Detecting Slag Formations with Deep Convolutional Neural Networks
This work addresses a domain-specific industrial automation problem for furnace operations, but it is incremental as it builds on existing deep learning methods with a minor architectural modification.
The paper tackled the problem of detecting slag formations in furnace images despite camera obstructions by introducing a convLSTM-layer into deep convolutional neural networks, resulting in sufficient performance for automating timely countermeasures with fewer outliers and lower variance in detection.
We investigate the ability to detect slag formations in images from inside a Grate-Kiln system furnace with two deep convolutional neural networks. The conditions inside the furnace cause occasional obstructions of the camera view. Our approach suggests dealing with this problem by introducing a convLSTM-layer in the deep convolutional neural network. The results show that it is possible to achieve sufficient performance to automate the decision of timely countermeasures in the industrial operational setting. Furthermore, the addition of the convLSTM-layer results in fewer outlying predictions and a lower running variance of the fraction of detected slag in the image time series.