Research On Scanning Imaging Technology Of Terminal-sensitive Bomb-borne Lidar

The line array LIDAR cooperates with the steady-state scanning mechanism of the terminal-sensitive bomb while rotating and falling,and can scan and image ground targets.The LIDAR range image can reflect the three-dimensional characteristics of the target,and is rich in information content,which is very suitable for the automatic recognition of ground armored targets by the terminal-sensitive projectile warhead.The research results in this article provide a reference for the combined use of line array LIDAR and terminal-sensitive projectile.First,the problem of matching the motion parameters of the terminal-sensitive projectile with the LIDAR 's own parameters was studied.The field properties of the view field under various parameter combinations were obtained through simulation.The influence of factors such as the width,span,and point cloud density of the view field was analyzed and considered After that,the line number,the view field,and sampling rate of the line array LIDAR were initially determined.Secondly,after analyzing the abnormal noise generation mechanism,a linear median quadratic filter algorithm based on the scanning path line was proposed.Whether the current target point is anomalous distance,median filter the target point,and evaluate the median reliability after secondary filtering to complete the distance anomaly noise suppression,and use root mean square error to judge multiple sets of data.The results show that this noise suppression algorithm has a significant improvement over traditional methods;then,a cross-scan line method based on scan lines is proposed,and a method of combining morphological filtering methods with distance image gradients is used to extract the edges of targets through gradient thresholds to segment and extract distance images by area sorting and graphic size,It can accurately convert the target from a more complex background to the horizontal plane and segment the target.Finally,a target recognition algorithm based on similarity model matching is proposed.The rectangular feature of the target is extracted using the minimum external rectangle,and the top surface of the target is extracted by the gradient threshold algorithm.The point cloud uses histogram statistics to calculate the height characteristics of the target,extracts the target's three-dimensional features to calculate the geometric similarity,and uses the similarity threshold to estimate whether the target to be identified is a real target.The results show that the recognition rate of this algorithm for different armored targets is above 85%.