| 摘要: | 科技的發達與進步，使得人們對於生活上的習慣有很重大的改變。其中交通工具的普遍更為顯著，而人們在追求技術進步的同時也在追求更符合人性的科技，因此對於行車安全方面的要求也就更加的重視。本論文整合了白天與夜間的前車偵測技術，首先我們提出一個白天夜晚的準則針對不同的時間使用不同的演算法。在判斷白天夜晚的準則中，我們在不同的影片取固定的影像數目 ，且每張大小為320?240影像裡我們取出固定數量和固定位置的像素點來代表整張影像，並求出取出像素的平均亮度和小於某個亮度門檻值的像素點數佔所有像素點的比例，依照這準則定義出門檻值來分別日間與夜晚，在日間的天候狀況下我們也分成了不同的情況，晴天、陰天、黃昏、收費站、車輛陰影、向陽、小雨和大雨。在日間偵測前方車輛的演算法中，先對整張影像做第二差分找出不同強度的垂直邊影像和水平邊影像。而為了適應不同的天候，在定義二值化門檻值前，利用車道線偵測的結果將影像上半部的背景濾除掉，最後定義出動態的門檻值做影像的二值化。最後將二值化的影像，找出具有連續像素點的水平邊，而水平邊分成位於車子底部陰影的負水平邊和車體上的正水平邊，我們根據水平邊的兩端點位置找出是否有顯著的垂直邊或是是否具有對稱的兩垂直邊，若符合此條件的水平邊，最後再根據水平邊和車道寬比例是否位在我們定義的道路寬比例中，若符合則產生候選車輛。在夜晚偵測的演算法中，我們使用車尾燈在影像中所呈現的特徵來偵測車尾燈。再以車尾燈的垂直距離、水平距離、移動軌跡、與兩車尾燈相關性等四項特徵配對，以偵測可能的候選車輛。最後再依據車尾燈對的距離變化量，計算與前車的碰撞時間 (time to collision, TTC) ，提供駕駛者警示。As the growing and progressive of technology, people change the styles of daily life, especially on the land vehicle. When people pursued advanced techniques as well as the humanistic technology, the traffic safety is also an important issue.We integrate the front car detection in day and night time, using different algorithm for the different time. The beginning we defined a criterion for judging what it day or night is, and decided what the algorithm we choose. The criterion judging day or night, we fetch fixed number of frame from different video. We fetch fixed pixels at fixed position for representing the whole image which size is 320×240, then calculating the average intensity and the percentage of bellow the threshold of pixels we fetch. According to the criterion, judging what it day or night is.In daytime forwarding collision warning system, we fetch vertical edge response image and horizontal edge response image by second difference mask. For adapting different weather, we filter the background at the upper half of the image according to the result of lane detection system before defining threshold. Then we generate the bi-level gradient image with appropriate threshold. Then find the continuous pixels belong to the horizontal edge, the horizontal edge we divided into positive and negative for representing the bottom of car and body of car respectively. We find whether it has enough of vertical edge patency or symmetric vertical edge pair at the end point of horizontal edge. If the horizontal edge confirmed, we calculate the proportion of width of the horizontal edge with lane with. If the proportion confirmed, the candidate vehicles generate.In nighttime forward collision warning system, we detect the tails of preceding vehicles. Then, we pair the lights using the features of the horizontal distance, the vertical heights, the trajectory, and correlation of a pair of lights. Finally, we estimate the time to collision (TTC) of the verified light pair and providing the warning for the driver. |
