PROSPECT: Precision Robot Spectroscopy Exploration and Characterization Tool

Abstract

Near Infrared (NIR) spectroscopy is widely used in industrial quality control and automation to test the purity and grade of items. In this research, we propose a novel sensorized end effector and acquisition strategy to capture spectral signatures from objects and register them with a 3D point cloud. Our methodology first takes a 3D scan of an object generated by a time-of-flight depth camera and decomposes the object into a series of planned viewpoints covering the surface. We generate motion plans for a robot manipulator and end-effector to visit these viewpoints while maintaining a fixed distance and surface normal. This process is enabled by the spherical motion of the end-effector and ensures maximal spectral signal quality. By continuously acquiring surface reflectance values as the end-effector scans the target object, the autonomous system develops a four-dimensional model of the target object: position in an $R^3$ coordinate frame, and a reflectance vector denoting the associated spectral signature. We demonstrate this system in building spectral-spatial object profiles of increasingly complex geometries. We show the proposed system and spectral acquisition planning produce more consistent spectral signals than naive point scanning strategies. Our work represents a significant step towards high-resolution spectral-spatial sensor fusion for automated quality assessment.

Figures (7)

• Figure 1: PROSPECT manipulator mounted to robot arm and scanning a surface with observed Visible-Near Infrared (VNIR) reflectance profile and intersecting region of points forming the signature.
• Figure 2: (Left) PROSPECT end-effector component diagram showing subcomponents. Individual motors and linkages are numbered. (Right) Identification of parameters and axes for kinematics and motion planning.
• Figure 3: Histograms of errors from motion capture pose tracking.
• Figure 4: Model of spectroscopic measurement and modeling of object reflectance. Light is emitted from the PROSPECT platform (yellow) and reflected (green) by the surface $S$ (blue). The conical acceptance profile is defined by offset distance from the surface $d$, subsurface penetration $\epsilon$, and acceptance angle $\theta_{max}$. The intersection of the acceptance cone with the point cloud yields a subset of points $S_c$ (pink) with which the observed spectral signature is associated.
• Figure 5: Experiments for sparse spectral-spatial modeling. Each subfigure also includes an RGB camera image of the scene, point cloud measurement, and spectral reflectance curves for extracted voxels. (a) Calibrated color checkerboard (Spyder) (b) Triangular sandstone slab (c) Gypsum boulder with Iron Oxide deposits.