Method and system for real-time generation of orthophotos based on SLAM technology

Abstract

The present invention discloses a real-time orthophoto generation method and system based on SLAM technology, which can perform the following processing on a single frame image in a single frame image sequence obtained by decomposing video stream data collected by a drone to obtain a digital orthophoto image in real time. The steps Including: (1) removing camera distortion of a single frame image and extracting feature points; (2) obtaining and optimizing the high-precision absolute position and posture of a single frame image; (3) optimizing the pose and pixel coordinates of the four corner points of a single frame image Convert to projected image coordinates, obtain the homography transformation relationship between the single frame image and the projected image, and generate an orthophoto. This invention does not rely on a high-precision positioning and attitude determination system. It can use the acquired UAV image data and GPS data to perform high-precision positioning and attitude determination in real time, and dynamically obtain orthophotos of the measurement area; realizing UAV data acquisition and The generation of high-precision orthophotos is performed simultaneously, which can meet the needs for high-precision real-time processing of data in emergency situations such as earthquake relief and emergency response.

Description

technical field [0001] The invention relates to the field of aerial photogrammetry, in particular to a method and system for real-time generation of orthophotos based on SLAM technology (simultaneous vision-based positioning and mapping technology). Background technique [0002] In recent years, drones have been widely used in surveying and mapping, exploration, emergency response, disaster relief and other fields due to their high efficiency, flexibility, and low cost. UAV aerial photogrammetry technology greatly reduces the cost of traditional aerial photogrammetry technology, making aerial photogrammetry technology from aristocratic to civilian, and has more general practical significance and application value. [0003] But on the other hand, the traditional implementation method of UAV image processing is to transmit the UAV aerial image to a computer or server through the USB interface for post-processing after the aerial photography mission is completed. Generally, it ...

AI Technical Summary

Problems solved by technology

This method relies on a high-precision POS system (Positioning and Attitude Determination System), and does not clearly provide a calculation method for the local average elevation, and this method only uses the rotation matrix composed of the angular elements of the camera's external orientation elements to transform the oblique image into a horizontal image. Correct the generated orthophoto accuracy is not high enough

In addition, this method calculates the size of the mosaic image according to the pre-divided aerial photography area information and the scale of the panoramic image. It is difficult to cope with unknown or inaccurate route planning information; generally, the image size of the panoramic image can only be estimated roughly or saturated, but the image size of the panoramic image cannot be accurately calculated; when generating a panoramic image The processing of the graph takes up a lot of memory resources

However, the problem with this method is that it relies on a high-precision POS system, or requires known high-precision POS data; it is suitable for the generation of large-scale orthophotos with low precision (meter level), such as using large-scale digital aerial photography Or satellite images to generate meter-level precision DOM

Images