Analysis of the accuracy of vehicle detection based on multiple linear regression

Abstract

Traffic volume surveys are a key factor in determining the state of road traffic and are a necessary basis for activities such as traffic planning, road construction, traffic control and management and engineering economic analysis. Most traffic volume surveys count vehicles separately by type, so accurate identification of vehicle types is particularly important. This paper is based on traffic volume survey data from eight detection methods for different vehicle types with separate detection counts. Statistical and correlation analyses were carried out separately, culminating in a multiple linear regression model, which yielded specific relationships between the effects of the different models when performing the tests.