Passive sensor evaluation for unmanned ground vehicle mud detection

Passive Polarized Vision for Autonomous Vehicles: A Review

This review article aims to address common research questions in passive polarized vision for robotics. What kind of polarization sensing can we embed into robots? Can we find our geolocation and true north heading by detecting light scattering from the sky as animals do? How should polarization images be related to the physical properties of reflecting surfaces in the context of scene understanding? This review article is divided into three main sections to address these questions, as well as to assist roboticists in identifying future directions in passive polarized vision for robotics. After an introduction, three key interconnected areas will be covered in the following sections: embedded polarization imaging; polarized vision for robotics navigation; and polarized vision for scene understanding. We will then discuss how polarized vision, a type of vision commonly used in the animal kingdom, should be implemented in robotics; this type of vision has not yet been exploited in robotics service. Passive polarized vision could be a supplemental perceptive modality of localization techniques to complement and reinforce more conventional ones.

Reflected near-infrared photography: Digging deeper into post-mortem examination

AIM

This research aims to investigate the utility of the Near Infra-Red (NIR) photographic technique in visualizing forensic evidence in a post-mortem examination.

MATERIAL AND METHODS

A total of twenty-six deceased (male: n = 15; female: n = 11) were brought to the Mortuary of King George's Medical University, Chowk, Lucknow, Uttar Pradesh, India (226003), and were used to investigate the applicability of the human visible spectrum (HVS) & near-infrared photography. In the investigation, a modified Nikon D5300 crop-framed digital single-lens reflex camera was used for NIR Photography in combination with a Micro-Nikkor 105 mm, f/2.8, (Nikon Inc., Melville, NY, USA), Hoya R72 (760nm-860nm) infrared filter and a Nikon D5600 crop-framed digital single-lens reflex camera in combination with Nikkor 50mm lens for Human Visible Spectrum (HVS) Photography.

RESULTS

The finding of the study reported that the application of the NIR photography would be the best of the investigative techniques for visualization and photo-documentation of forensically relevant post-mortem findings, such as - trace evidences (e.g., blood spots & soil particles on dark clothing), in external findings (e.g., contusion on victim's body) also for internal findings in a road traffic accident (RTA), gunshot, and drowning victim (e.g., mud particles in the trachea of drowning victim & localizing contusion of the scalp in road traffic accident cases).

CONCLUSION

Human visible spectrum (HVS) photographs taken with Nikon D5600 provides substantial evidence for documentation purpose, the best results for trace evidence & contusion visualization in Post-mortem examination of the deceased were achieved with the NIR Photography in combination with indirect sunlight & room light as an infrared light source. (At a wavelength of 760nm-860nm).

Deep Learning Based SWIR Object Detection in Long-Range Surveillance Systems: An Automated Cross-Spectral Approach

SWIR imaging bears considerable advantages over visible-light (color) and thermal images in certain challenging propagation conditions. Thus, the SWIR imaging channel is frequently used in multi-spectral imaging systems (MSIS) for long-range surveillance in combination with color and thermal imaging to improve the probability of correct operation in various day, night and climate conditions. Integration of deep-learning (DL)-based real-time object detection in MSIS enables an increase in efficient utilization for complex long-range surveillance solutions such as border or critical assets control. Unfortunately, a lack of datasets for DL-based object detection models training for the SWIR channel limits their performance. To overcome this, by using the MSIS setting we propose a new cross-spectral automatic data annotation methodology for SWIR channel training dataset creation, in which the visible-light channel provides a source for detecting object types and bounding boxes which are then transformed to the SWIR channel. A mathematical image transformation that overcomes differences between the SWIR and color channel and their image distortion effects for various magnifications are explained in detail. With the proposed cross-spectral methodology, the goal of the paper is to improve object detection in SWIR images captured in challenging outdoor scenes. Experimental tests for two object types (cars and persons) using a state-of-the-art YOLOX model demonstrate that retraining with the proposed automatic cross-spectrally created SWIR image dataset significantly improves average detection precision. We achieved excellent improvements in detection performance in various variants of the YOLOX model (nano, tiny and x).

Thermal Image Sensing Model for Robotic Planning and Search

This work presents a search planning system for a rolling robot to find a source of infra-red (IR) radiation at an unknown location. Heat emissions are observed by a low-cost home-made IR passive visual sensor. The sensor capability for detection of radiation spectra was experimentally characterized. The sensor data were modeled by an exponential model to estimate the distance as a function of the IR image's intensity, and, a polynomial model to estimate temperature as a function of IR intensities. Both theoretical models are combined to deduce a subtle nonlinear exact solution via distance-temperature. A planning system obtains feed back from the IR camera (position, intensity, and temperature) to lead the robot to find the heat source. The planner is a system of nonlinear equations recursively solved by a Newton-based approach to estimate the IR-source in global coordinates. The planning system assists an autonomous navigation control in order to reach the goal and avoid collisions. Trigonometric partial differential equations were established to control the robot's course towards the heat emission. A sine function produces attractive accelerations toward the IR source. A cosine function produces repulsive accelerations against the obstacles observed by an RGB-D sensor. Simulations and real experiments of complex indoor are presented to illustrate the convenience and efficacy of the proposed approach.

Handy method to calibrate division-of-amplitude polarimeters for the first three Stokes parameters

This paper presents a complete and original calibration framework for the three-CCD polarimetric cameras. These Division-of-Amplitude imaging polarimeters provide polarization images in real-time and open new applications in robotics. In order to fully exploit properties from polarization images, the sensor has to be calibrated leading sometimes to a tedious task that has to be undertaken with specific optical devices in a controlled environment. The proposed framework relies only on the use of a tablet and enables both to calibrate the geometric and the polarization settings of the camera. After rotating freely by hand the tablet in front of the camera, the system is automatically calibrated providing both the well-known geometric calibration matrix as well as the polarization calibration matrix. The last one is derived from the estimation of the orientation of the three polarizers, and the estimation of their relative values of degree of polarization and average transmittance.

VIS-NIR, SWIR and LWIR Imagery for Estimation of Ground Bearing Capacity

Ground bearing capacity has become a relevant concept for site-specific management that aims to protect soil from the compaction and the rutting produced by the indiscriminate use of agricultural and forestry machines. Nevertheless, commonly known techniques for its estimation are cumbersome and time-consuming. In order to alleviate these difficulties, this paper introduces an innovative sensory system based on Visible-Near InfraRed (VIS-NIR), Short-Wave InfraRed (SWIR) and Long-Wave InfraRed (LWIR) imagery and a sequential algorithm that combines a registration procedure, a multi-class SVM classifier, a K-means clustering and a linear regression for estimating the ground bearing capacity. To evaluate the feasibility and capabilities of the presented approach, several experimental tests were carried out in a sandy-loam terrain. The proposed solution offers notable benefits such as its non-invasiveness to the soil, its spatial coverage without the need for exhaustive manual measurements and its real time operation. Therefore, it can be very useful in decision making processes that tend to reduce ground damage during agricultural and forestry operations.