3D Gaussian Splatting (3D GS) is a transformative technique in explicit radiance field and computer graphics, characterized by the use of millions of learnable 3D Gaussians. Unlike mainstream neural radiance field approaches, which primarily use implicit, coordinate-based models, 3D GS offers real-time rendering capabilities and unprecedented levels of editability. This paper provides the first systematic overview of recent developments and critical contributions in 3D GS, exploring its underlying principles, practical applicability, and performance across various benchmark tasks. The survey highlights the potential of 3D GS in applications such as virtual reality, interactive media, and real-time cinematic rendering, while also identifying current challenges and suggesting future research directions. The introduction of 3D GS represents a fundamental shift in scene representation and rendering, enabling more efficient computation and rendering while maintaining high-quality image synthesis. The survey covers the background of radiance fields, scene reconstruction, and rendering, and delves into the principles of 3D GS, including its rendering process with learned 3D Gaussians and optimization details. It also discusses extended versions of 3D GS, such as data-efficient, memory-efficient, and photorealistic 3D GS, as well as improved optimization algorithms and applications in various domains like robotics, dynamic scene reconstruction, AI-generated content, autonomous driving, and medical systems.3D Gaussian Splatting (3D GS) is a transformative technique in explicit radiance field and computer graphics, characterized by the use of millions of learnable 3D Gaussians. Unlike mainstream neural radiance field approaches, which primarily use implicit, coordinate-based models, 3D GS offers real-time rendering capabilities and unprecedented levels of editability. This paper provides the first systematic overview of recent developments and critical contributions in 3D GS, exploring its underlying principles, practical applicability, and performance across various benchmark tasks. The survey highlights the potential of 3D GS in applications such as virtual reality, interactive media, and real-time cinematic rendering, while also identifying current challenges and suggesting future research directions. The introduction of 3D GS represents a fundamental shift in scene representation and rendering, enabling more efficient computation and rendering while maintaining high-quality image synthesis. The survey covers the background of radiance fields, scene reconstruction, and rendering, and delves into the principles of 3D GS, including its rendering process with learned 3D Gaussians and optimization details. It also discusses extended versions of 3D GS, such as data-efficient, memory-efficient, and photorealistic 3D GS, as well as improved optimization algorithms and applications in various domains like robotics, dynamic scene reconstruction, AI-generated content, autonomous driving, and medical systems.