A new inside-laser coaxial wire feeding processing is adopted that can ensure a good axiality between wire and laser beam through transmitting wire along the centerline of the laser beam. Basing on the optimization of process parameters and single beads cladding experiments, the causes of the uneven surface of the formed part is analyzed. By putting forward a method of the real-time control of the defocusing amount and the laser power, a fine multi-layer hollow rotary part is formed. The formed part has no subsidence damage and crack on the whole cladding layer. The size of the error is small and the hardness of the laser cladding is uniformly distributed.

Based on the technology of laser cladding manufacturing process with coaxial inside-beam powder feeding, screwy thin-walled part is accumulated on the substrate of 45# steel by controlling the angle of inclination, scanning paths, Z axis incremental and power. Through the inclined wall accumulation model and a large number of tests, limiting angle of the cladding accumulation is obtained. According to the forming process of screwy thin-walled part, two scanning paths are designed, which include continuous laser scanning between each layer and intermittent laser scanning between each layer, the better scanning path is gained by analyzing each path. A good forming part is obtained by controlling Z axis incremental and power simultaneously. The forming part is higher with smooth appearance and lower roughness, and wall thickness is uniformed.

In this paper, a method on monocular vision for spatial position is presented. The geometric model of digital camera is built and decomposed to intrinsic and extrinsic parameter matrixes under a certain assumption. Firstly, the intrinsic parameter matrix of camera is determined. Then, the extrinsic parameter matrix is solved according to the information of image. The edge detection operators are worked and in order to detect the accuracy of this method, the point of some feature points are obtained by using the principle of least squares. Compared with the conventional calibration method, this method is simple, fast and robust.