Influencing Factors And Experimental Study On Three-Dimensional Model Reconstructed By Section Sequence Images

With the rapid development of manufacturing industry,people have higher requirements on the performance and the diversification of product.Meanwhile the shorter production cycle,as well as the faster research and development are expected.In industry,it has important application to reconstruct a CAD model of the part.In the reverse design,researchers can identify and measure the contour of parts through images and then reconstruct 3D models.In terms of layer-cutting reverse technology,there are still some problems in the actual measurement and reconstruction process such as poor image quality caused by the bad shooting environment,messy points cloud and large fitting error,which lead to low reconstruction accuracy.To solve the above problems,the layer-cutting inverse method was adopted in this paper.The acquisition method of high-contrast image was studied.On this basis,a region-based image enhancement algorithm was proposed to remove the light spots and stray points in the image.Then the single and clear edges was obtained.In the process of surface fitting,the complex surface was divided into simple surface to fit separately,so as to improve the fitting precision.Specific research contents are as follows.Firstly,the image-acquisition system was studied theoretically and experimentally.The camera model,saturation,focal length and aperture were adjusted according to the shooting environment regulating.Next,the imaging quality under different light intensity was compared by the shooting experiment.The image gray histogram and gray scatterplot of the part image show that it can obtain high contrast image with adjusting saturation properly under the condition of weak light.Workpiece surface texture is an important factor for the imaging accuracy.Therefore,the four kinds of section texture processed by different manufacturing method-line cutting,turning,milling and grinding,were photographed under the same condition.By calculating the contour accuracy of the part in the image with Matlab,it was found that the physical shape of part affects the figure quality a lot.And it provided theoretical basis and experimental guidance for building experimental platform.Based on the existing experimental conditions,the milling method was selected to obtain the section of the parts,and the shooting experiment table was built.Moreover,the camera was controlled to capture the section image of 100 layers continuously.On this basis,an image enhancement algorithm based on region was put forward to remove the noise and light spots.The algorithm can realize the subarea of image and reallocate grayscale values in the different area to remove noise point and light spot effectively,meanwhile with not changing the edge pixel positioning.After that,sobel edge detection operator was used to calculate the clear and single edge of the part in picture.In the process of surface fitting and 3d reconstruction,100-layer contour point clouds data were simplified based on the distance between the points with the software-Imageware first.And then the point clouds were divided into eight simple-shape parts based on the geometric features.The algorithm with small error was selected to fit the different parts of the surface respectively,and the 3D model of the part was finally restored by cutting and splicing each surface.After the above steps,the 3D model of the given part is obtained and the reconstruction error is only 0.69%.A good reconstruction result was obtained,which provides a reference method for reverse engineering design.