Size Classification of Tomato Fruit Using Thresholding, Machine Learning, and Deep Learning Techniques

Abstract

The size of tomato fruits is closely related to the market segment and price. Manual sorting in tomato is very dependent on human interpretation and thus, very prone to error. The study presents thresholding, machine learning and deep learning techniques in classifying the tomato as small, medium and large based from a single tomato fruit image implemented using Open CV libraries and Python programming. Tomato images with different sizes are gathered where features like area, perimeter and enclosed circle radius are extracted. The experiment shows that using thresholding, a classification accuracy of 85.83%, 65.83% and 80% was achieved for area, perimeter and enclosed circle radius, respectively. For machine learning, the training accuracy rates were recorded as 94.00%-95.00% for SVM, 97.50-92.50% for KNN and 90.33-92.50% for ANN. Comparison of models revealed that SVM is the most model without over fitting. The deep learning approach, regardless of the algorithm, produced low performances with 82.31%-78.21%-55.97% trainingvalidation-testing accuracy for VGG16, 48.17%-41.44%-37.64% for InceptionV3 and 56.05%-44.96%-22.78% for ResNet50 models. Comparative analysis showed that machine learning technique bested the performance of the thresholding and deep learning techniques in classifying the tomato fruit size in terms of accuracy performance.

Hybrid Genetic Programming and Multiverse-based Optimization of Pre-Harvest Growth Factors of Aquaponic Lettuce Based on Chlorophyll Concentration

“…NPK, iron, and salinity concentration were previously optimized based on germination rate and root morphology for lettuce and rice [6]- [8]. Crops, such as corn, rice, tomato, and lettuce, are vulnerable to climate change resulting in lower biomass production [5], [9], [10]. The initial tendency to improve crop quality is through adjusting the irrigation and fertigation [11], air velocity [12], and spectral filters [13].…”

“…Current optimization approaches make use of crop features such as fresh weight and leaf shape index [29], dry biomass [30], carotenoid [31], stomatal conductance, size, and density [32], leaf starch content [24], root anatomy [33], and vitamin C and nitrate content [15]. Image phenotyping has been made possible through computer vision (CV) [5], [9]. The use of adaptive neuro-fuzzy inference systems (ANFIS) [1] and genetic algorithm (GA) [2] have been proven effective in optimizing nonlinear relationships of environmental stressors to canopy areas.…”

Lettuce Canopy Area Measurement Using Static Supervised Neural Networks Based on Numerical Image Textural Feature Analysis of Haralick and Gray Level Co-Occurrence Matrixs

“…Cultivation of vegetable crops inside a controlled environment chamber and extracting leaf canopy signatures that holds essential information for crop phenotyping has captivated agriculturist and scientist for quite a few generations (Burgos-Artizzu, Ribeiro, Guijarro, & Pajares, 2011;Calangian et al, 2018;de Luna, Dadios, Bandala, & Vicerra, 2019;Hang, Lu, Takagaki, & Mao, 2019;Loresco, Valenzuela, Culaba, & Dadios, 2019;Zou et al, 2019). In this study, it is shown that indoor hydroponic lettuce canopy area can be measured based on numerical image textural feature analysis of Haralick and gray level co-occurrence matrix (Table 1) as compared with morphological pixel feature (Calangian et al, 2018), leaf shape (Saleem, Akhtar, Ahmed, & Qureshi, 2019) and point cloud analysis (Berk, Stajnko, Belsak, & Hocevar, 2020).…”

“…Broader leaf corresponds to maturity as part of its phenological development. Several studies have already developed a non-destructive approach on estimating individual leaf area of carrot (Haug & Ostermann, 2015), cauliflower (Hamuda, Mc Ginley, Glavin, & Jones, 2017), cucumber (Xie et al, 2014), lettuce (Hernández-Hernández et al, 2016;Tian & Wang, 2009;Zou et al, 2019), maize (Burgos-Artizzu, Ribeiro, Guijarro, & Pajares, 2011), potato (Boyd, Gordon, & Martin, 2002), radish (Dang et al, 2018), sugar beet (Chebrolu et al, 2017) and tomato (Boulard, Roy, Pouillard, Fatnassi, & Grisey, 2017;de Luna, Dadios, Bandala, & Vicerra, 2019) considering the length and width of the leaf shape (Saleem, Akhtar, Ahmed, & Qureshi, 2019;Wang, Jin, Shi, & Liu, 2019).…”

FGrade: A Large Volume Dataset for Grading Tomato Freshness Quality

“…Opena et al [22] proposed an automated tomato classification system that used an artificial neural network (ANN) classifier and the artificial bee colony (ABC) algorithm used for training the model. Luna et al [23] proposed a classifier that is used to classify fruits in different classes based on size of images by thresholding, machine learning and deep learning models. Semary et al [24] proposed a method to classify infected fruits based on its external surface.…”