3D Object’s Pose Estimation Based On Colored Markers Information

Pose estimation of3D object is a hot research in the field of computer vision，whichis an important part of motion estimation of3D object and has a wide research in manyfield such as face recognition, aircraft operation and control. Due to the stability, theinvariant of rotation, translation and scale and simplicity of operation of the color, it getsmore attention and wide research in image processing. Therefore, low-cost color markerbased pose estimation is universal, but there are still many problems on accuracy andcalculation of the existed researches.Based on the above analysis, the paper takes the color markers and the extraction ofthe color markers feature as the main content. The pose estimation method based on colormarkers with high efficiency and accuracy is proposed. The main work of our paper isdescribed as follows.Firstly, this paper has summarized the procedures of solving the pose estimationusing the technology of the computer vision. On basis of relevant theories knowledge andthe camera model, we design the specific plan to solve the pose estimation, which isprepared for a further research.Secondly, according to the plan of the pose estimation that has been fixed already andthe requirement of the vision based pose estimation, preprocess the images obtained bycamera. Firstly, identify the color using color components in the HSV color space;Secondly, utilize the self-defining template to remove the noise for further extraction ofthe color. Markers with different color distinguish different feature points, which makesthe identification of the feature simple, avoids the complex procedure of feature matchingand increases the speed of image preprocessing; Last, calculate the center of color areathrough the method of coordinate mean values.Thirdly, with the perspective projection in the computer vision and the camera model,the pose estimation method is introduced in detail including the algorithm theory, cameracalibration and the detailed problem-solving processing.Last, we test the validity of our method. Using the MFC to implement the graphics user interface of the whole process, which show a series of functions, at the same time theresults and analysis the experiments are given in the paper.