As industrialization and urbanization continue to accelerate, ground electromagnetic interference is becoming increasingly severe, which significantly affects the data quality of geomagnetic observatories. Marine magnetic surveys serve as an important supplement to oceanic observations and play a crucial role in studying marine geology and seabed resources. For instance, the discovery of striped magnetic anomalies on the seabed through marine magnetic surveys has provided important evidence for plate tectonics theory. Airborne magnetic surveys are not limited by terrain and can quickly and economically obtain high-precision regional geomagnetic data. Compared to these observational methods, low-altitude satellite magnetic measurement has the advantages of large coverage, good dynamics, high precision, and the ability to conduct global payload detection with the same satellite, effectively compensating for the limitations of ground, marine, and airborne magnetic surveys. With the development of satellite technology, it is gradually becoming an important means for global geomagnetic detection and global geomagnetic field modeling.

Monitoring the changes in the Earth's magnetic field through low Earth orbit (LEO) satellite magnetic measurements is a complex process that involves data collection, in-orbit calibration, data processing, and the construction of geomagnetic field models. Here are some key steps for monitoring changes in the Earth's magnetic field:

1. **Data Collection**: LEO satellites, such as the Zhang Heng-1 satellite, are equipped with high-precision fluxgate magnetometers that can measure the local magnetic field and its low-frequency fluctuations. These data form the foundation for monitoring the Earth's magnetic field.

2. **In-Orbit Calibration**: To ensure the accuracy of the data, it is necessary to calibrate the satellite magnetic measurements while in orbit. This includes orthogonal correction of the vector magnetic field and coordinate transformation to ensure the accuracy of the measurement data.

3. **Data Processing**: The collected magnetic measurement data needs to be processed to separate different components of the Earth's magnetic field, such as the internal source field, external source field, and long-term variation of the magnetic field. Using spherical harmonic analysis methods, a global geomagnetic field reference model, such as the CGGM 2020.0 model, can be constructed. This model integrates the geomagnetic internal source field, external source field, and long-term variation items.

4. **Construction of Geomagnetic Field Models**: Through spherical harmonic function analysis, global geomagnetic field models that describe the main magnetic field and its variations can be constructed. These models can be used in fields such as scientific research, resource exploration, communication navigation, and natural disaster prevention and control.

5. **Monitoring and Analysis**: Using in-orbit monitoring, testing, and analysis of LEO satellite signals, real-time monitoring and analysis of satellite signals can be carried out, helping satellite regulatory and operational customers to better complete in-orbit monitoring, signal verification, and analysis of LEO satellite constellations.

6. **Technical Challenges and Solutions**: The LEO electromagnetic monitoring system faces the challenge of processing massive amounts of data. The application of deep learning technology provides a new solution to this problem. Deep learning technology, combined with traditional technology in aspects such as spectral perception, blind source separation, and passive positioning, has improved the efficiency of data processing.

Through these steps, scientists can effectively use LEO satellite magnetic measurement data to monitor changes in the Earth's magnetic field and continuously optimize monitoring technology to improve accuracy and efficiency.