基于场景复杂度的移动机器人目标搜索方法的研究

借鉴人类搜索经验,将场景复杂度概念应用到移动机器人目标搜索过程,提出了一种基于场景复杂度的移动机器人目标搜索方法.首先,通过分析影响场景复杂度的主要因素,给出了融合图像和激光深度信息的场景复杂度形式化定义;其次,结合视觉目标检测过程,描述了基于场景复杂度的目标搜索方法,该方法通过对各个场景按复杂度进行筛选来确定待搜索场景,并根据场景深度信息确定该场景的待搜索点.实验结果表明：本文提出的场景复杂度计算方法符合人的复杂度主观感受,较好地反映了场景的内在特征;基于场景复杂度的目标搜索方法可以有效地搜索目标,具有较好的鲁棒性.

Research on target search method for mobile robot based on scene complexity

Base on the concept of scene complexity which was from the human search experience, we present a novel target search method for mobile robot. In the method, formalization definition of scene complexity was firstly given by analyzing the main factors relative to scene complexity which combined image complexity with laser ranges complexity. Then, combined the processes of visual target detection, the complexity of several scenes were calculated according to the definition. Thirdly, the next search scene was selected by comparing the obtained scenes complexity and the next search point was determined by the laser ranges of the scene. Experimental results show that the method of scene complexity is in accord with the human evaluation and reflects the scene＇s intrinsic characteristics, and the proposed target search method is effective and robust.