近年来，国内外大视场测光巡天项目不胜枚举，对天文学的发展产生了革命性影响。对大视场测光巡天而言，流量定标精度通常决定了天体亮度测量的误差下限，因此开展高精度流量定标至关重要。然而，由于以下三种原因的存在，使得地面大视场测光巡天相
对流量定标存在“1% 精度瓶颈”：平场难以精确改正、地球大气不透明度存在秒到分钟量级的快速变化以及探测器电子学不稳定性。Gaia 全天高精度测光数据、Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) 等光谱巡天海量恒星光谱数据及Gaia
DR3 BP/RP (XP) 无缝光谱数据的释放，为开展高精度流量定标提供了极佳的参考数据集，为打破地面大视场测光巡天流量定标的“1% 精度瓶颈”提供了重要契机。

本文围绕将大视场测光巡天的流量定标精度提升至毫星等级别展开研究。借助恒星颜色回归（SCR）方法和基于Gaia DR3 XP 光谱的光度合成（XPSP）方法等高精度流量定标方法，本文对定标精度最好的地面巡天数据之一Pan-STARRS (PS1) DR1 五个宽波段数据、
SAGES-NOWT 三个宽波段数据、恒星参数敏感的J-PLUS DR3 12 个中、宽波段数据开展了高精度流量（重）定标。

针对PS1 巡天测光数据，本文基于LAMOST DR7 提供的恒星大气参数数据、修正后的Gaia EDR3 高精度测光数据以及Gaia DR3 提供的XP 无缝光谱数据，借助基于光谱的SCR 方法、基于测光的SCR′ 方法和XPSP 方法，对PS1 巡天DR1 测光数据中存在的星等
相关和空间位置相关的系统误差开展了认证、溯源与修正。此过程证实了PS1 的原始定标精度优于10 毫星等，验证了三种高精度流量定标方法结果之间一致性为1–4 毫星等，修正后PS1 测光数据为开展高精度科学研究及作为参考数据定标其他测光数据奠定了坚实
基础。另外，在开展PS1 重定标过程，本文考虑了宽波段颜色红化系数的温度和消光依赖性、发展了不依赖于光谱数据的SCR′ 方法，从提高消光改正精度和提升标准星数量方面进一步发展了SCR 方法。

针对SAGES-NOWT gri 波段测光数据，本文基于LAMOST DR7 提供的恒星大气参数数据和修正后的Gaia EDR3 高精度测光数据，借助基于光谱的SCR 方法和基于测光的SCR′ 方法对其开展了相对流量定标，并结合修正后的PS1 DR1 测光数据进行了绝对流量定标。通过将修正后的PS1 星等转换成的南山星等与南山定标好的星等、南山观测图像中共同源之间的差异进行比对，我们验证了其内、外部定标精度皆为1–2 毫星等。此过程中，我们还详细讨论了恒星平场和探测器通道间相对增益随时间的变化情况以及传统的恒星
平场改正方式存在的问题，为进一步提升流量定标精度奠定基础。

针对J-PLUS 巡天DR3 测光数据，本文基于LAMOST DR7 提供的恒星大气参数数据、修正后的Gaia EDR3 高精度测光数据以及修正后的Gaia XP 光谱数据，借助基于光谱的SCR 方法和基于修正后的Gaia XP 光谱的XPSP 方法，对其展开了重定标。结果表明，上述两种方法的一致性在uJAVA、J0378、J0395、J0410 波段约为3–5 毫星等，其他波段约1–2 毫星等。在此过程中，我们比对了恒星轨迹方法（SL）和基于Gaia DR3 原始XP 光谱的XPSP 方法得到的J-PLUS 原始零点与SCR 方法零点之间的差异，并对SL 方法和XPSP方法得到的零点中存在的系统误差进行了认证与溯源。

此外，平场是大视场测光巡天数据处理中最具挑战性的步骤之一，在校正仪器效应和提升流量定标精度方面起着重要作用。本文基于mini-JPAS 54 个中带滤光片的平场数据，对CCD 相机小尺度平场展开了深入研究。此研究借助考虑电荷收集效率、像元有效面积和硅感光区厚度的物理模型，揭示了CCD 小尺度平场的物理起源：由量子效率非均匀性和像元有效面积非均匀性共同主导，CCD 像元有效面积的变化尺度大于0.4%；阐明了其波长依赖性和空间依赖性。研究表明，当需要对天体进行高精度光度、天体测量或形状测量时，CCD 间像元有效面积的相对变化及小尺度平场的空间分布不均匀性值得充分考虑和研究。另外，不同波长处小尺度平场的参数敏感性不同，为用少数波长处的小尺度平场重构其他波长处的小尺度平场提供了可能。

本文对三个代表性的巡天项目测光数据开展了高精度流量（重）定标，对CCD 探测器小尺度平场开展了深入研究。本工作已成功打破了地面大视场测光巡天流量定标的“1%精度瓶颈”，并为实现1 毫星等级别的定标精度奠定了坚实基础。论文工作可为CSST、司天工程、Mephisto 等即将到来的巡天数据的流量定标提供方法、数据支持及经验参考。

In recent years, there have been numerous wide-field imaging surveys conducted both domestically and internationally, which have had a revolutionary impact on astronomy. For wide-field imaging surveys, the accuracy of photometric calibration usually determines the lower limit of the measurement error of celestial brightness. Therefore, conducting high-precision photometric calibration is crucial. However, the relative photometric calibration of ground-based wide-field imaging surveys faces a “1% accuracy bottleneck” due to the following three reasons: difficulty in accurately correcting the flat field, rapid changes in the Earth’s atmospheric transparency at the level of seconds to minutes, and instability of detector electronics. The release of high-precision photometric data from the Gaia, massive stellar spectra data from the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST), and BP/RP (XP) spectroscopic data from Gaia DR3 provide an excellent reference dataset for conducting high-precision photometric calibration, providing an important opportunity to break through the “1% accuracy bottleneck” of ground-based wide-field imaging surveys.

This article focuses on improving the photometric calibration accuracy of wide-field imaging surveys to the millimagnitude level. Using high-precision photometric calibration methods such as the Stellar Color Regression (SCR) method and the XP Synthetic Photometry (XPSP) method based on Gaia DR3 XP spectroscopy, this study carried out high-precision photometric (re)calibration on five broad-band data of Pan-STARRS (PS1) DR1, which is one of the best calibrated ground-based survey datasets, three broad-band data of SAGES-NOWT, and 12 medium and wide-band data of J-PLUS DR3.

For PS1 survey photometric data, using atmospheric parameters of stars from LAMOST DR7, corrected Gaia EDR3 photometric data, and XP spectra data from Gaia DR3, this study employs three different methods: spectroscopy-based stellar color regression (SCR), photometry-based stellar color regression (SCR′), and XPSP methods to certify, trace, and correct the magnitude- and spatial- dependent systematic errors in the PS1 DR1 photometric data. The study confirms that the original calibration precision of PS1 is better than 10 mmag and verifies the consistency of the three high-precision photometric calibration methods to within 1–4 mmag. The corrected PS1 photometric data lays a solid foundation for high-precision scientific research and as a reference for calibrating other photometric data.

For the SAGES-NOWT gri band photometric data, the article performs relative photometric calibration using the spectral-based SCR method and the photometric-based SCR′ method, based
on the stellar atmospheric parameters provided by LAMOST DR7 and the corrected Gaia EDR3 photometric data. Absolute photometric calibration is then carried out using the corrected PS1 DR1 photometric data. The internal and external calibration accuracy is verified to be between 1–2 mmag. 

Regarding the J-PLUS DR3 photometric data, the article recalibrates the data for 12 filters (seven narrowband filters sensitive to atmospheric parameters and five SDSS-like filters) using the spectral-based SCR method and the XPSP method based on the corrected Gaia XP spectrum, along with the stellar atmospheric parameters provided by LAMOST DR7 and the corrected Gaia EDR3 photometric data. The consistency between the two methods is about 3-5 millimagnitudes in the uJAVA, J0378, J0395, and J0410 bands, and about 1-2 millimagnitudes in other bands. The article also compares the differences between the J-PLUS original zero points obtained by the stellar trajectory method (SL) and the XPSP method based on Gaia DR3 XP spectra and the zero points obtained by the SCR method, and certifies and traces the systematic errors in the zero points obtained by the SL and XPSP methods. Additionally, the article discusses the changes in relative gain between stellar flat fields and detector channels over time and the problems with traditional stellar flat field correction methods, laying the foundation for further improving photometric calibration accuracy.

In addition, flat-fielding is one of the most challenging steps in processing large-field photometric survey data, playing an important role in correcting instrument effects and improving photometric calibration accuracy. In this paper, based on the mini-JPAS data with 54 intermediate-band filters, we conducted an in-depth study on small-scale flat fields of CCD cameras. By using a physical model that takes into account the charge collection efficiency, the effective pixel area, and the thickness of the silicon photosensitive region, we revealed the physical origin of the small scale flat field of CCDs: it is jointly dominated by the non-uniformity of quantum efficiency and the non-uniformity of effective pixel area, and the change scale of the effective pixel area of CCD pixels is greater than 0.4%. We also clarified its wavelength and spatial dependencies. The study shows that when high-precision photometry, astrometry, or shape measurement of celestial objects is required, the relative variation of the effective pixel area of CCD pixels and the spatial distribution non-uniformity of small-scale flat fields should be fully considered and studied. In addition, the parameter sensitivities of small-scale flat fields at different wavelengths are different, which provides the possibility of reconstructing small-scale flat fields at other wavelengths using small-scale flat fields at a few wavelengths.

Wide-field photometric survey projects are flourishing both domestically and internationally, and high-precision photometric calibration is crucial to fully exploit the potential of photometric survey data. This paper conducted high-precision flux (re)calibration on three representative survey projects’ photometric data and conducted an in-depth study on small-scale flat fields of CCD detectors. This work has successfully broken through the “1% accuracy bottleneck” of ground based large-field photometric survey photometric calibration and laid a solid foundation for achieving calibration accuracy at the level of 1 mmag. The results of this paper can provide method, data support, and experience reference for the photometric calibration of upcoming survey data such as CSST, Sitian project, Mephisto, etc.