<p>This paper investigates the adaptive image-based visual servoing (IBVS) control problem for excavators, explicitly accounting for hydraulic actuator dynamics. In contrast to conventional joint-space control methods that rely on known inverse kinematics, the proposed task-space strategy directly maps operational tasks to image-space trajectory tracking via visual feedback. The algorithm effectively decouples kinematic and dynamic loops, minimizing coupling effects induced by hydraulic servo system. Within the dynamic loop, the controller compensates for lumped uncertainties using a smooth, integrable compensation signal. An innovative barrier function-based constraint mechanism regulates joint and reference velocity signals, mitigating singularities and overly aggressive commands to ensure operational safety. Furthermore, as direct measurement of image-space velocity is infeasible, the method employs an outer-loop observer combined with an inner-loop filter to suppress noise from repeated differentiation. Finally, the asymptotic convergence of image-space tracking errors is demonstrated through Lyapunov stability theory, and simulation results validate the effectiveness of the proposed algorithm.</p>