Object localization of channel robot using laser triangulation

Trajectory tracking and Object Localization in robots are developing rapidly, but the tasks are becoming increasingly complex and significantly increasing the range of tasks for robotic systems. Cognitive tasks in domestic, industrial or traffic conditions require not only the recognition of objects but also their evaluation by classifying them without direct recognition. One of such spheres are tunnels that are physically difficult for humans to reach and require diagnostics. In such an environment, it is difficult to globally define the direction and goal, so it is necessary to interpret the locally obtained information. To solve such a problem, sensor fusion is widely applied, but sensors of different physical natures do not allow to obtain the necessary information directly, so there is a great need to use AI to interpret and control the received information and generate the robot's trajectory [1]. Local navigation systems require a wide range of sensors [4]. Various cameras and time-of-flight LiDAR lasers are widely used. For the aforementioned reasons, an economical local trajectory generation and tracking system is being developed, one of the most important components for object recognition is the laser triangulation method. The essence of this method is that the camera reacts to the projection of the laser light in front of it and interprets the obstacle depending on its distortion. In this way, the camera's resources are more concentrated, and at the same time, a simple RGB camera is enough for this method. Also, this method is perfect in the dark, when the laser light is more pronounced. In this paper, the laser triangulation method will be reviewed in detail, evaluating its advantages and disadvantages.