In order to allow robust obstacle detection for autonomous freight traffic using freight trains or shunting locomotives, several different sensors are required. Humans and other objects must be detected so that the vehicle can stop in time. Laser scanners deliver distance information and are popular in robotics and automation. Cameras deliver further pieces of information on the environment and are especially useful for the classification of objects, but do not deliver distance measurements. Thermal cameras are ideal for the detection of humans based on their body temperature if the surrounding temperature is not too similar. It is only the combination of these different sensors which delivers enough robustness. Therefore a sensor fusion and an extrinsic calibration has to take place. This article presents an approach fusing a 2D and an 8-layer 3D laser scanner with a thermal and a Red-Green-Blue (RGB) camera, using a triangular calibration target taking all six degrees of freedom into account. The calibration was tested and the results validated during reference measurements and autonomous and manually controlled field tests. This sensor fusion approach was used for the obstacle detection of an autonomous shunting locomotive.

It is the aim of this thesis to present a visual body weight estimation, which is suitable for medical applications. A typical scenario where the estimation of the body weight is essential, is the emergency treatment of stroke patients: In case of an ischemic stroke, the patient has to receive a body weight adapted drug, to solve a blood clot in a vessel. The accuracy of the estimated weight influences the outcome of the therapy directly. However, the treatment has to start as early as possible after the arrival at a trauma room, to provide sufficient treatment. Weighing a patient takes time, and the patient has to be moved. Furthermore, patients are often not able to communicate a value for their body weight due to their stroke symptoms. Therefore, it is state of the art that physicians guess the body weight. A patient receiving a too low dose has an increased risk that the blood clot does not dissolve and brain tissue is permanently damaged. Today, about one-third gets an insufficient dosage. In contrast to that, an overdose can cause bleedings and further complications. Physicians are aware of this issue, but a reliable alternative is missing. The thesis presents state-of-the-art principles and devices for the measurement and estimation of body weight in the context of medical applications. While scales are common and available at a hospital, the process of weighing takes too long and can hardly be integrated into the process of stroke treatment. Sensor systems and algorithms are presented in the section for related work and provide an overview of different approaches. The here presented system -- called Libra3D -- consists of a computer installed in a real trauma room, as well as visual sensors integrated into the ceiling. For the estimation of the body weight, the patient is on a stretcher which is placed in the field of view of the sensors. The three sensors -- two RGB-D and a thermal camera -- are calibrated intrinsically and extrinsically. Also, algorithms for sensor fusion are presented to align the data from all sensors which is the base for a reliable segmentation of the patient. A combination of state-of-the-art image and point cloud algorithms is used to localize the patient on the stretcher. The challenges in the scenario with the patient on the bed is the dynamic environment, including other people or medical devices in the field of view. After the successful segmentation, a set of hand-crafted features is extracted from the patient's point cloud. These features rely on geometric and statistical values and provide a robust input to a subsequent machine learning approach. The final estimation is done with a previously trained artificial neural network. The experiment section offers different configurations of the previously extracted feature vector. Additionally, the here presented approach is compared to state-of-the-art methods; the patient's own assessment, the physician's guess, and an anthropometric estimation. Besides the patient's own estimation, Libra3D outperforms all state-of-the-art estimation methods: 95 percent of all patients are estimated with a relative error of less than 10 percent to ground truth body weight. It takes only a minimal amount of time for the measurement, and the approach can easily be integrated into the treatment of stroke patients, while physicians are not hindered. Furthermore, the section for experiments demonstrates two additional applications: The extracted features can also be used to estimate the body weight of people standing, or even walking in front of a 3D camera. Also, it is possible to determine or classify the BMI of a subject on a stretcher. A potential application for this approach is the reduction of the radiation dose of patients being exposed to X-rays during a CT examination. During the time of this thesis, several data sets were recorded. These data sets contain the ground truth body weight, as well as the data from the sensors. They are available for the collaboration in the field of body weight estimation for medical applications.

This paper describes the estimation of the body weight of a person in front of an RGB-D camera. A survey of different methods for body weight estimation based on depth sensors is given. First, an estimation of people standing in front of a camera is presented. Second, an approach based on a stream of depth images is used to obtain the body weight of a person walking towards a sensor. The algorithm first extracts features from a point cloud and forwards them to an artificial neural network (ANN) to obtain an estimation of body weight. Besides the algorithm for the estimation, this paper further presents an open-access dataset based on measurements from a trauma room in a hospital as well as data from visitors of a public event. In total, the dataset contains 439 measurements. The article illustrates the efficiency of the approach with experiments with persons lying down in a hospital, standing persons, and walking persons. Applicable scenarios for the presented algorithm are body weight-related dosing of emergency patients.

In order to increase the robustness of localisation and victim detection in low visibility situations it is necessary to fuse several sensors. The most common sensor used in robotics is the 2D laser scanner which delivers distance measurements. In combination with a camera the gained information can be supported by visual information about the environment. Thermal cameras are ideal for finding objects with a certain temperature, but they do not deliver distance information. The difficulty in fusing these two sensors is, that a correspondence between each distance measurement and its corresponding pixel within the thermal image needs to be found. As the laser scanner only displays one plane, this is not an intuitive task. A special triangular calibration target, covering all six degrees of freedom and being visible for both sensors, was developed. In the end the transformation between each laser scan point and its corresponding thermal image pixel is given. This allows for assigning every laser measurement within the field of view a corresponding thermal pixel. The final application will enable detection of human beings and display the distance required to reach them.

Body weight is a crucial parameter when it comes to drug or radiation dosing. In case of emergency treatment time is short so that physicians estimate the body weight by the visual appearance of a patient. Further, visual body weight estimation might be a feature for person identification. This paper presents the anthropometric feature extraction from RGB-D sensor data (Red, Green, Blue and Depth), recorded from frontal view. The features are forwarded to an artificial neural network for weight estimation. Experiments with 233 people demonstrate the capability of different features for body weight estimation. To prove robustness against sensor modalities, a structured light sensor is used, as well as a time-of-flight sensor. An additional experiment including temperature features from a thermal camera improves the body weight estimation beyond.

The invention relates to a device (100) for optical detection of a person (201) of a weight with a camera (101) which is adapted to receive an image pickup of the person, wherein the image recording comprises depth information, and a processor (103), which is adapted to convert the image pickup in a 3D pixel matrix (300e), wherein the processor (103) is configured to determine eigenvalues ​​of the 3D pixel matrix (300e), and wherein the processor (103) is further configured to the weight of the person ( to determine 201) based on the eigenvalues.

This publication describes a 2D Simultaneous Localization and Mapping approach applicable to multiple mobile robots. The presented strategy uses data of 2D LIDAR sensors to build a dynamic representation based on Signed Distance Functions. Novelties of the approach are a joint map built in parallel instead of occasional merging of smaller maps and the limited drift localization which requires no loop closure detection. A multi-threaded software architecture performs registration and data integration in parallel allowing for drift-reduced pose estimation of multiple robots. Experiments are provided demonstrating the application with single and multiple robot mapping using simulated data, public accessible recorded data, two actual robots operating in a comparably large area as well as a deployment of these units at the Robocup rescue league.

Body weight adapted drug dosages are important for emergency treatments: Inaccuracies in body weight estimation may lead to inaccurate drug dosing. This paper describes an improved approach to estimating the body weight of emergency patients in a trauma room, based on images from an RGB-D and a thermal camera. The improvements are specific to several aspects: Fusion of RGB-D and thermal camera eases filtering and segmentation of the patient's body from the background. Robustness and accuracy is gained by an artificial neural network, which considers geometric features from the sensors as input, e.g. the patient's volume, and shape parameters. Preliminary experiments with 69 patients show an accuracy close to 90 percent, with less than 10 percent relative error and the results are compared with the physician's estimate, the patient's statement and an established anthropometric method.

This publication describes the application of a weight estimation method for emergency patients in clinical environments, named Libra3D. Based on algorithms from robotic vision and control, the approach applies point cloud processing and filtering, to segment a patient from the stretcher and to estimate his weight. The application is working with a low-cost structured light sensor, fused with a thermal camera.

This paper describes the application of a weight estimation method for emergency patients in clinical environments. The approach applies established algorithms for point cloud processing and filtering to data from a low-cost, structured light sensor. A patient's volume is estimated on the basis of their visible front surface. The approach is currently being tested in the workflow of the emergency room at the Universitätsklinikum Erlangen, Germany. Preliminary results show the accuracy of the approach in relation to other conservative means of weight measurements, for example, by physicians and anthropometric measurements.

This paper describes a data integration approach for arbitrary 2D/3D depth sensing units exploiting assets of the signed distance function. The underlying framework generalizes the KinectFusion approach with an objectoriented model respecting different sensor modalities. For instance, measurements of 2D/3D laser range finders and RGB-D cameras can be integrated into the same representation. Exemplary, an environment is reconstructed with a 3D laser range finder, while adding fine details from objects of interest by closer inspection with an RGB-D sensor. A typical application of this approach is the exploration in rescue environments, where large-scale mapping is performed on the basis of long-range laser range finders while hollows are inspected with lightweight sensors attached to a manipulator arm.

This paper describes a data fusion approach for 3D sensors exploiting assets of the signed distance function. The objectoriented model is described as well as the algorithm design. We developed a framework respecting different modalities for multi-sensor fusion, 3D mapping and object localization. This approach is suitable for industrial applications having need for contact-less object localization like bin picking. In experiments we demonstrate 3D mapping as well as sensor fusion of a structured light sensor with a Time-of-Flight (ToF) camera.