Comparison between neural networks and partial least squares for intra-growth ring wood density measurement with hyperspectral imaging

Fernandes, Armando; Lousada, J.; Morais, J.J.L.; Xavier, José; Pereira, João Luís; Melo‐Pinto, Pedro

doi:10.1016/j.compag.2013.03.010

Cited by 20 publications

(12 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For the validation group, the values are 0.895 and 47.1 × 10 −3 g cm −3 , respectively. The RMSEP obtained are better than those found in studies using hyperspectral imaging (Fernandes et al 2013a;Fernandes et al 2013b). According to Alves et al (2012) the RER should be ≥4 for screening calibration, ≥10 to be acceptable for calibration for quality control, and ≥15 for calibration for quantification.…”

Section: Discussioncontrasting

confidence: 62%

“…Fernandes et al (2013a) used this technology at a spatial resolution of 0.6 mm in a spectral range from 380 to 1028 nm and obtained a coefficient of determination value between the NIR/PLS method and the X-ray method of 0.810 (for close to 17,000 points) with an RMSEP value of 65.4 × 10 −3 g cm −3 . Using neural networks, these values improved to 0.821 and 62.5 × 10 −3 g cm −3 for R 2 and RMSEP, respectively (Fernandes et al 2013b). Kitamura and Tsuchikawa (2015) proposed measuring density profiles using attenuation in transmission of an infrared laser at 830 nm, using an avalanche diode as a detector.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Evaluation of intra-ring wood density profiles using NIRS: comparison with the X-ray method

Baettig

Cornejo

Guajardo

2017

Annals of Forest Science

View full text Add to dashboard Cite

& Key message Pith-to-bark wood density profiling is interesting in forestry science. By comparing it with the X-ray method, this study proved that a fiber optic NIR spectrometer with a high-precision displacement system could accurately measure intra-ring wood density with a spatial resolution of 0.5 mm. & Context Most near-infrared spectroscopy (NIRS) studies for wood density determination use samples that have been pulverized beforehand. Attenuation of ionizing radiation is still the standard method to determine wood density with high spatial resolution. However, there is evidence that NIRS could be an accurate and affordable method for determining intraring density in solid wood strips. & Aims In this study, we research whether the results published for intra-ring density predictions in wood can be improved when calibrated with X-ray microdensitometry. & Methods The measurements were made using a fiber optic probe with a separation between measurement points of 0.508 mm in a range between 1200 and 2200 nm. A total of 4520 density points were used to create partial least squares regression (PLSR). X-ray densitometry data were used as reference values. Twenty PLSR calibrations were randomly executed on 31 samples collected from 28 Pinus radiata D. Don trees. & Results Upon selecting 20 latent variables, the R 2 value was 0.873 for the training group and 0.895 for the validation group, while RMSEP values are 43.1 × 10 −3 and 47.1 × 10 −3 g cm −3 for the training and validation groups, respectively. The range error ratio (RER) was 13.7. & Conclusion The RER was high and almost in the range suggested for quantification purposes. Results are superior to wood density studies in the literature which do not employ spatial resolution and to those found in studies using hyperspectral imaging.

show abstract

Section: Discussioncontrasting

confidence: 62%

Section: Introductionmentioning

confidence: 99%

Evaluation of intra-ring wood density profiles using NIRS: comparison with the X-ray method

Baettig

Cornejo

Guajardo

2017

Annals of Forest Science

View full text Add to dashboard Cite

show abstract

“…The collection of cases presented in [17] shows that in both the fresh and processed food sectors, artificial intelligence and neural networks can be used for classification purposes. However, it is also on the experimental results of several researchers that the main difficulty is the heterogeneity of criteria when determining the classification technique used in one case in particular; that is the reason why studies with [19] have been developed to compare the technique of regression of partial least-squares and more adequate architecture of the network or neural networks neural networks applied in a study in particular. Figure 8 also shows that the error rate can be high, and only for some cases, there is a relatively low misclassification rate compared with the manual procedure.…”

Section: Discussionmentioning

confidence: 99%

“…Although current advances in computer capabilities have allowed the construction of sophisticated machines to replace human actions, there is still a large gap in those activities that require simulation of the five senses, especially when this is aimed at systems of classification and prediction based on pattern recognition [18,19,22,23].…”

Section: Introductionmentioning

confidence: 99%

“…Source: [17] In these systems, images are quantitatively characterized according to morphological characteristics and color parameters, among others; these are used to represent the food and to train the system. Using these researched data previously classified, the researcher trains the data into an intelligent classification system, providing the ability to classify the unknown cases and simulate the way a trained user performs the classification task [17][18][19].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation