Research On 3D Visual Perception And Guidance System Of Coded Binary Surface Structure Light

Using three-dimensional vision technology,robots can accurately and rapidly detect the three-dimensional spatial information in which they are located,which is one of the key technologies of robotic intelligence.At present,there are still some problems in the research of robot three-dimensional vision,such as low accuracy and poor self-adaptability.In this thesis,a three-dimensional visual perception and guidance system coded binary structured light is studied.The system has the characteristics of non-contact,high detection accuracy and self-adaptation.The main work of this paper can be summarized as follows:(1)The structure of three-dimensional vision system based on structured light is studied.The structure of three-dimensional perception system based on line structured light and surface structured light is analyzed respectively.The calibration principle and depth measurement principle of the structured light 3D vision system are expounded,and the structured light system used in this paper is calibrated and the results are obtained.Several coding methods of structured light and some previous work of structured light coding are summarized.(2)A multi-parameter coded binary surface structured light adaptive coding system is designed and the design scheme of the system is given.According to the theory of multidimensional matrix,the multi-parameter coding theory and coding pattern of structured light are studied.On this basis,the structured light coding color algorithm which minimizes the influence of scene noise,the coding density algorithm based on target region contour and the structured light intensity adjustment algorithm based on surface reflectance estimation are studied respectively.And a structured light decoding method based on support vector machine(SVM)classification is proposed for the designed encoding pattern.(3)A binary structured light 3-D reconstruction algorithm based on multi-parameter coding is studied to optimize the 3-D depth details of the detected object.Then the binary structured light system is applied to the detection scene of micro-nano scale,and a micro-nano three-dimensional detection system based on photon counting is designed.The experimental results show that the surface structured light generated by the multi-parameter encoding method has better environmental adaptability than the traditional encoding structured light,and the binary structured light three-dimensional reconstruction algorithm can effectively optimize the three-dimensional details of the target.The multi-parameter coded binary structured light three-dimensional detection system is preliminarily tested on an industrial robot,and the adaptive structured light coded three-dimensional detection for different targets was realized.The target detection resolution is 20 microns,and the accuracy rate is 98.06%,the time of locating and grasping each target is 0.8 seconds.It provides an effective and accurate visual inspection method for industrial robots.