The Processing Technology Of Point Cloud Data With Section Feature For Complex Surface

With rapid development of science and technology, higher shape complexity, geometric accuracy and physical properties of the product is required, which demands new challenge to modern manufacturing technology and level, closed-loop manufacturing model of "measurement-shape redesign-precision machining" is an important method for multi-source constraints re-designing complex surface-parts with high geometric accuracy and physical performance requirements, which can get parts meeting the final performance demands. In reverse engineering, the physical parts need a procedure of "data acquisition-data processing-surface reconstruction-CAD model", and then the CAD model is used to achieve the copy, improvement or innovative design of the products. With the improving requirements of fierce market competition to product development speed and manufacturing technology flexibility, CNC profile machining technique directly based on the point cloud data continues to develop, the parts is machined through a flow of "point cloud data-data processing-machining path-CNC machining", which increases the product development speed, improve the production efficiency. It can be seen that data processing is an essential procedure in the product manufacturing, point cloud processing technology is still an important part of modern manufacturing technology, is a significant foundation for precision machining of parts or object, and acquiring qualified products meeting our needs.In this paper, basing on the previous study of precision machining and surface re-design modeling technology for multi-source constraints re-designing complex surface-parts, according to the true measured point cloud data with section feature of common base structure, the point cloud processing technology of noise pre-processing, repairing and reduction compatible with the machining process principle have been deeply researched.As to noise pre-processing of point cloud, discrete curvature estimation method based on local space parabola is propounded, which realizes the curvature estimation of cross-sectional data in any space plane; According to strong noise points and low frequency random error, chord-angle denoising method based on discrete curvature and recursive identification filtering method with local feature preservation is put forward, preserving the local morphology of the original data well. As to point cloud repairing, an adaptive repairing method based on neighborhood characteristic of point distance is studied, hole and non-uniform distribution defects is eliminated, complete point cloud with reasonable distribution is acquired. As to point cloud reduction, feature points extraction is studied, then, according to default reduction ratio or target reduction number, equal-points sampling reduction of point cloud through partitioning optimization based on feature information is achieved, point cloud data with rectangular topology is acquired, which prepares for the following surface modeling. Finally, according to true measured point cloud of common base structure, data processing is conducted using the proposed method, point cloud data meeting our needs is obtained, which lays a good foundation for the follow-up surface modeling, surface re-design, and precision machining of the common base structure skin.The proposed approach in this paper is a complement of point cloud processing technology, it can be applied to the data modeling process of complex surface-parts with apparent cross-section characteristics, which prepares for the following surface reconstruction and precision processing, also, it provides a theoretical support and reference for the development of complex surface-parts measurement-machining integration equipment.