Tool wear detection based on machine vision is the core to realize intelligent and automated cutting processing. It is essential to solve the problem of low-efficiency and low-quality acquisition of tool wear image acquisition in the machine detection process. In this paper, the image acquisition method by using a 6-DOF (degree of freedom) serial robot with an industrial camera is developed. The digital twin unit of the detection system is set up, the UR10 robot kinematics model and acquisition trajectory planning method are proposed, and Unity3D software is used to arrange the digital twin environment of the detection system. The synchronous mapping function of the robot digital model and physical entity is given through Socket communication to realize complex tool edges image acquisition trajectory planning, accurate teaching of virtual scenes, digital monitoring of the detection process, and optimization of system models. The stability and efficiency of robot kinematics, trajectory planning, and interactive communication in the tool wear image detection environment are verified. The research results show that the proposed method can provide support for the digital monitoring and control technology of the tool wear on-machine detection process. [ABSTRACT FROM AUTHOR]