Image-Based Model for Assessment of Wood Chip Quality and Mixture Ratios

Abstract

This article focuses on fuel quality in biomass power plants and describes an online prediction method based on image analysis and regression modeling. The main goal is to determine the mixture fraction from blends of two wood chip species with different qualities and properties. Starting from images of both fuels and different mixtures, we used two different approaches to deduce feature vectors. The first one relied on integral brightness values while the latter used spatial texture information. The features were used as input data for linear and non-linear regression models in nine training classes. This permitted the subsequent prediction of mixture ratios from prior unknown similar images. We extensively discuss the influence of model and image selection, parametrization, the application of boosting algorithms and training strategies. We obtained models featuring predictive accuracies of R2 > 0.9 for the brightness-based model and R2 > 0.8 for the texture based one during the validation tests. Even when reducing the data used for model training down to two or three mixture classes—which could be necessary or beneficial for the industrial application of our approach—sampling rates of n < 5 were sufficient in order to obtain significant predictions.

A comprehensive review on wood chip moisture content assessment and prediction

“…Table 2 shows a list of such studies. The most relevant image-based study was conducted by Plankenbuhler et al [45], which proposed an imagebased method for wood chip quality assessment. The key objective of the study [45] was to determine the mixture ratios from blends of two wood chip species of different qualities and characteristics, as shown in Fig.…”

“…The obtained results demonstrated that the prediction accuracy of the brightness-based model (R 2 > 0.9) is greater than that of the texture-based model (R 2 > 0.8). Apart from RGB (Red Green Blue) images [45,77,78], laser scanned image [50,79] and hyperspectral images [36] have also been explored for wood chip quality control. Gillespie et al [36] combined the NIR hyperspectral imaging with chemometrics to determine the moisture content and specific energy of wood chips and achieved a prediction accuracy of R 2 = 0.94 and RMSE of 1.11%.…”

“…Scanned image [46] Determining size distribution of poplar wood chip Partial least square regression (PLS-R) Image and multivariate data [47] Sorting wood wastes Multi-layer perceptron (MLP) Image [78] Predicting mixture ratio of two wood chip species with varying properties Linear and non-linear regression Image [45] Sorting wood wastes A graphical-analytical method involving segmentation of binary image obtained through Otsu's method Laser scan [79]…”

Computer Aided Modeling of Wood Chips Transport by Means of a Belt Conveyor with Use of Discrete Element Method

“…They refer primarily to the tests on the raw material as an energy product at the various stages of processing and in various applications. The experimental tests analyze the impact of chips origin due to the calorific value and exhaust fumes emission, e.g., on account of tree species [14], mixture proportions [15], an acquiring method [14,16], a drying method [17], and pollution in the event of wood chips from recycling [18]. There are also the descriptions of the methods of processing wood chips to fuels in the form of ethanol fuel [19] or concentrated pellet [20].…”

Improving the quality of sorting wood chips by scanning and machine vision technology

“…High-quality wood chips are usually characterized by the type of wood, volume fraction of the chip content of the non-standard size, thick chips, uniform density, absence of bark and rot (Schön et al 2019;Plankenbühler et al 2020). The quality of wood chips is also affected by the grinding process.…”