Looking and listening through complex media

光的散射使雾、油漆、生物组织等复杂介质以及我们周围的大多数材料呈现出不透明的外观。这也是为什么我们不能看到内部或通过许多对象的原因：即使光不被吸收，多重散射也会将任何信息扰乱成一个均匀的光环，似乎包含很少的信息。但事实上，弹性光散射是一个完全确定性的过程，因此我们想要成像的物体的信息是保守的。事实上，这些信息是以一种混乱的形式存在的，我们不容易使用。近年来，这种混乱的信息可以为我们所用的想法开始进入科学界，今天波前整形使我们能够以前所未有的精度操纵散射光。适当的成像技术，利用波前整形和散射光的属性，我们的优势，仍然是少之又少，但基本的物理学是建立和成熟的发展，现实世界的应用。我们建议将波前整形技术应用到最近发展的飞行时间成像领域，其中检测单个光子的到达时间以重建我们没有直接视线的物体的位置和形状。这通常是通过发送超短激光脉冲从表面（例如墙壁）散射来实现的，它可以在隐藏的物体上反弹，再次向我们散射，然后被检测到。这种方法的问题在于，在每次反射时，光被分散到各个方向，因此最终只有非常小的一部分可用于检测。波前整形将允许我们控制光散射的位置，以便相对于当前方法将信噪比提高几个数量级。对散射光的控制也将使我们能够将激光直接聚焦在目标物体上并对其进行光栅扫描，从而使我们能够完成对隐藏物体的3D高光谱成像。我们将把同样的原理应用于“激光麦克风”问题，其中检测从振动表面反向散射的光，以远程重建产生振动的声音。利用波前整形，我们将增强后向散射光的方向性，增加信噪比，从而增加最大工作距离（> 100 m）和可以检测声音的材料范围。此外，我们将利用光散射的复杂特性，特别是散斑相关性，开发一种完全被动的技术（仅依赖于环境照明），以成像和跟踪隐藏在复杂环境中的物体。

Scattering of light is what gives complex media like fog, paint, biological tissues, and most materials around us their opaque appearance. It is also the reason why we can't see inside or through many objects: even if light is not absorbed, multiple scattering scrambles any information into a uniform halo which seems to contain very little information. But in fact, elastic light scattering is a fully deterministic process, so the information about the object we would like to image is conserved. What happens is that the information is present in a scrambled form, not readily usable by us.In recent years, the idea that this scrambled information can be used to our advantage started to make its way in to the scientific community and today wavefront shaping allows us to manipulate scattered light with an unprecedented accuracy. Proper imaging techniques that exploit wavefront shaping and the properties of scattered light to our advantage, are still few and far between, but the fundamental Physics is well established and ripe for the development of real-world applications.We propose to apply wavefront shaping techniques to the recently developed field of time-of-flight imaging, where the arrival time of single photons is detected to reconstruct position and shape of objects we have no direct line of sight with. This is usually accomplished by sending an ultrashort laser pulse to scatter from a surface (e.g. a wall) from where it can bounce on the hidden object, scatter again toward us, and then be detected. The problem with this approach is that at each bounce the light is dispersed in all directions, so that only a very small fraction is finally available for detection. Wavefront shaping will allow us to control where light is scattered, so as to increase the signal to noise ratio by orders of magnitude with respect to current approaches. The control over the scattered light will also allow us to focus the laser light directly on the target object and raster scan over it, thus allowing us complete 3D hyperspectral imaging of an otherwise hidden object.We will apply the same principle to the problem of "laser microphone", where the light backscattered from a vibrating surface is detected to remotely reconstruct the sound producing the vibration. Exploiting wavefront shaping we will enhance the directionality of the backscattered light, increasing the signal to noise, and thus both the maximum working distance (>100m) and the range of materials one can detect sound from.Furthermore, we will take advantage of the complex properties of light scattering, and in particular of speckle correlations, to develop a completely passive technique (relying only on ambient illumination) to image and track objects hidden in a complex environment.