Development of classification tools for multi-spectral LiDAR data

多光谱光探测和测距（LiDAR）传感器是一种新兴的技术，服务于民用、工业和军事领域。随着Teledyne Optech新型传感器“Titan”的推出，使用激光雷达技术的应用进入了一个新时代。Titan传感器是一种激光扫描仪传感器，可以同时在三种不同的波长（1550、1064和532 nm）下工作，并允许通过物体的激光能量反射来记录物体的特性。利用激光雷达传感器脉冲的反向散射能量（强度数据）在过去几年中得到了广泛的探索。最近的研究表明，强度数据在土地覆盖分类中具有潜在的应用价值。然而，需要对强度数据进行大量的辐射校正处理，并且需要特殊的分类算法来最大限度地利用激光雷达强度数据。拟议的项目将通过进一步开发和增强激光雷达强度数据的辐射校正模型，为泰坦技术做出贡献。研究小组将与工业伙伴合作，将其开发的辐射校正过程嵌入数据收集和交付过程中。拟议的项目还旨在开发新的图像和点分类算法，这些算法使用新传感器收集的距离和强度数据进行土地覆盖分类。特别重点是开发有效地从激光雷达多光谱数据自动圈定陆地/水的算法。该项目将对激光雷达行业产生直接影响，率先促进多光谱激光雷达数据的使用。它还将支持加拿大在激光行业的领导地位。

Multi-Spectral Light Detection and Ranging (LiDAR) sensors are a new and emerging technology serving several civil, industrial and military applications. With the launch of the new Teledyne Optech sensor "Titan", applications using LiDAR technology entered a new era. The Titan sensor is a laser scanner sensor which operates in three different wavelengths simultaneously (1550, 1064, and 532 nm) and allows recording of object properties from the reflectance of the laser energy of objects. The use of the back scatter energy of the LiDAR sensor pulses (intensity data) has been extensively explored over the last few years. Recent research showed the potential use of the intensity data in land cover classification. However, extensive processing is required to radiometrically correct the intensity data and there is a need for special classification algorithms to maximize the benefit from the use of the LiDAR intensity data. The proposed project will contribute to the Titan technology by further developing and enhancing the radiometric correction models of the LiDAR intensity data. The research team will collaborate with the industrial partner to embed its developments of the radiometric correction process within the process of data collection and delivery. The proposed project is also aiming to develop new image and point classification algorithms that use both the range and intensity data collected by the new sensor for land cover classification. Special focus is given to develop algorithms that work efficiently for land/water delineation automatically from the LiDAR multispectral data. The project will have a direct impact on the LiDAR industry, taking the lead on facilitating the use of multi-spectral LiDAR data. It will also support Canadian leadership in the laser industry.