Enhancing underwater images in epicontinental sea is a challenging problem owing to the influence of ocean currents, the refraction, absorption and scattering of light by suspended particles, and the weak illumination intensity. Recently, different methods have relied on the underwater image formation model and deep learning techniques to restore the underwater image, but they tend to degrade underwater image, interference of background clutter and miss boundary details of blue regions. Improved image fusion and enhancement algorithm based on a priori dark channel is proposed in this paper. Image edge features sharpening and dark detail enhancement by homomorphism filtering in CIELab color space is realized. In RGB color space, the multi-scale retinal with color restoration (MSRCR) algorithm is used to improve color deviation and enhance color saturation, and the contrast-limited adaptive histogram equalization (CLAHE) algorithm is used to de fog and enhance image contrast. Finally, according to the dark channel images of the three processing results, the final enhanced image is obtained by linear fusion of multiple images and multiple channels. Experimental results demonstrate the effectiveness and practicality of the proposed method on various datasets.

Delta parallel robot has been widely used in precision parts processing, handling, sorting, and precision assembly, with high efficiency and motion stability being its important performance indexes. Any small line segment corner in trajectory planning causes tangential discontinuous trajectory abrupt change, and the process vibration and shock caused by it seriously affect its high-speed and high-precision performance. In this paper, a trajectory planning method combining Cartesian space and joint space is proposed. Firstly, the vector method and micro-element integration method are used to establish the complete kinematics and dynamics equations of the Delta parallel robot, and the inverse kinematics/dynamics model-solving program is written based on Matlab software. Secondly, the end-effector trajectory of the Delta parallel robot is planned in Cartesian space, and the data points and inverse control points of the end-effector trajectory are obtained by the normalization method. Finally, the data points and control points are mapped to the joint space through the inverse kinematics equation, and the fifth-order B-spline curve is adopted for quadratic trajectory planning, which realizes the high-order continuous smoothing of trajectory planning. The simulation and experiment results show that the trajectory smoothing performance of continuous high-order curvature change can be improved using the proposed method, the tracking error of the end-effector is reduced by 21.12%, and the peak torques change of three joints is reduced by 3.5%, 11.6% and 1.6%, respectively.