Capturing aerial images by Unmanned Aerial Vehicles (UAV) allows gathering a general view of an agricultural site together with a detailed exploration of its relevant aspects for operational actions. Here we explore t...
详细信息
ISBN:
(纸本)9783319914794;9783319914787
Capturing aerial images by Unmanned Aerial Vehicles (UAV) allows gathering a general view of an agricultural site together with a detailed exploration of its relevant aspects for operational actions. Here we explore the challenging task of detecting cirsium arvense, a thistle-weed species, from aerial images of barley-cereal crops taken from 50m above the ground, with the purpose of applying herbicide for sitespecific weed treatment. The methods for automatic detection are based on object-based annotations, pointing out the RGB attributes of the Weed or Cereal classes for an entire group of pixels, referring to a crop area which will have to be treated if it is classified as being of the Weed class. In this way, an annotation belongs to the Weed class if more than half of its area is known to be covered by thistle weeds. Hence, based on object and pixel-level analysis, we compare the use of k-Nearest Neighbours (k-NN) and (feed-forward, one-hidden layer) neural networks, obtaining the best results for weed detection based on pixel-level analysis, based on a soft measure given by the proportion of predicted weed pixels per object, with a global accuracy of over 98%.
The development of data signal processing technology is changing with each passing day. For embedded systems, both the communication rate and the complexity of processingalgorithms are constantly being upgraded with ...
详细信息
The development of data signal processing technology is changing with each passing day. For embedded systems, both the communication rate and the complexity of processingalgorithms are constantly being upgraded with the development of embedded systems. The real-time performance of embedded data processing in many application fields and the reliability is also getting higher and higher. In the context that the embedded system composed of the traditional single processor is far from being able to meet the increasingly high data processing requirements of users, the application of multi-core DSP processing technology in the design of embedded systems is also more and more extensive. This study proposes an imageprocessing system based on multi-core DSP technology.
The proceedings contain 59 papers. The special focus in this conference is on Applications of Evolutionary Computation. The topics include: A Fast Metaheuristic for the Design of DVB-T2 Networks;a Genetic Algorithm fo...
ISBN:
(纸本)9783319775371
The proceedings contain 59 papers. The special focus in this conference is on Applications of Evolutionary Computation. The topics include: A Fast Metaheuristic for the Design of DVB-T2 Networks;a Genetic Algorithm for Community Detection in Attributed Graphs;maximizing the Effect of Local Disturbance in the Dynamics of Opinion Formation;accelerating the Computation of Solutions in Resource Allocation Problems Using an Evolutionary Approach and Multiagent Reinforcement Learning;achieving Optimized Decisions on Battery Operating Strategies in Smart Buildings;Phase-Space Sampling of Energy Ensembles with CMA-ES;many-Objective Optimization of Mission and Hybrid Electric Power System of an Unmanned Aircraft;evolving Controllers for Electric Vehicle Charging;Network Coordinated Evolution: Modeling and Control of Distributed systems Through On-line Genetic PID-Control Optimization Search;Task Classification Using Topological Graph Features for Functional M/EEG Brain Connectomics;piecemeal Evolution of a First Person Shooter Level;online-Trained Fitness Approximators for Real-World Game Balancing;recomposing the Pokémon Color Palette;mapping Chess Aesthetics onto Procedurally Generated Chess-Like Games;Evolving a TORCS Modular Fuzzy Driver Using Genetic algorithms;Self-adaptive MCTS for General Video Game Playing;deceptive Games;evolution of Convolutional Highway Networks;adapting Bagging and Boosting to Learning Classifier systems;an Automatic Feature Extraction Approach to image Classification Using Genetic Programming;feature Selection for Detecting Gene-Gene Interactions in Genome-Wide Association Studies;improving Evolutionary Algorithm Performance for Feature Selection in High-Dimensional Data;CGP4Matlab - A Cartesian Genetic Programming MATLAB Toolbox for Audio and imageprocessing;can the Relevance Index be Used to Evolve Relevant Feature Sets?;towards Evolutionary Super-Resolution;evolvable Deep Features.
Digital images are widely used and numerous application in different scientific fields use digital imageprocessingalgorithms where image segmentation is a common task. Thresholding represents one technique for solvi...
详细信息
ISBN:
(纸本)9781509060733
Digital images are widely used and numerous application in different scientific fields use digital imageprocessingalgorithms where image segmentation is a common task. Thresholding represents one technique for solving that task and Kapur's and Otsu's methods are well known criteria often used for selecting thresholds. Finding optimal threshold values represents a hard optimization problem and swarm intelligence algorithms have been successfully used for solving such problems. In this paper we adjusted recent elephant herding optimization algorithm for multilevel thresholding by Kapur's and Otsu's method. Performance was tested on standard benchmark images and compared with four other swarm intelligence algorithms. Elephant herding optimization algorithm outperformed other approaches from literature and it was more robust.
In this paper, imageprocessingalgorithms are designed using high level synthesis and implemented on different hardware platforms. These implementations are compared in terms of speed, latency and resource utilizatio...
详细信息
ISBN:
(纸本)9781538627457
In this paper, imageprocessingalgorithms are designed using high level synthesis and implemented on different hardware platforms. These implementations are compared in terms of speed, latency and resource utilization on various hardware platforms. The use of pipelining to reduce the latency is illustrated with an example. Using high level synthesis the designer has the opportunity to employ libraries similar to OpenCV and can take advantage of the productivity benefits of working at a higher level of abstraction, while creating high-performance hardware. Basic imageprocessingalgorithms like calculating histogram, histogram equalization, averaging filter and laplacian filter are chosen to explain hardware acceleration using high level synthesis. The workflow of implementing high level C / C++ / SystemC code in hardware using Vivado high level synthesis tool is explained along with implementation results of various imageprocessingalgorithms.
Thousands of years ago written language was introduced as a way of enhancing and facilitating communication. Fast forward to the twenty first century much has changed, especially the flow of data incrementing at fast ...
详细信息
The alignment of the sub-apertures is a major challenge for future segmented telescopes and telescope arrays. We show here that a focal plane wave-front sensor using only two images can fully and efficiently align a m...
详细信息
ISBN:
(数字)9781510619500
ISBN:
(纸本)9781510619500
The alignment of the sub-apertures is a major challenge for future segmented telescopes and telescope arrays. We show here that a focal plane wave-front sensor using only two images can fully and efficiently align a multiple aperture system, both for the alignment (large amplitude tip/tilt aberrations correction) and phasing (piston and small amplitude tip/tilt aberrations correction) modes. We derive a new algorithm for the alignment of the sub-apertures : ELASTICS. We quantify the novel algorithm performance by numerical simulations. We show that the residues are within the capture range of the fine algorithms. We also study the performance of LAPD, a recent real-time algorithm for the phasing of the sub-apertures. The closed-loop alignment of a 6 sub-aperture mirror provides experimental demonstration for both algorithms.
Internet of Things (IoT) is a key pillar in various smart-* domains, including smart home, smart building, intelligent transportation systems, industrial Internet, to name a few. Its potential in terms of value creati...
详细信息
ISBN:
(纸本)9781538643280
Internet of Things (IoT) is a key pillar in various smart-* domains, including smart home, smart building, intelligent transportation systems, industrial Internet, to name a few. Its potential in terms of value creation is extraordinary, according to economic estimates from many market sectors. At the same time, there are some technical challenges to face, which could refrain IoT adoption. In this study, the attention is focused on interoperability and data quality issues from a pragmatic standpoint. In particular, a real smart building testbed has been set-up by integrating two different IoT platforms: TLSensing and NetwOrked Smart Object (NOS). The former has been developed to enable the quick deployment of a IPv6 over the TSCH mode of IEEE 802.15.4e (6tisch) IoT network and to expose the data gathered from sensors through a web interface. The latter is a middleware that handles heterogeneous data sources and offers a unifying view to users. The resulting IoT system is able to: (i) gather data from multiple sources, including multimedia ones;(ii) expose network resources to user through a common interface;(iii) enforce security and data quality checks. The performance of this system has been experimentally analyzed to demonstrate a satisfactory responsiveness and ability to meet the aforementioned requirements.
Regenerated tissue is a one of the wide developing research topics of nowadays. The characterization of tissue in culture permits to regulate its final mechanical properties and to test the influence of pharmaceutical...
详细信息
Regenerated tissue is a one of the wide developing research topics of nowadays. The characterization of tissue in culture permits to regulate its final mechanical properties and to test the influence of pharmaceutical and cosmetic products on it. For this purpose, an embedded vision system was developed in order to characterize the mechanical properties of tissue in culture. Thus, the aim of the present study was to develop an embedded, automated and portable cell count algorithm which can quantify the fibroblast cells from the fluorescent images taken by developed FPGA driven embedded high definition (HD) camera module. The developed algorithm was integrated into "uCell" software and tested on fluorescent images. The automated cell counting results were compared with manual counting and showed that the algorithm is reliable in 95% percentage of cases. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
This article describes the problem of segmentation of the spine for lateral C spine radiographs. In this case, the most frequently used approach is the Active Shape Model. The use of the Active Appearance Model is con...
详细信息
ISBN:
(纸本)9783319669052
This article describes the problem of segmentation of the spine for lateral C spine radiographs. In this case, the most frequently used approach is the Active Shape Model. The use of the Active Appearance Model is considered in this paper. Segmentation quality of sample data is tested for selected preprocessing and predetecting edge algorithms: Sobel filter, Canny edge detection algorithm, and Statistical Dominance Algorithm. The particularly important issue of precise description of contours is considered and partially tested. The aim is to deliver a good quality preliminary step to syntactic analysis of vertebrae using the generalized shape language.
暂无评论