Zhejiang University

Canjun Yang

Ri Lin

Mingwei Lin

Xin Wu

Bohao He

Zilin Xing

Yuanchao Zhu

Zhangpeng Tu

Shuguang Chu

Anzhe Yi

Dejun Li

Weitao Wu

The present invention discloses a key joint motion estimation based method for estimating continuous human postures. A motion estimation block matching algorithm is applied to human key joint tracking, so as to obtain continuous human posture results. Meanwhile, the results are continuously corrected by using a deep neural network based human posture estimator. The present invention may estimate the continuous human postures in a video stream, where the human postures are specifically embodied as coordinate positions of human joints in a video frame. Compared with a posture estimation method completely relying on a deep neural network, the posture estimation method provided by the present invention has the advantages of high frame rate, low hardware requirements, and sequential continuity of recognition results; and compared with a posture estimation method completely relying on a motion estimation algorithm, the present invention may correct a cumulative error, to improve the estimation accuracy.

Joint motion estimation based method for estimating continuous human postures

A multi-modal robot system capable of realizing autonomous submarine docking and reconstruction, includes a non-contact charging platform, several functional modules and a docking robot, wherein a coil is arranged on the non-contact charging platform, and energy and signals are transmitted between the coil and the functional modules; each of the functional modules includes a propeller module, an energy module, an observation detection module, an operation tool module, a navigation control module, a bow detection module and a tail propulsion module; and according to an instruction of a docking and reconstruction task, the docking robot reconstructs a multi-modal underwater robot by docking specific functional modules. The multi-modal robot system of the present invention can reconstruct the operation modal of the robot according to actual needs, realize the reconstruction and assembly of the underwater robot with three motion modes of horizontal plane cruise, fixed-point operation and vertical profiling, and reduce cost.

Multi-modal robot system capable of realizing autonomous submarine docking and reconstruction

The present invention discloses a flexible dynamic capturing system for an underwater moving carrier, including a shell, a front-end flexible guide apparatus, at least one tail flexible hand clamping and fastening apparatus, and a bottom retracting apparatus fastened in the shell. The front-end flexible guiding apparatus includes several flexible arms, the several flexible arms are fastened on a first rigid base in a circumferential array. The flexible arms can bend to an inner side and outer side of a circumference respectively when driven. Each tail flexible hand clamping and fastening apparatus includes several flexible claws, and the several flexible claws are divided into two groups and fastened on a second rigid base. When the flexible claws are driven, the two groups of flexible claws can bend and wind relative to each other. The bottom retracting apparatus implements lifting or lowering of the front-end flexible guide apparatus and the tail flexible hand clamping and fastening apparatus. The flexible dynamic capturing system for the present invention can capture underwater moving carriers of different diameters, and can be retracted and stored with a compact volume.

Flexible dynamic capturing system for underwater moving carrier

Growing community of inventors

HANGZHOU, China

Canjun Yang