Local Threshold Values Research Articles

A quantitative indicator framework for stand level evaluation and monitoring of environmentally sustainable forest management

An indicator framework was designed as an operational science-based tool for the evaluation of the environmental aspects of sustainable forest management at stand level in Flanders (Belgium). The framework aims to assess the effects of forest management on forest composition, structure and functioning. It consists of seven principles and 19 criteria, to which 157 potential indicators, selected from literature, were assigned; 40 of these were considered as suitable by an expert panel, based on 10 operational selection criteria. All indicators were quantitative variables measurable in the field. After elaboration of a measurement protocol, the indicator framework was validated in 115 forest stands, distributed over the three main forest types of Flanders. The new indicator framework exhibited greater sensitivity to forest management practices and demonstrated better discriminating power than the method that is currently used by the Flemish forest administration to estimate the naturalness and environmental quality of a forest stand. Following a detailed cost calculation of each indicator and based on the sensitivity of each indicator to forest management practices, the indicator framework was further reduced to a final set of 29 indicators. This framework can also be applied in other regions, provided that local threshold values are defined to convert indicator values to indicator scores. The selection procedure and the possible contribution of this set to the forest management in Flanders are discussed.

The influence of skeletal muscle anisotropy on electroporation: in vivo study and numerical modeling

The aim of this study was to theoretically and experimentally investigate electroporation of mouse tibialis cranialis and to determine the reversible electroporation threshold values needed for parallel and perpendicular orientation of the applied electric field with respect to the muscle fibers. Our study was based on local electric field calculated with three-dimensional realistic numerical models, that we built, and in vivo visualization of electroporated muscle tissue. We established that electroporation of muscle cells in tissue depends on the orientation of the applied electric field; the local electric field threshold values were determined (pulse parameters: 8 × 100 μs, 1 Hz) to be 80 V/cm and 200 V/cm for parallel and perpendicular orientation, respectively. Our results could be useful electric field parameters in the control of skeletal muscle electroporation, which can be used in treatment planning of electroporation based therapies such as gene therapy, genetic vaccination, and electrochemotherapy.

A Threshold Adaptation based Voice Query Transcription Scheme for Music Retrieval

This paper presents a threshold adaptation based voice query transcription scheme for music information retrieval. The proposed scheme analyzes monophonic voice signal and generates its transcription for diverse music retrieval applications. For accurate transcription, we propose several advanced features including (i) Energetic Feature eXtractor (EFX) for onset, peak, and transient area detection; (ii) Modified Windowed Average Energy (MWAE) for defining multiple small but coherent windows with local threshold values as offset detector; and finally (iii) Circular Average Magnitude Difference Function (CAMDF) for accurate acquisition of fundamental frequency (F0) of each frame. In order to evaluate the performance of our proposed scheme, we implemented a prototype music transcription system called AMT2 (Automatic Music Transcriber version 2) and carried out various experiments. In the experiment, we used QBSH corpus [1], adapted in MIREX 2006 contest data set. Experimental result shows that our proposed scheme can improve the transcription performance.

Small Retinal Blood Vessel Tracking Using an Adaptive Filter

The problem of extracting small retinal blood vessels using a Canny edge detection method is addressed. It shows that using the Canny edge detector to detect retinal blood vessels, especially small vessels, can be significantly improved if an adaptive edge detection filter is incorporated. The filter is designed as a local dynamic hysteresis thresholding value generator. It adapts knowledge of the location of major vessels to define a small neighborhood and to generate the local hysteresis threshold values to detect meaningful edges, especially the edges of small blood vessels that may be missing, using Canny edge detector alone. The effectiveness of the adaptive edge detection filter is demonstrated by the preliminary experimental results obtained with the proposed method. A comparative test is also presented to highlight the performance differences between the Canny edge detector with the adaptive edge detection filter and the one without the filter.

DOCUMENT IMAGE BINARISATION USING A SUPERVISED NEURAL NETWORK

Advances in digital technologies have allowed us to generate more images than ever. Images of scanned documents are examples of these images that form a vital part in digital libraries and archives. Scanned degraded documents contain background noise and varying contrast and illumination, therefore, document image binarisation must be performed in order to separate foreground from background layers. Image binarisation is performed using either local adaptive thresholding or global thresholding; with local thresholding being generally considered as more successful. This paper presents a novel method to global thresholding, where a neural network is trained using local threshold values of an image in order to determine an optimum global threshold value which is used to binarise the whole image. The proposed method is compared with five local thresholding methods, and the experimental results indicate that our method is computationally cost-effective and capable of binarising scanned degraded documents with superior results.

MUSEMBLE: A novel music retrieval system with automatic voice query transcription and reformulation

So far, many researches have been done to develop efficient music retrieval systems, and query-by-humming has been considered as one of the most intuitive and effective query methods for music retrieval. For the voice humming to be a reliable query source, elaborate signal processing and acoustic similarity measurement schemes are necessary. On the other hand, recently, there has been an increased interest in query reformulation using relevance feedback with evolutionary techniques such as genetic algorithm for multimedia information retrieval. However, these techniques have not been exploited widely in the field of music retrieval. In this paper, we develop a novel music retrieval system called MUSEMBLE (MUSic enEMBLE) based on two distinct features: (i) A sung or hummed query is automatically transcribed into a sequence of pitch and duration pairs with improved accuracy for music representation. More specifically, we developed two new and unique techniques called WAE (windowed average energy) and dynamic ADF (amplitude-based difference function) onsets for more accurate note segmentation and onset/offset detection in acoustic signal, respectively. The former improved energy-based approaches such as AE by defining small but coherent windows with local and global threshold values. On the other hand, the latter improved the AF (amplitude function) that calculates the summation of the absolute values of signal differences for the clustering energy contour. (ii) A user query is reformulated using user relevance feedback with a genetic algorithm to improve retrieval performance. Even though we have especially focused on humming queries in this paper, MUSEMBLE provides versatile query and browsing interfaces for various kinds of users. We have carried out extensive experiments on the prototype system to evaluate the performance of our voice query transcription and genetic algorithm-based relevance feedback schemes. We demonstrate that our proposed method improves the retrieval accuracy up to 20–40% compared with other popular RF methods. We also show that both WAE and Dynamic ADF methods improve the transcription accuracy up to 95%.

Agronomical and environmental evaluation of a long-term experiment with cattle slurry and supplemental inorganic N applications in silage maize

During 19 years of research on silage maize grown a fertile sandy loam soil in Flanders, we compared the use of only inorganic nitrogen (S−) with a system of an annual application of dairy cattle slurry (S+, corresponding with an average rate of 180 kg N ha −1) supplemented with inorganic fertilizer nitrogen. The objectives were to study the efficiency of slurry-N, to determine the economic optimum of additional fertilizer-N use and to relate this optimum with the local threshold on residual soil nitrate-N. During the last 5 years, the economically optimal fertilizer rates in the S− and the S+ system were centred around 150 and 90 kg N ha −1, respectively. So, saving on fertilizer N amounted to about 60 kg N ha −1. At these N rates, the risk for trespassing the local legal threshold value for residual soil nitrate-N content (90 kg ha −1, 0–90 cm) was small (less than one out of ten seasons); economic and ecological optima of N fertilization concurred well. Meeting the standards of the EU nitrates directive and/or a doubled fertilizer N price would urge to decrease the applied amount of inorganic N by ca. 50 kg ha −1, resulting in yield losses of about 5% on the studied sandy loam soil with an important content of total N (5300 kg ha −1, 0–30 cm). The ratio of N-fertilizer replacement to the total amount of applied slurry-N increased during the experimental period to an average level of 59% during the last 5 years.

Predicting electrical breakdown in polymeric insulators. From deterministic mechanisms to failure statistics

Breakdown theories yield a specified time to breakdown for fields exceeding a critical value, whereas experiment shows that times-to-breakdown and breakdown fields are different on a sample-to-sample basis. This has forced the use of a statistical approach to failure, often without a clear physical understanding of the pertinent parameters. A general methodology that allows failure statistics to be derived from a given breakdown mechanism is presented. A number of case studies are used to illustrate the way in which mechanistic features can be related to the parameters of the failure statistic. In particular it is shown that the Gumbel statistic and not the Weibull function is the one appropriate to failure initiation by random defects. Both the formative stage and initiation stage of breakdown are considered, and it is shown that features can be identified in the failure statistic related to known thresholds in the active breakdown mechanism. Aging is examined through the three main control features: activation energy, field enhancement factor, and threshold factor. It is shown that distributions in each of these parameters could give rise to the observed statistics above the estimated characteristic threshold field. However, distributions in the material variables that define the local threshold value allow some failures to occur in the sample set, even at fields below the characteristic threshold level, whereas distributions in activation energy do not. The possibility of predicting the lifetime of individual specimens is briefly addressed in closing.

Local activity indicators for hard-switching adaptive filtering of images with mixed noise

We present a quantitative analysis of several local activity indicator properties and provide comparisons between them. We illustrate that the considered local activity indicators can be applied as adaptation parameters to locally adaptive hard-switching processing (filtering) of images. The cases where images are corrupted with additive or multiplicative noise having different probability density functions and, possibly, spikes are considered. The advantages of the designed adaptive hard-switching filtering algorithms are discussed and demonstrated for simulated data. Recommendations concerning the selection of local activity indicators and threshold values are given.

Evaluation of a new satellite-based method for forest fire detection

Advanced Very High Resolution Radiometer (AVHRR)-based fire detection methods are considered in this work in order to assess their effective usefulness in the framework of civil programmes for fire risk and damage mitigation. The discussion is divided into the evaluation of the most commonly used methods and the description of a new fire detection procedure which is proposed in this paper. Commonly used detection methods are based on using absolute threshold values for decision tests. These values usually match only with very local, uniform (in space and time) situations, and are often inadequate when applied to heterogeneous, or simply different, geographical areas or seasons. A high number of false alarms, so high as to make the satellite-based product not operationally utilizable, is the main disadvantage of the fixed-threshold approach. The new fire-detection procedure proposed here makes use only of historical AVHRR data in order to automatically produce local (in space and time) threshold values, suitable for fire-event detection also in very critical situations. High fire discrimination capabilities with low false-alarm rates, simple unsupervised implementation and, above all, flexibility for automatic extension to completely different geographic areas and observation conditions, are the main advantages associated with this new technique. Results obtained for different Italian areas have been successfully compared with ground observations made by the Italian Forestry Service. Tests made over a long observation period show that, on cloud-free regions, more than 75% of significant forest fires are detected with less than 20% of false alarms.

Experimental analysis of self-affine fractured rock surfaces through shadow length measurements

Geometrical properties of fractured basalt and sandstone surfaces are studied by measuring the length distribution of shadows appearing under grazing illumination. For self-affine surfaces, the distribution is predicted to follow a power law of the shadow length with an exponent depending only on the self-affine exponent and an upper cut-off length related to the incidence angle and the amplitude of the roughness. A local thresholding procedure is described and allows to detect small shadows inside large illuminated zones and small light patches surrounded by dark regions. The dependence of the distribution on the incidence angle and on the local threshold value is studied and a criterion, based on the total length of the illuminated zones, for optimizing these parameters is introduced. For both basalt and sandstone samples, self-affine exponents obtained from length distributions verifying the criterion are in agreement with theoretical expectations using classical profilometry results. Limitations in the lowest usable incidence angle values due to deviations of the surface from self-affine characteristics are discussed and the possible determination of a roughness amplitude characterization parameter is evaluated.

Defining threshold values for residual soil N levels

Nitrogen (N) leaching in the Netherlands mainly takes place during the period of precipitation surplus from September 15–April 15. Soil mineral N after harvest is consequently the major source of nitrate pollution of groundwater. To reduce losses of nitrate to groundwater, a national leaching threshold value of 35 kg N ha −1 yr −1 is defined based on the maximum nitrate concentration in groundwater of 50 mg l −1 and the average annual precipitation surplus of 300 mm. Based on local weather conditions in the study area in the Northern part of the country, the leaching limit should be set at 42 kg N ha −1 yr −1. In this paper, N profiles after harvest are defined which do not result in leaching quantities exceeding the national and local leaching thresholds. The spatial and temporal variability of N leaching is quantified using dynamic simulation for 65 soil profiles in a farmer's field. Leaching is simulated for five different N profiles at the end of the growing season for a 20-year period. For the given initial N range (15–120 kg N ha −1 m −1), space–time relations for the wet period are linear. Different results among the 65 soil profiles in both space and time are shown. The required N profile at September 15 is calculated for the national and local threshold values of 35 and 42 kg N ha −1 yr −1 respectively, using three risk levels of exceeding these thresholds. Spatial interpolation of the required N profiles provided N target maps which can be used to focus and evaluate N fertiliser management.

Multi-modal image segmentation using a modified Hopfield neural network

In most computer vision applications, it is required to segment objects from a background. In case of a muti-modal image the segmentation is an involved problem in comparison to a bi-modal image. This paper deals with an adaptive technique for choosing local threshold values for faithful image segmentation. The image segmentation is done by generating a threshold surface which is determined by interpolating the image gray levels at points where the gradient is high, indicating probable object edges. The interpolation of edge points is done using a modified Hopfield neural network and the results are compared with that of a potential surface interpolation method.

QUANTIFICATION OF POROUS MEDIA USING COMPUTERIZED TOMOGRAPHY AND A STATISTICAL SEGREGATION THRESHOLD

A statistical segregation threshold (SST) that determines the spatial distribution of solids and pores incomputerized tomography (CT) images of porous media is presented. The proposed algorithm combines basic imageprocessing techniques with consideration of measurement errors and CT scanning resolution. While traditional thresholdmethods use only one or two parameters to identify pore features in CT images, the SST characterizes the spatialdistribution of features by segregating image voxels into groups and treating each group differently according to its knownstatistical properties. This provides a major improvement when pores are smaller than the image resolution. By classifyingthe CT image into mixed-component, peak and tail-voxel fractions, the proposed algorithm provides a semi-automatic andtheoretically sound technique for component segregation. The SST algorithm results are insensitive to the local thresholdvalue and sample heterogeneity. Three acrylic test objects with various void sizes and water content are used to test anddemonstrate the SST.

Effects of degradation-induced absorption in GaAs-d.h.s. laser diodes

The absorbing lines observed in degraded GaAs-d.h.s. laser diodes cause increased local heating and higher threshold values. Since absorption is partly saturable by strong optical signals, the transient behaviour resembles that of passive Q-switched lasers.