Abstract: Substation automatic inspection technology based on electric inspection robot can quickly find the fault area and improve the inspection effect. However, due to the limited computing resources and energy of the robot terminal, how to process the image information collected by the inspection robot in real time and efficiently to realize rapid inspection is a challenge to be solved. Considering the presence of multiple robotic terminals offloading Computing tasks to multiple edge computing providers (ECP), This paper proposes a solution to unload computing tasks based on VCG (Vickrey-Clarke-Groves) auction mechanism to reduce task processing delay and maximize ECP revenue. Firstly, this paper establishes a heterogeneous network model in which multiple terminals unload tasks to multiple ECPs. In addition, considering that tasks may be divisible in a real task scenario, we modeled the subtasks as Directed acyclic graph (DAG), and proposed an offloading algorithm for computing tasks that took into account communication and computing resource latency. The algorithm considers the revenue of each ECP and designs the auction algorithm through the VCG auction mechanism to ensure the authenticity and effectiveness of the algorithm. Finally, simulation results show that the performance of this algorithm can quickly approximate the optimal unloading decision.