Runway line-oriented variable-resolution high-precision detection method and system

Abstract

The invention relates to a runway line-oriented variable-resolution high-precision detection method and system, and the method comprises the steps: 1, obtaining a to-be-detected image; 2, detecting a to-be-detected image by using the trained network model to obtain a primary detection result of runway line pixel points; step 3, according to the initial detection result, performing clustering processing on the to-be-detected image to obtain a new detection image; and 4, inputting a new detection image into the network model to obtain a secondary detection result of the runway line pixel points. According to the runway line-oriented variable-resolution high-precision detection method, the to-be-detected image is subjected to grid division, the pixel-level classification problem is converted into grid-level classification, the overall calculation amount is reduced, the detection speed is increased, the to-be-detected image is clustered according to the initial detection result to obtain the clustering center, and the detection accuracy is improved. And the pixels near the clustering center are redetected in a targeted manner, so that the precision is further improved.

Description

technical field [0001] The invention belongs to the technical field of UAV runway line detection, and in particular relates to a variable-resolution high-precision detection method and system for runway lines. Background technique [0002] At present, UAVs are widely used in various industries. Whether it is civilian or military fixed-wing UAVs, when performing tasks, before preparing to take off, they need to manually pull and correct the angle on the taxiway, so that the UAV can operate correctly. taxiing in the direction of the takeoff. This process requires considerable manpower and material resources. Therefore, a method for improving the detection accuracy of the runway line during the taxiing process of the fixed-wing UAV is proposed to solve the existing problems of considerable manpower and material resources, time cost and low detection accuracy. technical problem. [0003] The existing runway line detection technology is mainly derived from the lane line detecti...

AI Technical Summary

Problems solved by technology

[0004] The main problem with the first solution is that the segmentation needs to convert the problem into a pixel-level classification task, that is, a width×height detection task, while the resolution of the general runway line picture is relatively high, which will bring a lot of calculation consumption. , so that the real-time requirements of the runway line detection scheme cannot be realized

The main problem with the second scheme is that the detection process is too priori, and the coordinate axes need to be drawn up in advance, and the existing data set is not balanced, which will cause some deviations

Images