A method for detecting the inclination of intelligent towers based on machine vision
A technology of tilt detection and machine vision, which is applied in the field of image recognition, can solve problems such as high material cost, installation cost and maintenance cost, failure to find a suitable installation position, unfavorable tilt detection, etc., to achieve high safety, improve safety, and improve The effect of measurement accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0053]Firstly, the applicable platform of the method will be explained. In order to complete the necessary detection, the system needs to include a camera 2. The actual mobile platform that can be carried can be a lifting mechanism or a fixed set of equipment for each tower, but considering the cost and the need for unmanned and automated operation In this embodiment, the above-mentioned shooting device is preferably mounted under the aircraft 1; specifically, a multi-axis drone is selected as a mobile platform, and each camera 2 is connected through a pan-tilt, and the plane of the camera 2 is kept parallel to the horizontal plane.
[0054] A machine vision-based intelligent tower tilt detection method, such as figure 1 As shown, the main steps are as follows:
[0055] First of all, there is not only one way to manually control the aircraft 1 to move to the sky above the tower; it includes using GNSS positioning technology to navigate to the sky above the tower, and using vis...
Embodiment 2
[0080] This embodiment adjusts for S4 on the basis of Embodiment 1. Preferably, in S4, it is found that there will still be a small amount of raised plane in the top plane image of the tower that has the greatest influence, and the minimum value is selected as ; so in order to improve the solution The accuracy of the inclination angle is extended in this embodiment to select N depth values in the plane area in addition to selecting the depth maximum point A and the minimum value point B, and use N+2 values to fit the plane. Fitting methods include but not limited to the least squares method; obtain the equation of the fitting plane, and obtain the angle (face-to-face angle) between the plane and the horizontal plane to be the inclination of the tower.
Embodiment 3
[0082] First, machine vision is used for target recognition. After the tower is recognized, the relative position of the tower and the carrier is calculated using visual SLAM technology and path planning is carried out. Finally, the aircraft 1 is guided to move autonomously with the camera 2 to the sky above the tower, and is controlled by the flight control system. After the aircraft 1 reaches the sky above the tower, flip down the camera 2 through the gimbal to take the image of the top of the tower, and then proceed to step S1.
[0083] S1, take a top image of the tower to be measured 3 by the camera 2 positioned above the tower to be measured 3;
[0084] S2, based on the plane recognition model of machine learning training, identify the plane area at the top of the tower 3 to be measured in the image taken by S1; from the general structure of the tower, there must be at least one plane formed by points (or lines) at the top of the tower, specifically as The partition / inter...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


