Research On Surface Reconstruction Of Multiple Objects With Large Depth-of-field Based On Structured Light Field

Surface reconstruction has been widely used in human-computer interaction,industrial measurement,cultural relics protection and other fields.Structured light imaging has become a key research direction and application hotspot in the field of surface reconstruction due to its high speed and high accuracy.In the large depth of field multiobjects measurement scene,because the distance between the front and back of the target may exceed the depth of field range of the traditional measurement system,traditional structured light surface reconstruction system often needs to focus manually or mechanically.The process is complex and the cost is large.Combined with the light field camera,it is expected to obtain accurate multi-objects focusing results quickly by computational focus,which is a promising scheme.When the multi-objects are measured,only one shot is needed,then the fringe image is all-in-focus by light field imaging technology,and then the surface of the all-in-focus image is reconstructed.In the previous work of this paper,the focus stack algorithm is used in all-in-focus.When the target scene is discontinuous,the clearest region of the image will be lost,which makes the synthesized all-in-focus image ring at the edge of the target contour,resulting in surface reconstruction error.In order to solve this problem,this paper uses the relationship between the multi sub-aperture image mapped to the aperture and then mapped to the depth of field,and proposes an algorithm to adjust the aperture to change the depth of field through the selective calculation of light.Secondly,the adjustable depth of field has a maximum value due to the limitation of the view angle of the light field camera.In order to get the maximum application range of this method,this thesis further explores and discusses through experiments,and obtains the scheme of automatically setting the aperture size to obtain the appropriate range of depth of field,that is,through the similarity comparison between the surface reconstruction results of different depth of field ranges and the standard surface reconstruction results,to judge whether the appropriate range of depth of field has been reached.Experimental results show that,compared with the original focus stack all-in-focus algorithm,the image quality is better in the maximum depth of field range,the contour edge has no ringing effect,and the surface reconstruction result data of the generated all-in-focus fringe image is more accurate and the contour is clearer.The maximum depth of field obtained by this method is usually about 90% of the distance from the camera to the objects,and the applicability is relatively strong.The work of this thesis provides an effective way to quickly obtain large depth of field multi-objects surface reconstruction data through the combination of structured light imaging technology and light field in the industrial field in the future.