3D Modeling in Artificial Intelligence

Across all application areas, including geospatial applications, artificial intelligence (AI) is transforming the way in which IT solutions are created and operated from a basic standpoint. This article discusses AI-based strategies for 3D point clouds and geographic digital twins as generic components of geospatial artificial intelligence (GAI), as well as specific applications of these techniques. First, we discuss the meaning of the term "AI" and the technological advancements required to apply AI to information technology solutions, as seen from the standpoint of software engineering. Following that, we describe 3D point clouds as a key category of geodata and discuss their role in the creation of the foundation for geospatial digital twins. Finally, we discuss the feasibility of machine learning (ML) and deep learning (DL) approaches for 3D point clouds, as well as their limitations. To be more specific, we propose that 3D point clouds can be viewed as a corpus with qualities that are comparable to those of natural language corpora, and we formulate a "Naturalness Hypothesis" for 3D point clouds. This section introduces a workflow for reading 3D point clouds that relies on machine learning and deep learning approaches to generate domain- and application-specific semantics for point clouds without the need to develop explicit spatial 3D Modeling Services or explicit rule sets in the first instance. Finally, examples are shown of how machine learning and deep learning can be used to efficiently construct and manage base data for geospatial digital twins such as virtual 3D city models, indoor models, and building information models, among other things. Artificial intelligence (AI) is transforming the way in which information technology (IT) solutions are planned, built, and operated. In addition to particular application areas, artificial intelligence is finding its way into nearly all sectors and domains.. In its capacity as a collection of general-purpose technologies, artificial intelligence (AI) has enormous ramifications since it "may transform chances not only for economic growth but also for corporate profitability" (Purdy and Daugherty 2017).

Fundamental concerns in geographic domains include how artificial intelligence can be applied to spatial data and whether or not artificial intelligence must be designed particularly for spatial data. Using spatially explicit artificial intelligence, Janowicz et al. (2020) provide an overview of the field, which "utilises advancements in techniques and data cultures to support the creation of more intelligent geographic information as well as methods, systems, and services for a variety of downstream tasks." Geographical artificial intelligence (GeoAI), a newly emerging scientific discipline that "combines innovations in spatial science, artificial intelligence methods in machine learning (e.g. deep learning), data mining, and high-performance computing to extract knowledge from spatial big data" (Vopham et al. 2018), will in particular improve existing and develop new technologies for geospatial information systems (GIS).

The importance of artificial intelligence in geospatial domains has been recognised for many years, for example, in expert systems and knowledge-based systems (Openshaw and Openshaw 1997), geographically based problem solving (Smith 1984), and the analysis of social sensing data (Openshaw and Openshaw 1997). (Wang et al. 2018a). In this paper, we focus on artificial intelligence-based approaches for a specific category of 3D geodata, 3D point clouds, which are fundamental in photogrammetry, remote sensing, and computer vision (Weinmann et al. 2015) and have a wide range of applications in the construction of geospatial digital twins (Weinmann and colleagues 2015).

Several conceptual issues are associated with the concept of artificial intelligence, including the definitions of "natural," "human," and"general-purpose" intelligence. For want of a better phrase, the "most common mistake about artificial intelligence begins with the most common misconception about natural intelligence." "There is a common misperception that intelligence has only one dimension" (Kelly 2017). In the general public, the term artificial intelligence (AI) frequently conjures up connotations and expectations such as the ability to simulate or even surpass human intelligence. If artificial intelligence (AI) is viewed pragmatically as technological development, then "AI will increase human intelligence rather than replace it, in the same way that every tool amplifies our talents" (Lecun 2017). ELIZA was one of the artificial intelligence apps that served as an example of these debates. In 1966, Josef Weizenbaum created the world's first chatbot, which was a speech-based simulation of a psychologist's relationship with a patient. It "demonstrated the kind of danger potential that was encompassed inside such technological breakthroughs," according to Weizenbaum (Palatini 2014). Further discussion of this topic is provided by Copeland (1993), who examines the hurdles and obstacles that AI must overcome in order for "thinking" computers to be built. According to him, the key to artificial intelligence is the ability of computers to reason logically, to uncover meaning, to generalise and learn from prior experiences, and to be intelligent through the use of learning, thinking, problem solving, perception, and language, among other abilities.

In general, there is no clear distinction between artificial intelligence and non-AI technology. Consider, for example, an autopilot driving an aircraft: at first, it was regarded as artificial intelligence, but today it is considered a standard working technological component of most aircraft. That is to say, "when artificial intelligence (AI) becomes widely used, it is typically no longer acknowledged to be such" (Haenlein and Kaplan 2019). In this sense, the term "AI" is most commonly used to refer to technology that goes beyond the limitations of present technological capabilities. In this paper, we discuss artificial intelligence in a geospatial setting, with a particular emphasis on the possibilities of machine learning and deep learning for 3D point clouds.