To improve the management efficiency and reduce the operational risks, a modern airport should not only be able to track the vehicles driving on the runway, but also be capable of detecting the deformation on the runway. To build such a smart airport, this paper takes measures from four perspectives: data, algorithm, computing power and platform. In terms of data, the optic fiber sensors (OFSs) rather than the traditional electromechanical sensors are used to collect the airport environmental data (AED). Compared with the electromechanical sensors, OFSs are cheaper, more reliable, and easier to be deployed. In terms of algorithms, the intelligent algorithms such as Convolutional Neural Networks (CNN), fast Fourier transform (FFT) and K-means are applied to analyze the AED. Compared with the traditional algorithms which detect the vehicle traces and runway deformation directly by signal pressure and amplitude, these algorithms are more precise and adaptable. In terms of computing power, the domain-specific architecture (DSA) technique is applied to increase the computing performance while keeping high energy efficiency. By designing several specific FPGA accelerators dedicated to the algorithms, the large quantity of AEDs can be processed quickly in real time. In terms of platform, a real-world edge-cloud collaborative platform based on the improved KubeEdge and Huawei openLooKeng is built. This platform can provide low-latency and high-performance computing, as well as data fusion for the AED processing in the airport. The work of this article has been practically applied to the Ezhou Huahu International Airport, and the real-world experimental results show that the proposed approaches have high detection accuracy, real-time data processing capability, low cost and also high energy efficiency.