In order to test the performance and application effect of roadside perception method based on camera-radar fusion in the multi-source heterogeneous data environment of expressways, a camera-radar fusion perception algorithm is implemented and deployed in the Yanqing-Chongli intelligent expressway of China. Firstly, a fusion perception and positioning algorithm based on mmWave radar and camera was developed and deployed in the test nodes on the roadside facilities of the expressway. Secondly, the positioning accuracy of the algorithm was tested and analysed. Then the structured data results of the fusion perception algorithm are further analysed and processed to realize the management of target information library. Finally, some practical scenario applications have been proposed based on previous fusion perception results. The results show that the proposed algorithm can achieve lane level positioning and the deployed applications have achieved good results.