Visual Control of Prey Pursuit

视觉控制猎物追踪

• 批准号: 9728032
• 负责人: Cole Gilbert
• 金额: $ 9.56万
• 依托单位: Cornell Univ - State: AWDS MADE PRIOR MAY 2010
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-08-15 至 2000-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9728032&HistoricalAwards=false
• 关键词: Visual; Control; Prey; Pursuit

IBN 97-28032 GILBERT Detection of targets while moving is difficult for systems that track targets visually, because the observer's own movements degrade the target image by blurring it. The blur causes a decrease in visual contrast between the target image and the background that makes it difficult to localize the target. Such motion-induced problems in target detection exist for moving animals, such as visually-hunting predators, and also for man-made systems, such as optically-guided robots. Predatory animals have evolved various strategies for coping with degradation of target images during pursuit. Tiger beetles are diurnal, visually-hunting predators that pursue their moving prey by running discontinuously, i.e., after initially localizing its prey, a beetle runs very fast (over 0.5 meter per second), then stops briefly (a few hundred milliseconds), then runs fast again toward the prey. This pattern of stop-and-go is repeated three or four times until the continuously moving prey is caught. In preliminary experiments, Dr. Cole Gilbert filmed beetles pursuing live prey or experimenter-controlled prey dummies. Frame-by-frame analysis indicated that a beetle stops only long enough to relocalize the prey. While a beetle is stopped, the image of the prey is not degraded and thus may be easier to see and localize. Subsequently, while the beetle is quickly running toward the prey, its image is degraded, perhaps so much so that the beetle loses track of it and must stop again. This hypothesis is supported by results from experiments in which beetles pursued prey dummies of very high contrast. In such pursuits, beetles were able to run continuously after the prey without stopping, suggesting that the visual contrast of the target image determines the type of pursuit the beetle can afford. Dr. Gilbert will investigate visually-guided pursuit by tiger beetles using a newly acquired digital camera capable of much higher speeds than his previous equipment . In particular, he will more rigorously test the hypothesis that viewing conditions determine whether the beetle performs continuous or stop-and-go pursuit. In short, is the beetle's switch in behavior an adaptive response to the amount of degradation of the target image? Various levels of target contrast and velocity will be used to produce different amounts of degradation and the dynamic properties of the resulting pursuits will be analyzed. The results from the project will elucidate the strategy used by the tiger beetle to cope with degradation of target images caused by the beetle's own high velocity. Moreover, the results may suggest useful strategies for man-made exploratory machines that often need to efficiently survey large areas quickly. The project may reveal that in some situations, the beetle's strategy of moving quickly then stopping to gather visual information may be more efficient than moving slowly enough to avoid image blur so that visual information can be gathered continuously.

IBN 97-28032 GILBERT 对于视觉跟踪目标的系统来说，在移动时检测目标很困难，因为观察者自己的运动会模糊目标图像，从而降低目标图像的质量。 模糊会导致目标图像与背景之间的视觉对比度降低，从而难以定位目标。 对于移动的动物（例如视觉狩猎捕食者）以及人造系统（例如光导机器人）来说，目标检测中存在这种由运动引起的问题。 掠食性动物已经进化出各种策略来应对追逐过程中目标图像的退化。 虎甲虫是昼夜性的视觉捕食者，它们通过不连续的奔跑来追捕移动的猎物，即，在最初定位猎物后，甲虫跑得非常快（每秒超过 0.5 米），然后短暂停止（几百毫秒），然后再次快速跑向猎物。 这种走走停停的模式重复三到四次，直到捕获不断移动的猎物。 在初步实验中，科尔·吉尔伯特博士拍摄了甲虫追逐活猎物或实验者控制的猎物假人的视频。逐帧分析表明，甲虫只停留足够长的时间来重新定位猎物。 当甲虫被阻止时，猎物的图像不会退化，因此可能更容易看到和定位。 随后，当甲虫快速跑向猎物时，它的图像会退化，甚至可能会导致甲虫失去踪迹而必须再次停下来。这一假设得到了实验结果的支持，在这些实验中，甲虫追逐对比度非常高的虚拟猎物。 在这种追逐中，甲虫能够不停地追赶猎物，这表明目标图像的视觉对比度决定了甲虫可以承受的追逐类型。 吉尔伯特博士将使用新购买的数码相机来研究虎甲虫的视觉引导追踪，该数码相机的速度比他以前的设备快得多。 特别是，他将更严格地检验这样的假设：观察条件决定甲虫是进行连续追击还是走走停停的追击。 简而言之，甲虫的行为转变是对目标图像退化程度的适应性反应吗？ 不同级别的目标对比度和速度将用于产生不同程度的退化，并且将分析由此产生的追踪的动态特性。 该项目的结果将阐明虎甲虫用于应对甲虫自身高速造成的目标图像退化的策略。 此外，这些结果可能为通常需要快速有效地勘测大面积的人造勘探机器提出有用的策略。 该项目可能表明，在某些情况下，甲虫快速移动然后停下来收集视觉信息的策略可能比缓慢移动以避免图像模糊更有效，从而可以连续收集视觉信息。