Combination of UAV video, GIS and AI and the applications (Ⅰ)
[Introduction] The combination of drone video, GIS, and AI can help perceive dynamic targets, abnormal events, and spatial information, thus empowering videos with data mining and geographic intelligence capabilities. Based on video spatialization technology, SuperMap iDesktopX realizes the integration of video and GIS data. At the same time, using artificial intelligence technology, it realizes geographical intelligent perception through drone video and spawns more video GIS applications.
As an emerging important means of surveying and mapping, UAVs have the advantages of long endurance, low cost, maneuverability and flexibility, etc., which bring great convenience to urban planning and construction, and also provide data support for social security, UAV remote sensing, smart transportation and other industries. The combination of video, GIS, and AI can empower videos with data mining and geographic intelligence capabilities through intelligent perception of dynamic targets, abnormal events, and spatial information.
Based on video spatialization technology, SuperMap iDesktopX realizes the integration of video and GIS data. At the same time, using artificial intelligence technology, it realizes geographical intelligent perception through UAV video and spawns more video GIS applications.
Drone video access and spatialization
SuperMap supports access to videos collected by DJI’s professional and consumer drones, as well as videos collected by equipment from other UAV manufacturers. Both offline video and online video streaming are supported.
1. DJI industry-level drones: M300, M30, Phantom 4 RTK, DJI Airport, etc.;
2. DJI consumer drones: Mavic series, DJI Air 2S, etc.;
3. Equipment from other UAV manufacturers: the captured videos include drone trajectory and camera attitude files.
Online video streaming supports access to HLS (m3u8), RTSP, RTMP, HTTP, and HTTP-FLV protocol video streams.
Video spatialization refers to assigning actual geographical spatial information to each frame of the video by establishing a mapping model of pixel coordinates and geographical coordinates. Video spatialization is the most critical issue in the integration of video and GIS. SuperMap has found the "key password" for the communication between video pixels and geographical coordinates, and solved the problem in mutual conversion between video pixel coordinates and geographical coordinates. Through seven parameters: timestamp, camera longitude, latitude, camera height, pitch, yaw, roll, and combined with the camera field of view, the drone video can be spatialized.
Drone offline video can be spatialized based on camera attitude files (*.srt, .csv). Camera attitude files can be derived from video subtitle files and drone flight record files ( DJIFlightRecord * .txt). When offline video import subtitle files or camera trajectory CSV files, video spatialization can be done.
While connecting to the drone's online video stream, SuperMap iDesktopX is also connected to the camera parameter stream to perform real-time spatialization of the video stream. It can be connected to the cloud API, enable drone video streaming, and push the video stream images using the RTSP/RTMP protocol; at the same time, it can obtain the drone's position and camera attitude information through the cloud API, and forward it to iServer data streaming service. By simultaneously accessing the video stream and parameter stream, the video stream can be spatialized in real-time.
Recommended by LinkedIn
Integration of videos, maps, and scenes
The content presented in the video has a strong sense of reality. When given spatial information, video can be displayed with geographical data, which makes up for the shortcomings of traditional 2D GIS such as lack of vividness and complex virtual reality modeling, and improves the real and immersive visual experience of GIS so that GIS and video can reinforce each other.
Assigned with geographical coordinates, the video can be integrated with and displayed with map. According to different map scales, the video data can be displayed in three modes: point, range area, and video real scene. The real scene of the specified camera can be viewed on the map, making it easy to understand real-time video information and traditional maps more vivid.
The spatialized video is added to the 3D scene, and can be integrated with terrain, oblique photography model and other data and displayed together. It supports simultaneous playback of multiple videos to achieve real-time monitoring of key areas, further enhancing the realism of the immersive visual experience.
Video map is a new type of map that adopts the original perspective of the video, uses the video as the base map, assigned with geographical spatial coordinates, and then overlays other GIS data. Like ordinary maps, it can overlay GIS data such as points, lines, areas, and text. In the video map, bubble tags are used to accurately display geographical annotations, and thematic maps are used to convey geographical information more intuitively, which will improve the readability of the video and make the video display more vivid.
UAV near-ground video has the characteristics of high resolution, high timeliness, and convenient data acquisition. Based on high-definition video with spatial information and superimposed with GIS data, ground type survey and evidence, map patch change verification, etc. can be carried out, which make up for the problems like insufficient resolution of remote sensing images and untimely updates. At the same time, it can reduce delays in field surveying and mapping work caused by external factors such as weather, reduce the workload of extra field surveying and mapping, thus reducing field production costs, and improving work efficiency.
To be continued...
GIS Analyst | Colorado School of Mines | MSc
11moDefinitely, UAVs have revolutionized surveying and mapping with their efficiency. Combining them with GIS elevates their utility, allowing for real-time, detailed mapping, crucial in areas like urban planning and disaster management. Incorporating AI can take this a step further, enabling the analysis of UAV imagery for patterns and insights, particularly useful in precision agriculture and environmental monitoring. This synergy not only saves costs and time but also enhances accuracy and safety. The potential applications are vast and exciting!