Research On Key Techniques Of Parameter Measurement Of Square Rectangular Steel Tube Based On Line Structure Light

Cold-formed rectangular steel tubes have been widely used in the production field.With the development of technology,the demand for high-standard rectangular steel tubes in the production field is increasing.And the on-line detection of parameters method using in the production process of rectangular steel tubes has problems such as low precision and high cost of detection system.Therefore,improving the on-line measurement accuracy and measurement efficiency of the rectangular steel tube size parameters is a problem that must be solved quickly.Based on the above,this paper takes the design of low-cost,high-precision and stable online measurement system as the starting point.It focuses on the key techniques of line structure light measurement technology,which uses to measure the width,height,corner radius,diagonal and roughness parameters of rectangular steel tubes.The specific work is as follows:(1)The morphological skeleton method is selected as the stripe centerline extraction algorithm by analyzing the existing laser stripe centerline extraction algorithm and considering the accuracy and efficiency requirements of the on-line measurement system.Aiming at the defect that the morphological skeleton method will produce a large number of bone spurs,an improved spur removal algorithm is proposed.The experimental results compared with other algorithms show that the improved bone spur removal algorithm improves the accuracy and efficiency of stripe centerline extraction and meets system requirements.(2)The imaging model of the camera is established,which clarifies the conversion relationship of the four coordinate systems between the camera and the rectangular steel tubes.Then,the internal and external parameters of the camera and the laser plane calibration principle are derived based on the target.The constraint equation of the three-dimensional reconstruction of the strip center point is constructed.Finally,the three-dimensional reconstruction of the rectangular steel tube point cloud model is completed.(3)According to the characteristic that convolutional neural network can extract the local features of the images,a fast splicing algorithm for rectangular steel tube point cloud model is studied.The network model takes the automatically generated rectangular steel tube depth image as input,and uses the corresponding generated splicing parameters as the predicted output,through end-to-end iterative training until the accuracy requirement is met.The experimental results show that the network model can accurately predict the splicing parameters between two point clouds.And the prediction time is independent of the number of point clouds,which ensures the highly splicing efficiency.(4)A parameter measurement algorithm based on template and normal vector is proposed by analyzing the characteristics of the splicing rectangular steel tube point cloud model.The algorithm uses the shape template on the model as the boundary,and collects the required point data by constructing the normal vector in the template.The experimental results show that the algorithm can meet the requirements of the online measurement system.(5)A parameter measurement system based on line structure light is designed and built.Firstly,the basic framework and equipment model of the online measurement system are determined based on technical indicators and environment of the actual online measurement.Secondly,according to the production model of the rectangular steel tube,the parameters of the camera and the laser projector,the relative positions of the rectangular steel tube,the camera and the laser projector are determined.Finally,through the measurement experiment on the standard rectangular steel tube,which show that the system can meet the error requirements of the national standard and meet the accuracy and stability of the online measurement in actual production.