Study on Remote Sensing Image Coverage Frequency Based on Spatial Grid

The coverage frequency analysis of remote sensing images is helpful to analyse the data acquisition capability of satellites or the vigorous users’ demand. In the current research, thermal values are mostly calculated using an individual point or surface target. In this paper, an index RSICF is proposed, which is calculated based on two shapes, image vectors, and spatial grid, to measure the frequency of remote sensing image coverage. The study uses GF-1 satellite image vectors to calculate RSICF around China to verify its availability. The result shows RSICF can reflect the distribution of historical image frequency respectively and can be introduced by creating a heatmap of remote sensing image coverage.


Introduction
At present, thermal values are mostly calculated using an individual point or surface target.[4] Due to the rapid development of commercial remote-sensing satellites, users' demands are spread worldwide.Most satellite operators are focused on effective data collection.[3] Most of the effective remote sensing image coverage is estimated based on the meteorological conditions for consecutive years.However, the scale of data acquisition efficiency calculated according to the meteorological conditions is relatively large, which is difficult to reflect the historical effective data in smaller regions.[6] The remote sensing image data query system of most satellite operators can retrieve the historical valid data, but it is difficult to distinguish frequencies for many valid image vector overlays in a long-time range.Currently, most studies focus on the production of heatmaps for points or shapes respectively.[2][5] [7] This paper proposes a method to calculate the remote sensing image coverage frequency (RSICF) based on a geospatial grid and historically valid image vectors.
RSICF is to calculate the coverage of historical satellite images in a certain region.It can reflect regional historical satellite data acquisition or vigorous user demand, which helps to improve the operation and management of remote sensing satellites and improve the efficiency of data acquisition in a specific region.This method is more suitable for a small range of historical valid data, and users can intuitively view the distribution of valid data.

Method
The coverage frequency analysis of remote sensing images based on spatial grid is to analyse the coverage of remote sensing images with appropriate grid size (n×n) and image vectors.The coverage of the grid is determined by calculating the overlapping area (A) between the remote sensing image vector and the grid.The image coverage frequency of a grid (ȡ) in the region can be determined by calculating the coverage of remote sensing images in a long-time series and normalizing the value.When choosing the grid scale, factors such as area size and width of remote sensing images should be taken into consideration and be consistent in the same area.
The effective coverage (m) of a grid is calculated as follows: where i is the number of long-time remote sensing images, ‫ܣ‬ is the overlapping area between the number i image vector and the certain grid, and n is the grid size.If m=1, it means effective coverage once.If m<1, it means that the grid is not completely covered, which is considered as invalid entire coverage, then making m=0.The sum of m is the effective coverage of a certain grid.The value of m is related to the time period, satellite revisit period and the size of the grid.
RSICF (ȡ) of a grid is calculated as follows: where k is the number of the grid and m is the effective coverage of remote sensing image, ݉ is the effective coverage frequency of the number k grid, and ݉ ௫ and ݉ are the maximum and minimum values of effective coverage frequency.

Experiment
The test area is located in China and its surrounding areas (6000 km × 6000 km).The grid size is 10 km and the number of grids is 600×600.The image data are the standard image product (35 km in width, half of the swath) obtained by PMS sensor of Gaofen-1 satellite in 2022, with 61242 scenes in total.
According to the strategic guidelines of the National Medium-and Long-Term Science and Technology Development Program (2006-2020), CHEOS (China's High-Resolution Earth Observation System) is approved to build a high-resolution earth observation system based on satellites, stratospheric airships and aircraft, and to improve relevant ground processing systems for application demonstration.Compared to other related observation systems, CHEOS will provide all-weather, all-time, and global earth observation capability [1].By the year 2020, a new generation of advanced earth observation systems has been established that can observe land, atmosphere, and oceans, providing information and strategic decision-making for various fields such as resource monitoring, agriculture, disaster prevention, and environmental monitoring.
The Gaofen-1 satellite (GF-1) was successfully launched at the Jiuquan Satellite Launch Center on April 26, 2013.It is the first satellite of the High-Resolution Earth Observation System (referred to as the Gaofen project).GF-1 is equipped with two sensors: PMS and WFV.The imaging characteristics of GF-1 include a high resolution of 2 m and a width of 800 km, as shown in Table 1.
According to the method, we calculate the value m in the test area, and the result is ݉ ௫ =21 and ݉ =0, which means that the most effective coverage of the area is 21, and the least is 0.
The distribution of grids and images is shown in Figure 1.RSICF in the test area is shown in Figure 2. Partial Enlarged View of RSICF is shown in Figure 3.
Considering the distribution of ground receive stations, users' demand and meteorological conditions, the coverage frequency in the northern region is greater than in the southern region.The results are basically consistent with the actual situation.

Conclusion
This paper innovatively grids a specific area, and determines the coverage frequency of remote sensing images of the study area by calculating the coverage frequency of the grid.The determination of RSICF can reflect the coverage of images in specific areas.We can adjust the grid based on the size of the research area and image swath width, to obtain a more accurate value of RSICF.It can provide a basis for effective data acquisition capabilities for the operation management of remote sensing satellites.We can use the RSICF to analyse the distribution of users' demand, effective data acquisition capabilities and historic data validity evaluation, and so on.
However, the calculation of the RSICF index requires a lot of computational power.In order to provide researchers with a more intuitive view of image coverage, we will study automated calculations for RSICF and display the real-time index on specific web pages in different styles.In the future, we should focus on grid size selection based on Multi-source image vectors with different swath widths and different sizes of research areas.

Figure 3 .
Figure 3. Partial Enlarged View of RSICF in test area.