Image recognition method and unmanned aerial vehicle system

Abstract

The invention provides an image recognition method and an unmanned aerial vehicle system. A training image marked with a specified range is received, and a plurality of features are captured from the training image through a basic model to obtain a feature map. Next, frame selection is carried out on each point on the feature map, so that a plurality of initial detection frames are obtained, and a plurality of candidate areas are selected from the initial detection frames based on a specified range; and then, classifying the obtained candidate regions to obtain a target block, extracting feature data corresponding to the target block from the feature map, and adjusting parameters of the basic model based on the extracted feature data. According to the invention, an image with high resolution can be provided, the method has the advantage of time elasticity, and the accuracy of image identification is improved.

Description

【Technical field】 [0001] The present invention relates to an image recognition technology, and in particular, to an image recognition method using an unmanned aerial vehicle and an unmanned aerial vehicle system. 【Background technique】 [0002] With the advancement of science and technology, more and more advanced technologies are applied to the detection and prevention of natural disasters. Landslide detection has always been an important topic in disaster prevention. It can be roughly divided into pre-prevention and post-exploration. Pre-prevention is to use various sensors to predict areas that may collapse and issue warnings in advance. The detection method will vary depending on the sensor used. The sensors can be roughly divided into three types: rangefinders, wireless sensors and image finder. The rangefinder is, for example, a lidar radar (LiDAR) or infrared light or the like. The wireless sensor is, for example, a motion measurement sensor including an accelerome...

AI Technical Summary

Problems solved by technology

However, the biggest disadvantage of using satellite images at present is that the image acquisition will be limited by the operation of satellites, and the resolution of satellite images is not high enough

At present, although many satellite images claim that the sampling frequency of the same location is one day, that is, the satellite can return to the original position every other day, but to obtain the captured images still needs to go through data transmission and image processing

In addition, there is usually a time gap of several days in obtaining the current satellite images, and there is no way to assess the landslide situation in real time after the natural disaster

Moreover, the resolution of satellite images is limited at present, and there is still room for improvement in the accuracy of landslide detection.

Images