withConstruction of Digital ChinaWith the continuous advancement of the digital transformation and upgrading of the whole society, the demand for more refined and high-precision basic geographic information is accelerating, and the demand for high-precision, high-resolution, and multi-modal information is acceleratingDigital elevation modelwith3D geospatial solid modelThe need for data and its intelligent analysis applications is becoming more and more urgent, from decentralized 4D products to integrated onesReality 3D modelImperative.
At the same time,Reality 3D conceptThe connotation is also evolving, according to the "Explanation of Terms for New Basic Surveying and Mapping and Real 3D China Construction Technical Documents" issued by the competent department of natural resources in 2022Live 3DAs a proper noun term, itsDefinitionsFor "real 3D is a real, three-dimensional, temporal reflection and expression of human production, life and ecological space within a certain range of digital space, is a new type of basic surveying and mapping standardized products, is an important new type of national infrastructure, for economic and social development and information technology of various departments to provide a unified space base". In August 2023, the state issued the "Overall Design Plan for the Construction of Real 3D China", which was clarifiedReal-life 3D ChinaIt is a new type of spatio-temporal information infrastructure that substantive, three-dimensional, and automatic dynamic description and expression of China's land space, and is a new generation of national basic geographic information system.
Current Mainstream Modeling Methods:
(1) Oblique photography
Oblique photography is the main means of real-life 3D modeling, and it is also the most cost-effective modeling method for large-scale scene modeling. This method is economical and efficient, but the deformation of the model is large, especially for low-rise buildings, and various features in the model are connected, and there is a lack of semantic information.
(2) Laser scanning
Compared with oblique photography, lidar has better stability and point cloud quality, but the cost is also higher than oblique photography. Lidar visualizes the landscape by revealing subtle elevation changes on the ground. The advantage lies in high-speed sampling, with lidar emitting hundreds of thousands, if not millions, of pulses per second from the air to the ground, and it is this dense point cloud that allows us to obtain true geomorphology.
(3) Manual modeling
At present, manual modeling is still a very important modeling method. Take a look at the various digital twin projects, where the models are largely modeled by hand. Manual models are controlled by humans, which can control the structure and materials of the model more exquisitely, and with the powerful rendering capabilities of the game engine, they can often show very good visual effects. Of course, the biggest problem with manual modeling is cost and speed, which is why manual modeling is often used in small scenarios.
FastModeling
Due to the constraints of high cost, long production cycle and large amount of data for city-level real-world 3D fine modeling, it is urgent to use new technologies and data sources to carry out rapid construction of urban 3D models. It is mainly for urban buildings and some structures to carry out single 3D model data production lod13. Monomer model. Automatic or semi-automatic modeling can be achieved through existing digital line drawing, laser point clouds, oblique mesh models, etc., and the approximate material library texture can be given to produce a monolithic model result.
Currently, flyYan Aviation Remote Sensing Technology***(hereinafter referred to as "Feiyan Remote Sensing") uses the above four modeling methods, processes the oblique images and point cloud data obtained through proprietary software, constructs shape information through point cloud data, and provides texture images for oblique images, so as to quickly and automatically construct 3D models. The difference is that in most cases, airborne laser and oblique images are acquired separately, and the two sets of data obtained need to be manually registered, and the cost of both acquisition and post-processing is highFlying Swallow remote sensingAdopt self-developed onesAIMS multimodal aerial camera, a single flight can simultaneously acquire high-resolution oblique imagery, point cloud data, hyperspectral and other data such as hyperspectral. A set of multi-mode heterogeneous, time-unified.
1. Aerial remote sensing data with basically the same imaging conditions can realize the fine perception of all elements of ground objects, which can not only save the cost of data acquisition, but also use multi-modal data to collaboratively improve the automation and intelligence level of processing, interpretation and application, and form a series of data products with rich characteristics and strong relevance, which can be efficiently assistedReal-life 3D construction。At the same time, users are allowed to customize specific services and data products according to their own needs, as well as data sharing and development applications, supporting urban planning, architectural design, smart transportation, environmental protection, tourism promotion and other fields.
theseReal-life 3D intelligent platformIt is conducive to promoting social development, improving the level of decision-making, optimizing resource allocation, promoting the construction of smart cities and promoting scientific research and innovationLive 3DdevelopmentIt is important to achieve sustainable development and build a better society.