Realtime multi-aircraft tracking in aerial scene with deep orientation network

Robotic vision-based crack detection in concrete bridges is an essential task to preserve these assets and their safety. The conventional human visual inspection method is time consuming and cost inefficient. In this paper, we propose a robust algorithm to detect cracks in a pixel-wise manner from real concrete surface images. In practice, crack detection remains challenging in the following aspects: (1) detection performance is disturbed by noises and clutters of environment; and (2) the requirement of high pixel-wise accuracy is difficult to obtain. To address these limitations, three steps are considered in the proposed scheme. First, a local pattern predictor (LPP) is constructed using convolutional neural networks (CNN), which can extract discriminative features of images. Second, each pixel is efficiently classified into crack categories or non-crack categories by LPP, using as context a patch centered on the pixel. Lastly, the output of CNN—i.e., confidence map—is post-processed to obtain the crack areas. We evaluate the proposed algorithm on samples captured from several concrete bridges. The experimental results demonstrate the good performance of the proposed method.

Section: Proposed Methodsmentioning

confidence: 99%

Pixel-Wise Crack Detection Using Deep Local Pattern Predictor for Robot Application

Wang

2018

“…Moreover, the adaption of camera parameters, as well as motion parameters estimation based on the integration of video measurements and data obtained by UAVs, have always to be seriously taken into account [32]. The relentless motivation for image classification tasks based on deep learning and aerial images captured by UAVs has led to an abundance of research including vehicles [24,33,34], aerial vehicles [35,36,37,38,39], roads [40], buildings [41,42], cracks [43], birds [44], cattle [45], and wilt [46] detection. Another remarkable approach for object detection in very high-resolution aerial images has been proposed in [47].…”

Section: Related Researchmentioning

confidence: 99%

Unsupervised Human Detection with an Embedded Vision System on a Fully Autonomous UAV for Search and Rescue Operations

Lygouras

Santavas

Taitzoglou

et al. 2019

185

Unmanned aerial vehicles (UAVs) play a primary role in a plethora of technical and scientific fields owing to their wide range of applications. In particular, the provision of emergency services during the occurrence of a crisis event is a vital application domain where such aerial robots can contribute, sending out valuable assistance to both distressed humans and rescue teams. Bearing in mind that time constraints constitute a crucial parameter in search and rescue (SAR) missions, the punctual and precise detection of humans in peril is of paramount importance. The paper in hand deals with real-time human detection onboard a fully autonomous rescue UAV. Using deep learning techniques, the implemented embedded system was capable of detecting open water swimmers. This allowed the UAV to provide assistance accurately in a fully unsupervised manner, thus enhancing first responder operational capabilities. The novelty of the proposed system is the combination of global navigation satellite system (GNSS) techniques and computer vision algorithms for both precise human detection and rescue apparatus release. Details about hardware configuration as well as the system’s performance evaluation are fully discussed.

“…The authors of the paper are mainly interested in applicability of the FGT method in the field of multirotor UAVs operating in various tasks and on different missions [42,43,44,45,46,47], including multi-UAV applications, where the highest level of reliability, safety and maneuvering precision is required. To shed a light on the scope of possible applications, a general idea has been presented in Figure 1 comprising a set of current applications.…”

Section: Uav Applications—practical Problemsmentioning

confidence: 99%

Real-Time Model-Free Minimum-Seeking Autotuning Method for Unmanned Aerial Vehicle Controllers Based on Fibonacci-Search Algorithm

Giernacki

Horla

Báča

et al. 2019

The paper presents a novel autotuning approach for finding locally-best parameters of controllers on board of unmanned aerial vehicles (UAVs). The controller tuning is performed fully autonomously during flight on the basis of predefined ranges of controller parameters. Required controller properties may be simply interpreted by a cost function, which is involved in the optimization process. For example, the sum of absolute values of the tracking error samples or performance indices, including weighed functions of control signal samples, can be penalized to achieve very precise position control, if required. The proposed method relies on an optimization procedure using Fibonacci-search technique fitted into bootstrap sequences, enabling one to obtain a global minimizer for a unimodal cost function. The approach is characterized by low computational complexity and does not require any UAV dynamics model (just periodical measurements from basic onboard sensors) to obtain proper tuning of a controller. In addition to the theoretical background of the method, an experimental verification in real-world outdoor conditions is provided. The experiments have demonstrated a high robustness of the method to in-environment disturbances, such as wind, and its easy deployability.