Barna Keresztes

Seismic Fault Detection Based on a Curvilinear Support

In this paper, we present a new approach for seismic fault detection. Our goal is to increase the detection accuracy by computing some classical attributes on a support founded on an a priori knowledge about the faults. Two forms of support are proposed: one approximating the fault by a set of linear sub-segments of fixed length, the other founded on a more complex curved support which aims to describe the whole fault system. In the second case, computing all the possible configurations to detect the real location of the faults is illusory; then, we propose a fault detection algorithm based on a stochastic approach. One interest of this approach is the possibility of using a common support for different fault detection operators. Then a whole detection framework can be proposed which acts like a decision fusion process.

Near infrared hyperspectral dataset of healthy and infected apple tree leaves images for the early detection of apple scab disease

This dataset presents two series of hyperspectral images of healthy and infected apple tree leaves acquired daily, from two days after inoculation until an advanced stage of infection (11 days after inoculation). The hyperspectral images were calibrated by reflection correction and registered to match the geometry of one reference image. On the last experiment day, scab positions are provided.

Seismic fault detection using marked point processes

A stochastic model for fault detection in 2D sections of seismic blocks is presented. The model is a marked point process in which each fault is modeled by a parabola.

Potential of on-board colour imaging for in-field detection and counting of grape bunches at early fruiting stages

In order to enable the wine industry to anticipate in field work and marketing strategies, it is necessary to provide early assessments of vine productivity. The proposed method is designed for the detection and the measurement of grape bunches between the flowering season and the early fruition stages, before ‘groat-size’. The method consists of determining the affiliation of a pixel to a grape cluster based on colorimetric and texture features, using an SVM supervised classifier. The eventual affiliation of the pixels is achieved with an average reliability above 75%, which lets us envision in the near future the possibility of estimating the real number of grape bunches.