitioned in this embodiment into one or more modules that are stored in the memory and executed by the processor to implement the invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which are used for describing the execution process of the computer program in the mapping terminal equipment for the identification of the post-fire forest felling distribution.

The drawing terminal equipment for recognizing the forest felling distribution after fire can be computing equipment such as a desktop computer, a notebook computer, a palm computer and a cloud server. The mapping terminal equipment for the post-fire forest felling distribution identification can comprise, but is not limited to, a processor and a memory.

The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like, said processor being the control center of the mapping terminal device for the identification of the distribution of post-fire deforestation, and various interfaces and lines are used to connect the various parts of the mapping terminal device for the identification of the distribution of post-fire deforestation.

The memory can be used for storing the computer programs and/or modules, and the processor can be used for realizing various functions of the mapping terminal equipment for the post-fire forest felling distribution identification by running or executing the computer programs and/or modules stored in the memory and calling the data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the mobile phone, and the like. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.

On the basis of the embodiment of the mapping method for the identification of the post-fire forest cutting distribution, another embodiment of the present invention provides a storage medium, where the storage medium includes a stored computer program, and when the computer program runs, a device on which the storage medium is located is controlled to execute the mapping method for the identification of the post-fire forest cutting distribution according to any one of the embodiments of the present invention.

In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, which may be in source code form, object code form, an executable file or some intermediate form, and so on. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.

In conclusion, according to the mapping method and device for identifying the post-fire forest felling distribution, the cloud-free optimal image of the long-time sequence year is obtained by image synthesis of the acquired remote sensing image data of the target area; carrying out pixel screening on the long-time sequence annual cloud-free optimal image to obtain a mature forest pixel, effectively solving the difficulty that the basis for judging the forest disturbance characteristics is insufficient, meanwhile, carrying out standardization processing on the spike-cap transformation brightness index and the spike-cap transformation humidity index of the mature forest pixel, and calculating the forest disturbance index based on the standard spike-cap transformation brightness index and the standard spike-cap transformation humidity index; meanwhile, a stable and reliable threshold value of the index of the disturbance of the felled forest is established by amplifying the difference between the brightness of the spike-cap conversion and the humidity index of the spike-cap conversion between the area of the unblanked forest after fire and the area of the felled forest after fire and the area of the mature forest after fire, each pixel in the area of the forest after fire is classified year by year, the forest felling behavior after fire can be effectively identified, the area of the forest after initial fire corresponding to different years is obtained, the distribution image of the forest after initial fire when felling is generated, more accurate dynamic information of the forest after fire can be provided for a forest manager, and the problems that only single disturbance factor space-time characteristics can be extracted by drawing of the forest disturbance at present, multiple disturbance factor spectrums are easily confused, and the range of the forest after fire is lack of extraction is difficult are solved.

The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and substitutions can be made without departing from the technical principle of the present invention, and these modifications and substitutions should also be regarded as the protection scope of the present invention.

Claims

1. A mapping method for recognizing the cutting distribution of the forest after fire is characterized by comprising the following steps:

obtaining an un-felled sample area after fire and a felled sample area after fire, carrying out layered random sampling treatment on the un-felled sample area after fire and the felled sample area after fire to obtain sample pixels of the un-felled sample area after fire and sample pixels of the felled sample area after fire, calculating and setting a threshold value of the index of the forest disturbance occurrence based on the sample forest disturbance corresponding to the sample pixels of the un-felled sample area after fire and the sample pixels of the felled sample area after fire;

2. A mapping method for recognition of post-fire forest cutting distribution according to claim 1, wherein after generating the image of the space-time distribution of the initial post-fire forest cutting, further comprising:

3. A mapping method for identification of post-fire deforestation distribution as claimed in claim 1 wherein after obtaining initial post-fire deforestation areas corresponding to different years, further comprising:

4. The mapping method for recognition of the post-fire forest felling distribution according to claim 1, wherein the extracting of the non-fire forest region and the fire forest region from the long time series year cloud-free optimal image based on a preset fire index threshold and a differential fire index threshold specifically comprises:

5. The mapping method for recognition of the deforestation distribution after fire as claimed in claim 1, wherein the step of performing pixel screening on the forest region without fire to obtain mature forest pixels in the long time series year cloud-free optimal image specifically comprises the following steps:

6. The mapping method for recognition of post-fire forest felling distribution according to claim 1, wherein the standard spike-cap conversion luminance index and the standard spike-cap conversion humidity index corresponding to each pixel in the over-fire forest area in different years are calculated by acquiring the mature forest pixels year by year and based on the mean value and the standard deviation corresponding to the spike-cap conversion luminance index and the spike-cap conversion humidity index, and specifically comprises the following steps:

acquiring all target area remote sensing images corresponding to the current year in the synthesized long-time sequence annual cloud-free optimal images year by year, calculating the spike-cap transformation brightness index and the mean value and standard deviation corresponding to the spike-cap transformation humidity index of all mature forest pixels in each scene target area remote sensing image, and obtaining the spike-cap transformation brightness index mean value, the spike-cap transformation humidity index mean value, the spike-cap transformation brightness index standard deviation and the spike-cap transformation humidity index standard deviation corresponding to each scene target area remote sensing image;

7. The mapping method for recognition of the post-fire forest felling distribution as claimed in claim 1, wherein the setting of the threshold of the index of occurrence of felled forest disturbance based on each sample pixel of the sample area not felled after fire and the sample forest disturbance index corresponding to each sample pixel of the sample area felled after fire specifically comprises:

8. The utility model provides a drawing device of distribution discernment is felled in forest after a fire which characterized in that includes: the system comprises a remote sensing image data acquisition module, a long-time sequence annual cloud-free optimal image synthesis module, a mature forest pixel screening module, a forest disturbance index calculation module, a deforestation index threshold value setting module and an initial fire-post forest deforestation space-time distribution image generation module;

the long-time sequence year cloud-free optimal image synthesis module is used for setting an initial date range of image synthesis, acquiring all target area remote sensing images within the initial date range in multiple years, and carrying out image synthesis on all target area remote sensing images to obtain a long-time sequence year cloud-free optimal image;

9. A terminal device, characterized in that it comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the charting method of the identification of the post-fire forest felling distribution according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, comprising a stored computer program, wherein the computer program, when executed, controls an apparatus on which the computer-readable storage medium is located to perform a charting method for identification of a post-fire forest felling distribution according to any one of claims 1 to 7.