Filipe Neves dos Santos

Towards a Reliable Monitoring Robot for Mountain Vineyards

Crop monitoring and harvesting by ground robots on mountain vineyards is an intrinsically complex challenge, due to two main reasons: harsh conditions of the terrain and reduced time availability and unstable localization accuracy of the GPS system. In this paper is presented a cost effective robot that can be used on these mountain vineyards for crop monitoring tasks. Also it is explored a natural vineyard feature as the input of a standard 2D simultaneous localization and mapping approach (SLAM) for feature-based map extraction. In order to be possible to evaluate these natural features for mapping and localization purposes, a virtual scenario under ROS/Gazebo has been built and described. A low cost artificial landmark and an hybrid SLAM is proposed to increase the localization accuracy, robustness and redundancy on these mountain vineyards. The obtained results, on the simulation framework, validates the use of a localization system based on natural mountain vineyard features.

A visual place recognition procedure with a Markov chain based filter

Recognizing a place with a visual glance is the first capacity used by humans to understand where they are. Making this capacity available to robots will make it possible to increase the redundancy of the localization systems available in the robots, and improve semantic localization systems. However, to achieve this capacity it is necessary to build a robust visual place recognition procedure that could be used by an indoor robot. This paper presents an approach that from a single image estimates the robot location in the semantic space. This approach extracts from each camera image a global descriptor, which is the input of a Support Vector Machine classifier. In order to improve the classifier accuracy a Markov chain formalism was considered to constraint the probability flow according the place connections. This approach was tested using videos acquired from three robots in three different indoor scenarios - with and without the Markov chain filter. The use of Markov chain filter has shown a significantly improvement of the approach accuracy.

Towards a Reliable Robot for Steep Slope Vineyards Monitoring

Develop ground robots for crop monitoring and harvesting in steep slope vineyards is a complex challenge. Because of two main reasons: harsh condition of the terrain and unstable localization accuracy got from Global Positioning Systems (GPS). This paper presents a hybrid SLAM (VineSLAM) considering low cost landmarks to increase the robot localization accuracy, robustness and redundancy on these steep slope vineyards. Also, we present a cost-effective robot to carry-out crop monitoring tasks in steep slope vineyard environment. Test results got in a simulated and in a real test case supports the proposed approach and robot.

Agricultural Wireless Sensor Mapping for Robot Localization

Crop monitoring and harvesting by ground robots in steep slope vineyards is an intrinsically complex challenge, due to two main reasons: harsh conditions of the terrain and reduced time availability and unstable localization accuracy of the Global Positioning System (GPS). In this paper the use of agricultural wireless sensors as artificial landmarks for robot localization is explored. The Received Signal Strength Indication (RSSI), of Bluetooth (BT) based sensors/technology, has been characterized for distance estimation. Based on this characterization, a mapping procedure based on Histogram Mapping concept was evaluated. The results allow us to conclude that agricultural wireless sensors can be used to support the robot localization procedures in critical moments (GPS blockage) and to create redundant localization information.

Automatic Eye Localization; Multi-block LBP vs. Pyramidal LBP Three-Levels Image Decomposition for Eye Visual Appearance Description

This manuscript presents the performance evaluation of our algorithm that precisely finds human eyes in still gray-scale images and describes the state of the founded eye. This algorithm has been evaluated considering two descriptors - Pyramid transform domain (PLBP) and Multi-Block Histogram LBP (BHLBP), which are extended versions of the Local Binary Pattern descriptor (LBP). For the classification stage, two types of supervised learning techniques have also been evaluated, Support Vector Machine (SVM) and Multilayer Perceptron (MLP). The proposed method is assessed on the Face Recognition Grand Challenge (BioID) and (CAS-PEAL-R1) databases, and experimental results demonstrate improved performance than some state-of-the-art eye detection approaches.

Online Vision-Based Eye Detection: LBP/SVM vs LBP/LSTM-RNN

Eye detection is a complex issue and widely explored through several applications, such as human gaze detection, human-robot interaction and driver’s drowsiness monitoring. However, most of these applications require an efficient approach for detect the ocular region, which should be able to work in real time. In this paper, it is proposed and compare two approaches for online eye detection. The proposed schemes, work under real variant illumination conditions, using the conventional appearance method that is known for its discriminative power especially in texture analysis.

Towards Extraction of Topological Maps from 2D and 3D Occupancy Grids

Cooperation with humans is a requirement for the next generation of robots so it is necessary to model how robots can sense, know, share and acquire knowledge from human interaction. Instead of traditional SLAM (Simultaneous Localization and Mapping) methods, which do not interpret sensor information other than at the geometric level, these capabilities require an environment map representation similar to the human representation. Topological maps are one option to translate these geometric maps into a more abstract representation of the the world and to make the robot knowledge closer to the human perception. In this paper is presented a novel approach to translate 3D grid map into a topological map. This approach was optimized to obtain similar results to those obtained when the task is performed by a human. Also, a novel feature of this approach is the augmentation of topological map with features such as walls and doors.

EyeLSD a Robust Approach for Eye Localization and State Detection

Improving the safety of public roads and industrial factories requires more reliable and robust computer vision-based approaches for monitoring the eye state (open or closed) of human operators. Getting this information in real time when humans are driving cars or using hazardous machinery will help to prevent accidents and deaths. This paper proposes a new framework called EyeLSD to localize the eyes and detect their states without face detection step. For EyeLSD aims, two novel descriptors are proposed: enhanced Pyramidal Local Binary Pattern Histogram (ePLBPH) and Multi-Three-Patch LBP histogram (Multi-TPLBP). The performance of EyeLSD with ePLBPH and Multi-TPLBP is evaluated and compared against other approaches. For this evaluation three independent and public datasets were used: BioID, CAS-PEAL-R1 and ZJU datasets. The set EyeLSD, ePLBPH and Multi-TPLBP have a greater performance when compared against the state-of-the-art algorithms. The proposed approach is very stable under large range of eye appearances caused by expression, rotation, lighting, head pose, and occlusion.

Mining the usage patterns of ROS primitives

The Robot Operating System (ROS) is nowadays one of the most popular frameworks for developing robotic applications. To ensure the (much needed) dependability and safety of such applications we forecast an increasing demand for ROS-specific coding standards, static analyzers, and tools alike. Unfortunately, the development of such standards and tools can be hampered by ROS modularity and configurability, namely the substantial number of primitives (and respective variants) that must, in principle, be considered. To quantify the severity of this problem, we have mined a large number of existing ROS packages to understand how its primitives are used in practice, and to determine which combinations of primitives are most popular. This paper presents and discusses the results of this study, and hopefully provides some guidance for future standardization efforts and tool developers.

FOREST FIRE DETECTION WITH A SMALL FIXED WING AUTONOMOUS AERIAL VEHICLE

Abstract In this work a forest fire detection solution using small autonomous aerial vehicles is proposed. The FALCOS unmanned aerial vehicle developed for remote-monitoring purposes is described. This is a small size UAV with onboard vision processing and autonomous flight capabilities. A set of custom developed navigation sensors was developed for the vehicle. Fire detection is performed through the use of low cost digital cameras and near-infrared sensors. Test results for navigation and ignition detection in real scenario are presented.