To enhance the pipeline leak detection system's efficiency, this paper presents a data-driven approach based on a collaborative cloud-side system. The increasing complexity of the leak detection process and the growing scale of pipeline transportation create challenges for data acquisition, transmission, and processing. Our proposed approach addresses these challenges by introducing an adaptive data compression and acquisition algorithm, which effectively reduces data redundancy and enables efficient collection of large volumes of pressure data. Fine-grained task division is performed for each link in the cloud-side collaborative system, based on the requirements of the cloud-side collaborative scheduling strategy. We propose a task for the pipeline leak detection system under the cloud-side collaborative system, based on the computation delay and transmission delay of the divided subtasks. The topological model of the pipeline leak detection system under cloud edge collaboration is also presented, and the optimization objective is defined as the edge controller utilization under the task execution time constraint. Furthermore, we use a genetic algorithm to solve the optimal scheduling strategy under the time constraint. And then we verify the effectiveness of the pipeline cloud edge collaborative leak detection method, achieving a rapid pipeline leak time alarm.