Shallow lake hydrological connectivity change prediction method based on thermal infrared remote sensing

Abstract

The invention relates to the unmentioned technical field of exploration or detection by using an optical device, and discloses a shallow lake hydrological connectivity change prediction method based on thermal infrared remote sensing. And the judgment is carried out based on the water temperature and the quadratic effect caused by the hydrological connectivity change on the water temperature and the water temperature change. The method comprises the following steps: extracting a shallow lake contour from a hyperspectral image to prevent a land thermal infrared remote sensing image from participating in identification; then collecting a thermal infrared remote sensing image within a period of time, monitoring an area in which the heating and cooling rate and the highest temperature are both higher than those of the periphery in the contour, and finding out nodes which are shallow in water, slow in flow velocity and easy to cause hydrological connectivity change; and then the change conditions of the temperature rise rates of the nodes are monitored, and the nodes to be disconnected are found out, so that the lake area in which the hydrological connectivity problem is about to occur is found out, and direct and exact guidance is provided for environmental workers by adopting a manual interference means.

Description

Technical Field

[0001] The present invention relates to the field of exploration or detection technology using optical devices, and in particular to a method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing. Background Art

[0002] Hydrological connectivity refers to the ease with which various substances such as sediment, nutrients and organisms migrate, transmit or diffuse between spatially heterogeneous landscapes or patches using water as a carrier. The fundamental factor is the connectivity of water as a carrier. Hydrological connectivity is an important indicator to characterize the pattern and functional stability of wetlands. The reduction of connectivity usually means the degradation of wetland ecological functions and the disturbance of internal energy flow and nutrient cycle.

[0003] Shallow lakes (less than 6 meters deep and without thermocline) are natural water bodies prone to hydrological connectivity problems due to their shallow water and strong material exchange with the bottom mud. Shallow lakes also have certain morphological particularities, and are often composed of multiple smaller lake areas connected by waterways. A mild drought is enough to cause a significant change in the water level of shallow lakes, weakening or even completely interrupting the connectivity between one part of the lake and other parts. Even the activities of wild animals (typically otters) are enough to affect the hydrological connectivity of shallow lakes. Lake areas with hydrological connectivity problems become dead water areas, with water quality rapidly deteriorating, large numbers of organisms dying, and even drying up completely before long, causing ecological problems.

[0004] However, the status of hydrological connectivity is difficult to obtain through remote sensing means. Among the factors that affect hydrological connectivity, such as water body contour, water depth, flow rate and flow velocity, only the water body contour can be directly reflected in remote sensing data and images. Other factors can only be inverted and calculated through indirect means, and the accuracy is poor, which is prone to misjudgment. Therefore, the hydrological connectivity results in the prior art are usually expressed as a relatively vague parameter such as the connectivity index, which contains probabilistic parameters. It is impossible to accurately judge the hydrological connectivity of a lake area, which is not conducive to providing work guidance for environmental workers. If it is an on-site survey, the workload is too large.

[0005] The inventors found that the absorbance of the bottom mud is significantly stronger than that of the lake water, and the specific heat capacity of the lake water is relatively high. Therefore, for shallow waterways with slow flow, the rates of heating and cooling are relatively high, and at the time of the highest water temperature every day, the water temperature is higher than that of the surrounding areas. And if the waterway gradually shrinks, the heating / cooling rate will become higher. In other words, waterways that are prone to changes in hydrological connectivity have unique characteristics in temperature, and their changes will also be reflected in temperature changes. The water temperature data can be easily obtained directly and continuously by means of thermal infrared remote sensing. Summary of the invention

[0006] The present invention provides a method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing.

[0007] The technical problem to be solved is that the hydrological connectivity status is difficult to obtain through remote sensing means, which is not conducive to providing work guidance for environmental workers.

[0008] In order to solve the above technical problems, the present invention adopts the following technical solution: a method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing, comprising the following steps: Step 1: Obtain the current hyperspectral image of the shallow lake and the thermal infrared remote sensing image of the shallow lake in the past period of time. The thermal infrared remote sensing image is collected on a sunny day, and the thermal infrared remote sensing image of each day is used as a sampling unit. The sampling units are sorted in chronological order; Step 2: Extract the outline of shallow lakes from hyperspectral images; Step 3: Extract water temperature information from the area within the shallow lake contour in the sampling unit, and find the time points with the highest and lowest temperatures of the shallow lake in each sampling unit throughout the day; Step 4: Find the shallow water slow flow nodes and the areas prone to flow interruption in the shallow lakes. The shallow water slow flow nodes are shallow lake areas that meet the following conditions: Condition 4.1: At the time of the highest temperature in a shallow lake on the same day, the temperature is higher than the surrounding area; Condition 4.2: At the time of the lowest temperature in a shallow lake on the same day, the temperature is equal to or lower than the surrounding area; Condition 4.3: On the same day, the heating rate is higher than the surrounding area, and the cooling rate is equal to or lower than the surrounding area; The easily interrupted flow area is a shallow lake area completely surrounded by shallow water slow flow nodes, or a shallow lake area completely surrounded by shallow water slow flow nodes and land; Step 5: Compare the heating rates of each shallow water slow flow node around the area prone to flow interruption at the same time in each sampling unit, and record the shallow water slow flow nodes with gradually increasing heating rates as nodes that are about to stop flowing. If all the shallow water slow flow nodes around an area prone to flow interruption are nodes that are about to stop flowing, then the area prone to flow interruption shall be recorded as an area that is about to stop flowing.

[0009] Furthermore, in step 2, the SMDPSO method is used to extract the contours of shallow lakes from the hyperspectral image.

[0010] Furthermore, in step one, at least 7 sampling units need to be obtained, and the time interval between the earliest sampling unit and the latest sampling unit shall not exceed 20 days.

[0011] Furthermore, if each sampling unit is continuous in time, then each sampling unit must contain thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is used; If the sampling units are continuous in time, then in each sampling unit, one sampling unit contains the thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is adopted; the remaining sampling units contain the thermal infrared remote sensing images of one hour before and after noon, and drone sampling is adopted.

[0012] Furthermore, in step 4, a preliminary judgment is first made based on condition 4.1, then areas that do not meet condition 4.2 are excluded, and finally areas that do not meet condition 4.3 are excluded.

[0013] Further, in step 1, a total of n sampling units are taken; In step 5, a temperature change scatter plot of each sampling unit is drawn for each shallow water slow flow node around the easy-to-break flow area. Based on the temperature change scatter plot, a polynomial function with time as the independent variable and temperature as the dependent variable is fitted. The derivative function is obtained by taking the derivative of the polynomial function. The derivative function of the nth sampling unit of the mth shallow water slow flow node is recorded as , and in the time period from time a to time b every day, ,like , then the mth shallow water slow flow node is the node that is about to stop flowing; Where T is temperature and t is time.

[0014] Furthermore, the prediction method further includes step 6: artificially intervening in the area that is about to dry up and the area that has dried up, deepening at least one node that is about to dry up around the area that is prone to dry up, and artificially digging to connect the area that has dried up with the main body of the shallow lake; The dry-up area is a shallow lake area completely surrounded by land.

[0015] Furthermore, the hyperspectral image includes an image in a thermal infrared band, and the image in the thermal infrared band of the hyperspectral image is used as a sampling unit.

[0016] Compared with the prior art, the method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing has the following beneficial effects: In the present invention, the hydrological connectivity is no longer determined by inverting water parameters that are difficult to obtain directly through remote sensing means, but the hydrological connectivity of shallow lakes is determined based on the secondary effects of changes in hydrological connectivity on water temperature and water temperature changes. First, the outline of the shallow lake is extracted from the hyperspectral image to avoid the involvement of thermal infrared remote sensing images of land in the identification; then the thermal infrared remote sensing images within a period of time are collected, and the areas within the outline where the heating and cooling rates and the highest temperatures are higher than the surrounding areas are found. Such areas correspond to areas with shallow water and slow flow rates, and are nodes that are prone to causing changes in the hydrological connectivity of shallow lakes. Then monitor the changes in the daily heating rates of these nodes over multiple days. If the heating rate is getting higher and higher in the same time period, it indicates that the water at the node is getting shallower and / or the flow is getting smaller and smaller, and it is about to be disconnected, thereby finding the lake area where hydrological connectivity problems are about to occur, and providing direct and accurate guidance to environmental workers using manual intervention methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to the present invention; DETAILED DESCRIPTION

[0018] Taking the Xiongan New Area Basin Ecological Health Protection and Quality Improvement Technology System Project: Xiongan New Area Baiyangdian and Upstream Basin Water Conservation Function Improvement Technology as an example, Figure 1 As shown, a method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing includes the following steps: Step 1: Obtain the current hyperspectral image of the shallow lake and the thermal infrared remote sensing image of the shallow lake in the past period of time. The thermal infrared remote sensing image is collected on sunny days, and the thermal infrared remote sensing image of each day is used as a sampling unit. The sampling units are sorted in chronological order; when sorting, it is generally carried out in the order of the 1st day, the 2nd day, the 3rd day..., and the two sampling units do not have to be taken on two consecutive days.

[0019] The reason why thermal infrared remote sensing images are collected from sunny days (cloud cover less than 10%) is to use the secondary effect of water temperature changes under sunlight to determine hydrological connectivity. Hydrological connectivity usually becomes problematic during droughts, in which case it is easy to obtain multiple consecutive sunny days.

[0020] Step 2: Extract the outline of shallow lakes from hyperspectral images; This is done to avoid the use of thermal infrared remote sensing images of land in the identification, because the temperature conditions of various nodes to be extracted from thermal infrared remote sensing images are often closer to those of land. When extracting the contours of water bodies from hyperspectral images, more bands such as visible light will be counted, and the spectral difference between land and water is obvious in these bands.

[0021] Step 3: Extract water temperature information from the area within the shallow lake contour in the sampling unit, and find the time points with the highest and lowest temperatures of the shallow lake in each sampling unit throughout the day; These time points are for subsequent judgment. Since the heat of the lake water mainly comes from solar energy, and the sun's heating of the lake water starts and stops at the same time on sunny days, the highest and lowest temperatures at different locations on the lake are consistent throughout the day.

[0022] Step 4: Find the shallow slow-flow nodes and areas prone to interruption of flow in shallow lakes. Shallow slow-flow nodes are shallow lake areas that meet the following conditions: Condition 4.1: At the time of the highest temperature in a shallow lake on the same day, the temperature is higher than the surrounding area; Shallow water means that it is heated by the lake bottom, and the absorbed solar energy is distributed over less water, while slow flow means that the location does not cool down as hot water flows away and cold water replenishes it; thus it can have a higher maximum temperature than the surrounding deeper water.

[0023] In addition, the temperature of a certain area refers to the representative temperature of the area. This is very intuitive in thermal infrared remote sensing images. The color of an area is basically one color, and the temperature corresponding to the color is also the representative temperature.

[0024] Condition 4.2: At the time of the lowest temperature in a shallow lake on the same day, the temperature is equal to or lower than the surrounding area; This condition is to exclude interference factors such as geothermal and artificial heat sources. Because if there is a heat source other than solar energy at a certain location, the temperature may be higher even if the water is deep at that location.

[0025] Condition 4.3: On the same day, the heating rate is higher than the surrounding area, and the cooling rate is equal to or lower than the surrounding area; Condition 4.1 is a judgment based on data collected at a certain time point, which is relatively fast, but is easily affected by accidental factors and measurement errors. Condition 4.3 is to eliminate these influences with data in two time periods. That is to say, when it is implemented, in step 4, a preliminary judgment is first made based on condition 4.1, then the areas that do not meet condition 4.2 are excluded, and finally the areas that do not meet condition 4.3 are excluded. This is both accurate and reduces the workload.

[0026] The area prone to flow interruption is a shallow lake area completely surrounded by shallow water and slow flow nodes, or a shallow lake area completely surrounded by shallow water and slow flow nodes and land; that is, the area prone to flow interruption is connected to other lake areas by shallow and slow-flowing waterways, which are easily disconnected.

[0027] Step 5: Compare the heating rates of each shallow water slow flow node around the area prone to flow interruption at the same time in each sampling unit, and record the shallow water slow flow nodes with gradually increasing heating rates as nodes that are about to stop flowing. If all the shallow water slow flow nodes around an area prone to flow interruption are nodes that are about to stop flowing, then the area prone to flow interruption shall be recorded as an area that is about to stop flowing.

[0028] The reason why we judge here by the change of the heating rate, rather than the change of the cooling rate, is that the heat source of the heating is relatively single and synchronized in different locations of the shallow lake. For cooling, there are too many interference factors.

[0029] In step 2, the SMDPSO method is used to extract the outline of shallow lakes from hyperspectral images.

[0030] The SMDPSO method, also known as the spectral matching method based on discrete particle swarm optimization, is a highly accurate water body contour extraction algorithm, which is more suitable for water bodies with irregular contours such as shallow lakes.

[0031] In step 1, at least 7 sampling units need to be obtained, and the time interval between the earliest sampling unit and the latest sampling unit should not exceed 20 days. The number of sampling units is to meet the needs of judging the change in the warming rate, and if the time interval between sampling units is too long, the influence of seasonal changes will interfere with the results.

[0032] If the sampling units are continuous in time, each sampling unit must contain thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is used; If the sampling units are continuous in time, then in each sampling unit, one sampling unit contains the thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is adopted; the remaining sampling units contain the thermal infrared remote sensing images of one hour before and after noon, and drone sampling is adopted.

[0033] Steps 1 to 4 do not require data from multiple days, only one day; and step 5 does not require data for the entire day, only a period of time is required each day. Satellite sampling is convenient for obtaining data for the entire day, but the accuracy is relatively low. Drone sampling has higher accuracy, but due to the impact of endurance, only data for a period of time can be obtained. Therefore, if the sampling unit is continuous in time, that is, if there are multiple consecutive sunny days, satellite sampling and drone sampling can be combined to obtain the advantages of both. However, if the sampling unit is not continuous in time, it means that it is necessary to be out in the field for a long time (drone sampling requires field work), and there is nothing to do when it is not sunny, which is a torture for the operator in actual operation. Of course, if the operator can tolerate it, satellite sampling and drone sampling can also be used at the same time in this case.

[0034] In step 1, a total of n sampling units are taken; In step 5, a temperature change scatter plot of each sampling unit is drawn for each shallow water slow flow node around the easy-to-break flow area. Based on the temperature change scatter plot, a polynomial function with time as the independent variable and temperature as the dependent variable is fitted. The derivative function is obtained by taking the derivative of the polynomial function. The derivative function of the nth sampling unit of the mth shallow water slow flow node is recorded as , and in the time period from time a to time b every day, ,like , then the mth shallow water slow flow node is the node that is about to stop flowing; Where T is temperature and t is time.

[0035] This set of calculation methods fully excludes the influence of water temperature, light duration, and surrounding changes on the results. Because the temperature rise rate within the same temperature rise period is selected for integration, the above factors are excluded in the definite integral process; then these definite integrals are judged by comparing the data of each day with the data of the last day. If there is an improvement in hydrological connectivity on any day during this process, it will be reflected in the results. This ensures that the node that is about to be cut off in step 5 is indeed an area where manual intervention is necessary.

[0036] The prediction method also includes step 6: artificially intervening in the area that is about to stop flowing and the area that has stopped flowing, deepening at least one node that is about to stop flowing around the area that is prone to stop flowing, and artificially excavating to connect the area that has stopped flowing with the main body of the shallow lake; The de-flowed area is a shallow lake area completely surrounded by land.

[0037] Hyperspectral images include images of thermal infrared bands, and the images of thermal infrared bands of hyperspectral images are used as a sampling unit.

[0038] In many cases, hyperspectral images include images of thermal infrared bands, and this part of the data can also be used, thereby reducing the workload of data collection.

[0039] The embodiments described above are merely descriptions of preferred implementation modes of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing, characterized in that: The following steps are involved: Step 1: Obtain the current hyperspectral image of the shallow lake and the thermal infrared remote sensing image of the shallow lake in the past period of time. The thermal infrared remote sensing image is collected on a sunny day, and the thermal infrared remote sensing image of each day is used as a sampling unit. The sampling units are sorted in chronological order; Step 2: Extract the outline of shallow lakes from hyperspectral images; Step 3: Extract water temperature information from the area within the shallow lake contour in the sampling unit, and find the time points with the highest and lowest temperatures of the shallow lake in each sampling unit throughout the day; Step 4: Find the shallow water slow flow nodes and the areas prone to flow interruption in the shallow lakes. The shallow water slow flow nodes are shallow lake areas that meet the following conditions: Condition 4.1: At the time of the highest temperature in a shallow lake on the same day, the temperature is higher than the surrounding area; Condition 4.2: At the time of the lowest temperature in a shallow lake on the same day, the temperature is equal to or lower than the surrounding area; Condition 4.3: On the same day, the heating rate is higher than the surrounding area, and the cooling rate is equal to or lower than the surrounding area; The easily interrupted flow area is a shallow lake area completely surrounded by shallow water slow flow nodes, or a shallow lake area completely surrounded by shallow water slow flow nodes and land; Step 5: Compare the heating rates of each shallow water slow flow node around the area prone to flow interruption at the same time in each sampling unit, and record the shallow water slow flow nodes with gradually increasing heating rates as nodes that are about to stop flowing. If all the shallow water slow flow nodes around an area prone to flow interruption are nodes that are about to stop flowing, then the area prone to flow interruption shall be recorded as an area that is about to stop flowing.

2. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1 is characterized in that: In step 2, the SMDPSO method is used to extract the outline of shallow lakes from hyperspectral images.

3. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1 is characterized in that: In step 1, at least 7 sampling units need to be obtained, and the time interval between the earliest sampling unit and the latest sampling unit shall not exceed 20 days.

4. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 3 is characterized in that: If the sampling units are continuous in time, each sampling unit must contain thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is used; If the sampling units are continuous in time, then in each sampling unit, one sampling unit contains the thermal infrared remote sensing images of the entire process from 0 o'clock to 24 o'clock, and satellite sampling is adopted; the remaining sampling units contain the thermal infrared remote sensing images of one hour before and after noon, and drone sampling is adopted.

5. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1 is characterized in that: In step 4, a preliminary judgment is first made based on condition 4.1, then areas that do not meet condition 4.2 are excluded, and finally areas that do not meet condition 4.3 are excluded.

6. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1, characterized in that: In step 1, a total of n sampling units are taken; In step 5, a temperature change scatter plot of each sampling unit is drawn for each shallow water slow flow node around the easy-to-break flow area. Based on the temperature change scatter plot, a polynomial function with time as the independent variable and temperature as the dependent variable is fitted. The derivative function is obtained by taking the derivative of the polynomial function. The derivative function of the nth sampling unit of the mth shallow water slow flow node is recorded as , and in the time period from time a to time b every day, ,like , then the mth shallow water slow flow node is the node that is about to stop flowing; Where T is temperature and t is time.

7. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1, characterized in that: The prediction method further includes step 6: artificially intervening in the area that is about to dry up and the area that has dried up, deepening at least one node that is about to dry up around the area that is prone to dry up, and artificially digging to connect the area that has dried up with the main body of the shallow lake; The dry-up area is a shallow lake area completely surrounded by land.

8. The method for predicting changes in hydrological connectivity of shallow lakes based on thermal infrared remote sensing according to claim 1, characterized in that: The hyperspectral image includes an image of a thermal infrared band, and the image of the thermal infrared band of the hyperspectral image is used as a sampling unit.

Images