A pose recognition system and method for a cantilever roadheader

Abstract

The present invention proposes a cantilever-type roadheader pose recognition system and method, which are used to solve the technical problems of large pose recognition errors and small measurable work space during long-distance excavation in the prior art. The implementation steps are as follows: Set basic parameters; establish three reference coordinate systems; adjust the initial position of the cantilever roadheader; two explosion-proof cameras collect calibration images and upload them; two explosion-proof cameras collect laser target surface images and upload them; intrinsically safe inclination sensors Collect the inclination data of its own measurement axis relative to the horizontal plane and upload it; the industrial computer calculates the inverse solution program of the cantilever roadheader; the invention uses a laser pointing device combined with a cuboid structure laser target to reduce the time required for long-distance excavation. Due to the pose recognition error caused by the two cross laser transmitters, all six sides of the cuboid laser target can be used to receive the laser projection of the laser pointer, which increases the measurable working space of the cantilever roadheader.

Description

technical field [0001] The invention belongs to the technical fields of excavation equipment application and machine vision application, and relates to a system and a method for identifying a pose of a cantilever type roadheader. Background technique [0002] The cantilever roadheader is a kind of mining equipment, which is widely used in the construction of underground roadways in coal mines. [0003] The traditional method of controlling the forward direction of the cantilever roadheader in the coal mine is to guide the operation by visually viewing the laser spot on the cutting section. This method is labor-intensive and dangerous for the operator, and the interference of human factors leads to over-excavation, under-excavation, etc. The phenomenon of digging often occurs, which reduces the production efficiency of enterprises. Therefore, the automation and intelligent transformation of cantilever roadheaders in coal mines is imperative. In order to realize the automati...

AI Technical Summary

Problems solved by technology

[0006] The disadvantage of this invention is that firstly, the laser pointing device adopts two cross beam emitters, which require that the laser beams emitted by the two cross laser emitters must be parallel to each other and in the same direction as the design axis of the roadway. The installation error of the cross laser transmitter will cause the parallelism error of the two laser beams emitted. In the case of long-distance excavation, the distance between the cross laser transmitter and the laser target is very long. The influence caused by the parallelism error of the laser beam in the pose calculation model established by the industrial control computer will be amplified, which will eventually lead to the yaw angle, pitch angle, The roll angle and the offset of a fixed point on the fuselage on the roadway section deviate greatly from the actual situation, and the system recognition accuracy is reduced

Secondly, the laser target adopts two semi-transparent PC boards with frosted surface, which are fixedly installed on the left front and right rear of the horizontal panel of the roadheader body. This kind of flat laser target can only receive laser projection on one surface. When the yaw angle, pitch angle, roll angle and the offset of a fixed point on the machine body on the roadway section are large, it is easy to cause the phenomenon of missing the target. The measurable workspace is too small, the measurable workspace is the set of all possible values ​​of the yaw angle, pitch angle, roll angle of the roadheader and the offset of a fixed point on the roadway section when the roadheader does not miss the target, so Requires frequent TBM adjustments to allow laser targets to receive laser projections again

Images