Multi-laser radar three-dimensional imaging artificial intelligence ore identification method and device

Abstract

The invention discloses a multi-laser radar three-dimensional imaging artificial intelligence ore identification method and device, relates to the technical field of photoelectric detection, photoelectric imaging and photoelectric measurement, and is high in applicability, good in real-time performance and good in identification effect. The method comprises the following steps: adopting multiple laser radars with different installation angles to respectively emit laser signals to ores, and respectively obtaining 3D point cloud data information of the ores according to signals returned by the ores, wherein the number of the laser radars is at least two; fusing the multi-laser radar 3D point cloud data information according to the calibration parameters obtained by calibrating the double laser radars to obtain fused 3D point cloud data; reconstructing the fused 3D point cloud data to obtain a 3D image of the ore. extracting key features from a 3D image of ore, wherein the key features include 3D texture features, 3D shape features and echo intensity features; identifying the key features by using a pre-trained convolutional neural network model to obtain an ore identification result.

Description

technical field [0001] The invention relates to the technical fields of photoelectric detection, photoelectric imaging and photoelectric measurement, in particular to a multi-laser radar stereoscopic imaging artificial intelligence ore identification method and device. Background technique [0002] In the process of ore mining, a large amount of solid miscellaneous materials and waste materials will be produced. For example, a large amount of gangue waste will be produced in coal mining. Mixing in coal will reduce the quality of coal combustion and increase the emission of pollutants. Therefore, identification and sorting are extremely important. link. At present, the main sorting method is still manual sorting, which is inefficient for sorting. A production line with a heavy load may require 4 or more people to sort without interruption for 24 hours. The work intensity is high, and the harsh environment on site is harmful to personnel health. [0003] At present, the imagi...

AI Technical Summary

Problems solved by technology

[0005] The high-definition camera method and the infrared camera method are due to the harsh on-site environment, the dust and impurities on the surface of the object lead to the loss of texture information, and the gray scale of some ores is similar, while the X-ray method is expensive and harmful to human radiation, which makes the actual application effect of these methods bad

[0006] The resolution of a single laser radar is limited, and the density of collected information is low. At some specific angles, due to laser reflection, there may be data loss or outliers. When there is moisture on the surface of the object to be measured, the surface forms a mirror-like surface. Due to the angle, the laser beam emitted by the lidar is totally reflected, so that part of the laser beam cannot return to the lidar receiver, resulting in the loss of part of the collected information, which has a great impact on the recognition effect

[0007] Therefore, the current mainstream ore identification method has poor applicability, poor real-time performance, and poor identification effect.

Images