Bolt defect space positioning method based on multi-view geometry

Abstract

The invention provides a bolt defect space positioning method based on multi-view geometry. An original image is shot by an unmanned aerial vehicle; a multi-view geometry theory and a deep learning algorithm are used; image recognition techniques are used to carry out target object extraction and Sift key point extraction; multi-view geometric operation is performed on the key points; pairing points are generated, cross query is performed, homonymous point sets are generated, internal and external orientation elements and lens distortion parameters of an aerial triangulation result of an imagewhere each point is located are calculated; based on a space forward intersection principle, object space coordinates of homonymous points are calculated; the bolt defects are positioned and displayed on the three-dimensional model of the iron tower through clustering and de-weighting, so that acceptance personnel can clearly see the number and positions of the defects, the situation of manual missing inspection is avoided, the accurate inspection of the bolt defects of the iron tower is ensured, the bolt installation acceptance quality is improved, and the continuous power supply operation of a power line is ensured.

Description

technical field [0001] The invention belongs to the technical field of power line inspection, and in particular relates to a method for spatially locating bolt defects based on multi-view geometry. Background technique [0002] At present, in the process of inspection and acceptance of power lines, the installation of iron tower bolts requires that the bolts penetrate uniformly and beautifully, and the bolts and nuts must be tightened. The judgment of this item is mainly by observing whether the nuts fit or checking with a torque wrench. The requirements are consistent with the drawings, and less missing bolts are not allowed. The bolts and screws are regular cylinders, and the nuts are regular hexagons. Uninstalled bolts can be directly identified by observation, but because there are many iron towers, it is difficult to check and accept, so Most of them use drones for aerial photography, but most of them use images to observe, and human judgment and analysis are used to de...

AI Technical Summary

Problems solved by technology

[0002] At present, in the process of inspection and acceptance of power lines, the installation of iron tower bolts requires that the bolts penetrate uniformly and beautifully, and the bolts and nuts must be tightened. The judgment of this item is mainly by observing whether the nuts fit or checking with a torque wrench. The requirements are consistent with the drawings, and missing bolts are not allowed. The bolts and screws are regular cylinders, and the nuts are regular hexagons. Uninstalled bolts can be directly identified by observation, but due to the large number of iron towers, it is difficult to check and accept, so Most of them use drones for aerial photography, but most of them use images to observe, and human judgment and analysis are used to determine whether there are defects. Due to the distortion of the photos taken, the irregularity of the overlap of the photos, etc., it is impossible to accurately judge the existence of defects. Operation poses a great risk

Images