This paper presents a deep-learning based framework for addressing the problem of accurate cloud detection in remote sensing images. This framework benefits from a Fully Convolutional Neural Network (FCN), which is capable of pixellevel labeling of cloud regions in a Landsat 8 image. Also, a gradient-based identification approach is proposed to identify and exclude regions of snow/ice in the ground truths of the training set. We show that using the hybrid of the two methods (threshold-based and deep-learning) improves the performance of the cloud identification process without the need to manually correct automatically generated ground truths. In average the Jaccard index and recall measure are improved by 4.36% and 3.62%, respectively.

No abstract

Progressive tool wear due to abrasive carbon fibres is one of the main issues in machining of CFRP and responsible for the short tool life. Because of occurring wear during machining, the tool’s micro-geometry changes continuously resulting in higher process forces and an increasing risk for workpiece damages. In this paper, a novel analytical model is presented in order to predict the wear-related change of the micro-geometry in orthogonal machining of CFRP depending on the fibre orientation and the initial tool geometry. For this purpose, a concept called the wear rate distribution is introduced which represents a measure to quantify the wear rate along the active micro-geometry. Based on experimental investigation, it is shown that the shape of an arbitrary wear rate distribution between two closely spaced wear states can be approximated and parameterised with a “line - curve - line” approach. Using the authors’ previously published analytical force model, the wear rate distribution can be calculated as function of five wear parameters that are used to parameterise the active micro-geometry of an arbitrary wear state. Based on an iterative solver, this is used to simulate the tool wear progression during machining. For model validation, the simulation is compared to experimental data in terms of the cutting edge profiles, the amount of worn tool material and the process forces. Accordingly, the wear model is capable to reproduce the most important wear characteristics, e.g. the cutting edge rounding, the decreasing clearance angle and the increasing contact length at the flank face.

No abstract