Research On Binocular Camera Self-calibration Methods

Binocular stereo vision technology simulates the principle of human using the binocular cues to perceive the surrounding scene, so it can achieve the understanding and recognition of the three-dimensional information by the sensing devices. Camera calibration technique is one of the key technologies in binocular stereo vision. It is an important part to extract three-dimensional information. The accuracy and speed of the calibration will directly affect the results of the subsequent three-dimensional measurement,This paper is under the background of three-dimensional measurement based on binocular stereo vision system, and focus on the binocular camera self-calibration technology. It primarily introduces the theoretical basis of binocular stereo vision camera self-calibration, summarizes the existing camera calibration methods, and we propose a self-calibration method for binocular cameras.The main jobs of this paper are as follows:1. The paper analysis the camera projection model, binocular stereo vision model and binocular stereo vision measuring principle. It also introduces some ordinary camera calibration methods. Based on this, we summarize the advantages of the camera self-calibration method.2. Corner detection is a key step in the camera calibration. For the X-type corner in the calibration template of this method, we use a sub-pixel corner detection method based on Harris algorithm. So the accuracy of the corner detection can be improved to sub-pixel level, and the detected corner can be closer to the actual position.3. In the binocular stereo vision system, the camera self-calibration not only gets the intrinsic and extrinsic parameters of the two cameras but also the structural parameters between the two cameras. This paper applies the camera self-calibration method based on the vanishing points to the binocular system calibration. When it is used in the actual calibration, we just need to take four groups of images with two orthogonal sets of parallel 4. lines from different angles, and don’t require the size and location information of the straight lines. The contrastive experimental results demonstrate that this method is simple and effective.