Road traffic sign posts are instructions for road use, which can ensure the orderly and safe travel of vehicles. Together, the traffic sign posts or traffic sign boards on the road and the decoration are a beautification of the road. According to the setting and construction of different types of traffic sign posts, correct information can be provided to road users, and vehicles can arrive at the destination smoothly under the guidance of traffic sign boards.

Therefore, how to ensure the safety of traffic sign post materials is an important means to improve road safety. At present, some experts have given good research results. For example, the literature proposed a multi standard building information detection method, using the sliding window to get the sample module, input the sample module to CNN and complete the classification of guangben in combination with SVM, and select some influences for testing. The test results show that the method can accurately detect the policy, but the method takes too long to detect. In this paper, a data detection method based on density is proposed, which uses the principle of sliding window to divide the data by sliding time window, screen all data according to information entropy, and detect the remaining data. The detection result of this method is inaccurate and needs to be improved.

In this paper, according to the improved data detection method, sliding and rolling are combined to reduce the amount of accounting, rolling verification is used to optimize the window, improve the speed of data matching, and complete the detection. The test results show that the method has high detection accuracy, but the detection time is too long. In view of the defects of the above methods, a method for corrosion resistance detection of traffic sign post data based on gradient motion method (GMM) is proposed. The anti-corrosive detection method of traffic sign column data is compared with the building information detection method based on multiple standards, the data detection method based on density and the improved data detection method for missing reporting rate