Research On 3D Recovery Based On Only One Online Measuring Image

The research of 3D shape recovery is a hot spot both inside and outside ourcountry. A new method of 3D shape measurement that can be used in the field ofindustrial online measurement is proposed in this paper. This method is able tomeasure the workpiece 3D shape fast and accurately with the lowest cost, so that itwill guarantee the workpiece qualification much better. The method of 3D shape recovery mentioned in this paper only need single framereal-time image to gain the object 3D dimensions in 0o～180o. The conventional SFS(Shape from Shading) algorithm is studied deeply in the researching process, whosemerits are drawn. Then the new method of 3D shape recovery based on single image,which is not only more simple and practical, but accelerates the 3D calculation greatly,is designed. This method needs only optical source and image pick-up system in termsof hardware. The uncomplicated principle of optical lens imaging is used in operatemode. The core idea of this research is that, according to the 3D cue left in 2Dimage--the shading of image, the surface vertical vector corresponding with eachpixel which is calculated in axis of optical source is converted to the axis of imagewith the conversion of coordinates to gain the object's 3D dimensions. A great numberof experimental objects are used to test the feasibility of the new algorithm in thepaper. The accuracy of 3D recovery is analyzed profoundly in the first time. Foranalyzing the accuracy of the results of 3D recovery more comprehensively, "theoverall evaluation of pixel error" and "the overall evaluation of pixel sphericity" areproposed and used for the first time. The multi-angle accuracy evaluation of artificialvase is did by "the overall evaluation of pixel error"; the sphericity evaluation of ballis did by "the overall evaluation of pixel sphericity", according to the principle ofleast square or other principles, then the results of evaluation are compared with themeasuring results which are obtained from high precision roundness measuringequipment. The factors, which may affect the accuracy of 3D recovery, such as thedepth of field, the stability of optical source and etc., are found from large quantity ofexperimental data. The measurement range and resolution are analyzed in detail, andthen the path of increasing accuracy is given in the paper. Many applications of the method of 3D recovery in different fields are introduced,among which the application in the field of defect detecting test offer a more correctand comprehensive evaluation mode for the evaluation of the quality of products thathas an important meaning for improving the quality of products. Some applications inother fields are given by the type of data in the paper, such as the military targetlocation, the analysis of bubble in chemical experiment, the recovery andrepresentation of historical relics in response to photogragh, the field of rapidprototyping and model production, the 3D transforming of living material, the literalthree-dimensional identification, and some special fields where human being can notreach. This method can also be used as the auxiliary analyzing method of other 3D IImethods, using in the fields like the shaping of tooth to help to analyze local 3Dinformation. According to the analysis of this paper, the method of 3D shape recovery basedon single image researched in the paper is a new fast effective simple research methodthat has a wide applicability. As the improvement of the stability of optical source andthe resolution of camera is going on, the accuracy of the new method will be furtherimproved, so the method will have a wide application foreground.