單鏡頭車輛測速系統

Abstract

現今交通工具上，由於強大的機械動力，更常常帶給了人不可挽回的傷害，如車輛高速行駛時會容易失控而發生意外，因此有相關交通速限法規的制定，而道路容易發生車禍的路段也常看到架設超速照相機來取締超速車輛。本論文要探討的就是利用攝影機進行車輛行駛速度的量測，目前測速設備最常看到的就是使用移動式或固定式雷達量測，但由於雷達測速設備相當昂貴，近年來逐漸有人提出使用攝影機進行車輛測速的研究。 而使用攝影機測速也有一定的難度，如車輛移動距離的判定、鏡頭本身畸變的誤差、路面傾角影響等，往往降底了攝影機測速的準確度，本論文主要探討攝影機鏡頭本身畸變誤差對於測速所造成的影響，也探討路面傾角的影響，提出車輛測速方法將其準確度提升，讓車速量測誤差值低於5%。當攝影機鏡頭完全沒有對影像畸變補償時，本論文使用泰勒展開式推導出影像畸變補償多項式，針對影像畸變補償。若攝影機影像畸變補償不足時，本論文使用畸變座標內插到理想座標方法對影像畸變補償，並在上述影像畸變補償後再考量路面傾角之影響，計算出準確車輛移動距離並計算車速。本論文所提技術針對低價位camera 以個人電腦進行實驗，經過本論文所提技術，可達到即時處理之目標，而車速估計之誤差範圍也可以控制在低於5%的誤差值內，已具備實際應用之可行性。

On today's vehicles, since strong mechanical power often brings irreparable harm, such as an accident resulting from going out of control when traveling at a high speed, therefore the regulations relevant to the speed limit are made and speed cameras are installed to ban the vehicles over the speed limit at the sections prone to accidents on the road. This study investigates the use of cameras to measure the vehicle speed. Currently, the most common equipments used to measure the speed come in mobile or stationary radar type. In recent years, many studies using the camera to measure the speed of the vehicle are proposed because the radar equipment is quite expensive. However, it is difficult to cameras for measuring the speed, such as the judgment of the moving distance of the vehicle, the deviation from the image distortion caused by lens, and the affection from the inclination of the road surface. All of them often reduce the accuracy in the use of cameras to measure the speed. This paper focuses on the influence of the measured speed resulting from the image distortion caused by the camera lens itself, and the impact of the road inclination as well. Measurement of the vehicle speed is proposed to enhance the accuracy, and allows the error less than 5 percent. When the camera lens is completely without image distortion compensation, this paper uses Taylor series to derive the polynomial of the image distortion compensation for image distortion compensation. If the image distortion compensation for the camera is not enough, this paper uses the method with the distorted coordinate interpolated into the ideal coordinate for the image distortion compensation. At the same time, the inclination of the road is considered after the image distortion compensation to calculate the correct distance of the movement of the vehicle and the speed of the vehicle. In this paper, the proposal uses a low-cost camera with a "PC" for the experiments. Through the proposed technique, real-time processing can be achieved. The error range of the vehicle speed measurement can also be controlled within the values below 5%. The feasibility has been available in practical applications.