Research On Target Recognition And Location Technology Of Wall Material Destacking Robot

With the acceleration of The integration process of urban and rural areas in China,the development momentum of China’s wall material industry is rapid,and depalletizing is an indispensable process in the production process of wall materials.At present,wall material depalletizing is still mainly manual depalletizing,manual depalletizing labor intensity,low efficiency,and high labor costs,the use of robot automatic depalletizing instead of manual depalletizing can greatly improve work efficiency,so the study of wall material automatic depalletizing robot has a very important research significance.In this thesis,the operation process of wall material depalletizing machine is studied for the identification and positioning technology of brick stacks.Based on the working environment full of smoke and dust in the field of wall material production,lidar is used as a visual sensor for target recognition to collect data from brick stacks.Then,according to the linear and corner point characteristics of brick stacking,the target recognition and positioning are studied.The main contents are as follows:1)According to the stacking posture of the brick stack,a data collection scheme is designed to place the lidar on the side of the brick stack and scan the brick stack layer by layer from top to bottom,which can quickly and accurately obtain the point cloud data of each layer of the brick stack.According to the designed data acquisition scheme,a motion control system is designed.The servo motor is controlled by the motion controller to drive the lidar to realize the up and down point and length motion on the vertical moving guide rail.2)In view of the designed data acquisition scheme,a motion control system equipped with a data acquisition system is designed,and the lidar is driven by the servo motor to drive the lidar on the vertical movement guide rail to achieve fixed length movement of the upper and lower points to achieve the layer-by-layer scanning of the brick stack by lidar.3)Considering that it will be affected by inevitable random interference in the data acquisition process,the point cloud data information collected by lidar is analyzed and pre-processed.First,the collected data is analyzed and converted into point cloud data in a Cartesian coordinate system.Then,according to the existing LOF algorithm,it is easy to produce false positives in the process of anomaly rejection,and the fusion and improved LOF algorithm is adopted to eliminate the anomalies in the data when the anomaly point misjudgment is avoided.The data is smoothed by using a median filtering algorithm.4)For the brick stack contains a large number of straight and corner point features characteristics.This paper proposes an algorithm for data segmentation and corner point feature extraction.Aiming at the excessive segmentation phenomenon of the existing Douglas-Puke algorithm in the process of data segmentation,the corner point feature determination method is introduced,which can remove the excess suspected corner points obtained by excessive segmentation,obtain the actual corner point characteristics of the brick stack,and obtain its corner point coordinates.Achieve target identification and positioning of brick stacks.5)According to the actual depalletizing process,a simulation experimental environment is set up,the brick stacks are scanned layer by layer to collect data,and the single-layer brick stacks are removed brick by brick,which verifies the feasibility of the corner point feature extraction algorithm and the accuracy of corner point positioning for different brick stack postures.Finally,the human-machine interface is designed based on Pyqt5,which enhances the intuitiveness and convenience of data acquisition,analysis and motion control functions.