论文部分内容阅读
利用高分辨率遥感影像实现道路车辆监测在交通管理系统中占有重要地位,然而受成像条件及车辆本身颜色多样性的影响,很难从遥感影像中直接提取道路车辆。为此,本文提出了一种结合遗传算法和数学形态学的遥感影像道路车辆提取方法。首先利用模糊聚类方法建立影像分割模型,并利用遗传算法对其进行求解,从而实现影像中道路、车辆及环境的精确分割;然后对分割结果中不同层进行赋值,实现分割结果的二值化;最后利用数学形态学操作从分割的道路—车辆的二值结果中剔除道路,从而实现道路车辆的提取。利用本文方法对航空影像进行了道路车辆提取试验。试验结果表明,本文方法不仅能够准确识别地面车辆位置,还能精确提取其轮廓线。
The use of high-resolution remote sensing images to achieve road vehicle monitoring occupies an important position in traffic management systems. However, it is difficult to extract road vehicles directly from remote sensing images under the influence of imaging conditions and the color diversity of vehicles. To this end, this paper presents a remote sensing image road vehicle extraction method combining genetic algorithm and mathematical morphology. Firstly, the fuzzy clustering method is used to establish the image segmentation model, and the genetic algorithm is used to solve the problem. In this way, the accurate segmentation of road, vehicle and environment can be achieved. Then the different layers in the segmentation result are assigned to realize the binarization Finally, the mathematical morphology operation is used to remove the road from the segmented road-vehicle binary result to realize the extraction of road vehicles. Using the method of this paper, we carried out the road vehicle extraction test on aerial video. The experimental results show that this method not only can accurately identify the position of the ground vehicle, but also accurately extract its contour line.