Research On Automatic Measurement Technology Of Complex Parts Based On Feature Recognition

The automatic measurement of three-dimensional topography of complex parts requires the measurement system to have high measurement efficiency,automation,and the ability to complete three-dimensional topography measurement of various types of parts.At present,some measurement systems combining an industrial robot and an optical scanner can realize automatic measurement of three-dimensional topography of complex parts,but the above-mentioned traditional method can’t complete the topography measurement of parts with different shapes due to its fixed single measurement path.In order to solve this problem,a new method for automatic measurement of complex parts is proposed.It introduces visual recognition technology to identify the parts to be measured,and plans scanning paths for parts with different shapes to achieve automatic measurement of the three-dimensional shape of parts.Based on the measurement principles of industrial robots and optical 3D scanners,the schematic design of the researched automatic measurement system is developed,and the mathematical model of the measurement system is established on this basis.According to the system model established above,a template matching and feature recognition method based on edge contours is proposed.The edge contour information of the target image is extracted using Canny edge detection operator,and the edges are made into a template to match the part image to achieve the feature type;at the same time,an adaptive scanning path planning method of feature type is proposed.According to the type of geometric features of the part under test,combined with the projection raster triangle method and scanning constraints,the corresponding scanning path is adaptively planned.An automatic measurement set-up for complex parts was developed using ATOS optical scanner,Yaskawa industrial robot and measuring tool,moreover,the proposed measurement method was verified by experiments.The experimental results show that: during the measurement process,the visual recognition system can accurately identify the feature information of the part to be tested and figure out the recognition result;the scanning path planned for the multiple feature type parts can adaptively measure the 3D topography.Furthermore,the deviation of the measured data compared with the standard CAD digital model is carried out,which proves that the proposed automatic measurement method based on feature recognition can realize the topography measurement of parts with multiple feature types,and the measurement accuracy is within the allowable range.