CAJ | 학술논문

针对UAV（Unmanned Aerial Vehicle）侦察目标识别中的实时性和鲁棒性的要求，提出一种基于SURF（Speeded Up Robust Features）的快速目标识别算法。对UAV侦察图像进行预处理，采用不变矩构造遗传算法的适应度函数，利用遗传算法的全局搜索能力快速地提取可能包含目标的ROI（Region Of Interesting）区域。在ROI区域和模板图像中提取SURF特征点，采用最近邻的匹配算法搜索匹配对，从而精确确定目标的位置。仿真结果显示，该算法可以明显地提高目标识别的实时性并具有相当的鲁棒性。
침대UAV（Unmanned Aerial Vehicle）정찰목표식별중적실시성화로봉성적요구，제출일충기우SURF（Speeded Up Robust Features）적쾌속목표식별산법。대UAV정찰도상진행예처리，채용불변구구조유전산법적괄응도함수，이용유전산법적전국수색능력쾌속지제취가능포함목표적ROI（Region Of Interesting）구역。재ROI구역화모판도상중제취SURF특정점，채용최근린적필배산법수색필배대，종이정학학정목표적위치。방진결과현시，해산법가이명현지제고목표식별적실시성병구유상당적로봉성。
In order to solve the problem of robust identification of target in the UAV reconnaissance image in real time, a fast recognition algorithm based on Speeded Up Robust Feature(SURF)is proposed in this paper. A similar matching function using moment invariant features is constructed as the fitness function in the genetic algorithm. Then the region of interest which may contain target is selected using genetic algorithm. Speeded up robust features are taken between the ROIs and the match pair is found using nearest matching algorithm. The match pair is used to determine the location of the target accurately. The result of simulation indicates that the algorithm is not only effective on enhancing the real-time ability of identifying target, but also robust.