Almonds classification using supervised learning methods

2019

Preprint

After harvesting almond crop, accurate measurement of almond kernel sizes is a significant specification to plan, develop and enhance almond processing operations. The size and mass of the individual almond kernels are vital parameters usually associated with almond quality, particularly head almond yield. In this study, we propose a novel methodology that combines image processing and machine-learning ensemble that accurately measures the size and mass of whole raw almond kernels (classification -Nonpareil) simultaneously. We have developed an image-processing algorithm using recursive method to identify the individual almond kernels from an image and estimate the size of the kernels based on the occupied pixels by a kernel. The number of pixels representing an almond kernel was used as its digital fingerprint to predict its size and mass. Various popular machine learning (ML) models were implemented to build a stacked ensemble model (SEM), predicting the mass of the individual almond kernels based on the features derived from the pixels of the individual kernels in the image. The prediction accuracy and robustness of image processing and SEM were analyzed using uncertainty quantification. The mean error in estimating the average length of 1000 almond kernel was 3.12%. Similarly, mean errors associated with predicting the 1000 kernel mass were 0.63%. The developed algorithm in almond imaging in this study can be used to facilitate a rapid almond yield and quality appraisals.

Section: Machine Learning In Almond Researchmentioning

confidence: 99%

Size and mass prediction of almond kernels using machine learning image processing

2019

Preprint

“…Halac, Sokic, and Turajlic () implemented various supervised machine learning methods, specifically multiclass support vector machines, and artificial neural networks to classify different type of almonds. Their method relied on the principal component analysis to identify the most significant shape and color parameters.…”

Section: Introductionmentioning

confidence: 99%

Stack ensembled model to measure size and mass of almond kernels

Vidyarthi

J Food Process Engineering

2020

After harvesting almond crop, accurate measurement of almond kernel sizes is a significant specification to plan, develop and enhance almond processing operations. The size and mass of the individual almond kernels are vital parameters usually associated with almond quality and yield. In this study, we propose a novel methodology that combines image processing and machine‐learning ensemble that accurately measures the size and mass of whole raw almond kernels (classification—Nonpareil) simultaneously. We have developed an image‐processing algorithm using recursive method to identify the individual almond kernels from an image and estimate the size of the kernels based on the occupied pixels by a kernel. The number of pixels representing an almond kernel was used as its digital fingerprint to predict its size and mass. Various popular machine learning (ML) models were implemented to build a stacked ensemble model (SEM), predicting the mass of the individual almond kernels based on the features derived from the pixels of the individual kernels in the image. The prediction accuracy and robustness of image processing and SEM were analyzed using uncertainty quantification. The mean error in estimating the average length of 1,000 almond kernel was 3.12%. Similarly, mean errors associated with predicting the 1,000 kernel mass were 0.63%. The developed algorithm in almond imaging in this study can be used to facilitate a rapid almond yield and quality appraisals.

“…Image processing and analysis is one such technique which has been implemented to perform size measurement of kernels, faster with higher throughput. SHAPE, SeedCount, GrainScan, ImageJ and Smartgrain are few examples of software which were used for image processing and analysis for different grains and nuts (Bejagam, Singh, An, & Deshmukh, 2018;Halac, Sokic, & Turajlic, 2017;Li et al, 2009;Mirzabe, Khazaei, Chegini, & Gholami, 2013;Singh et al, 2020;Tanabata, Shibaya, Hori, Ebana, & Yano, 2012;Teimouri, Omid, Mollazade, & Rajabipour, 2014;USDA-AMS, 2018;Whan et al, 2014;Williams, Munkvold, & Sorrells, 2013). The details of these techniques were well described in our previous publications (Singh et al, 2020;Vidyarthi et al, 2020).…”

Section: Practical Applicationsmentioning

confidence: 99%

Prediction of size and mass of pistachio kernels using random Forest machine learning

Vidyarthi

J Food Process Engineering

et al. 2020

The size and mass of the pistachio kernels are important parameters in pistachio industry worldwide, which are generally associated with pistachio classification and consumer's approval. In this study, we proposed a technique that accurately measures the size and mass of raw pistachio kernels simultaneously using image processing and machine‐learning ensemble. An image‐processing algorithm applying recursive method was used to detect a pistachio kernel from an image and estimate its size based on the pixels occupied. The number of pixels representing a pistachio kernel was used as its digital impression to estimate its size and mass. We employed Random Forest (RF) model to predict the mass of the individual pistachio kernels based on the patterns derived from the pistachio pixels in the image. The mean measured length (18.002 mm) of 100 pistachio kernels were close to the RF predicted length (18.608 mm). Similar close correlation for individual measured pistachio kernel mass and predicted mass was noticed having a bootstrapped residual of 0.036 ± 0.004 g with a class interval of 0.0358–0.0362 at 95% confidence level. Compared to the regression model, bootstrapping and uncertainty quantification showed a good prediction accuracy and robustness of the RF model. Practical Applications The size of pistachio kernels is a crucial factor in defining pistachio grades commercially. Current pistachio grading in industry is still performed manually, which is tedious, time consuming, labor intensive and subject to human error and inconsistency. The developed machine learning algorithm based on Random Forest combined with image processing in this study can be a very useful tool in pistachio industry for rapid classification of pistachio worldwide.