Fruit Sorting Research Articles

Segmentation using fuzzy cluster‐based thresholding method for apple fruit sorting

Apple fruit sorting has been an important postharvest process carried on for the sorting of diseased apple fruits. A fuzzy cluster-based thresholding (FCBT) method for segmenting the region of interest from an apple image has been proposed for sorting apples in this study. As the first step, the acquired RGB colour image of an apple fruit was converted into a greyscale image. Then, five different fuzzy cluster bins with overlapped pixel ranges were taken and greypixel values were binned into them. A cluster with the maximum number of pixels was selected for calculating the threshold value. The region of interest from the apple image was then segmented using the proposed FCBT value. Features extracted from the segmented images were given as input to a fully-connected deep neural network for a classification. The performance of the FCBT method was compared with similar greyscale thresholding methods like Otsu's and Kapur's methods. The visual segmentation accuracy and the execution speed showed that the FCBT outperformed the other methods in segmenting the diseased area. A fully-connected deep neural network model with the FCBT image extracted features as input values gave a 98.33% accuracy rate in sorting the apple images.

A hybrid deep learning-based fruit classification using attention model and convolution autoencoder

Image recognition supports several applications, for instance, facial recognition, image classification, and achieving accurate fruit and vegetable classification is very important in fresh supply chain, factories, supermarkets, and other fields. In this paper, we develop a hybrid deep learning-based fruit image classification framework, named attention-based densely connected convolutional networks with convolution autoencoder (CAE-ADN), which uses a convolution autoencoder to pre-train the images and uses an attention-based DenseNet to extract the features of image. In the first part of the framework, an unsupervised method with a set of images is applied to pre-train the greedy layer-wised CAE. We use CAE structure to initialize a set of weights and bias of ADN. In the second part of the framework, the supervised ADN with the ground truth is implemented. The final part of the framework makes a prediction of the category of fruits. We use two fruit datasets to test the effectiveness of the model, experimental results show the effectiveness of the framework, and the framework can improve the efficiency of fruit sorting, which can reduce costs of fresh supply chain, factories, supermarkets, etc.

PENERAPAN REKAYASA MESIN SORTIR SEBAGAI PENENTU KEMATANGAN BUAH JERUK DAN TOMAT MERAH BERBASIS IMAGE PROCESSING

The embedded systems in the industrial, especially image processing, is increasingly leading to the study of production automation systems such as fruit sorting. Post-harvest sorting system implemented by the industry is manual, so it’s not effective. The solution was to conduct research aimed at modifying post-harvest sorting tools by engineering tomato and orange sorting machines based on their color. The method uses image processing. It’s the most efficient alternative in terms of cost and complexity of hardware design, does not require many sensors, but produces an accurate output. The camera is placed on the mechanical sorting machine system, taking images to determine the sorting execution after the fruit color type are recognized. The results of the research were carried out through several tests, namely: light intensity, color image data, and organoleptics. Light intensity test showed that the position of the tool had a value of 0.78% of the outside light disturbance. Color image shows the range of ripeness values (R/G) for raw tomatoes 0<=1.04; half ripe tomatoes 1.04<=1.39; ripe tomatoes 1.39<=3.59; raw orange 0<=0.92; undercooked oranges 0.92<=0.98; and ripe oranges 0.98<=1.66. Organoleptic test from five observers had the same results as the reading on the fruit sorting tool. Keywords : engineering, fruit maturity, oranges, sorting machines, tomatoes

Vision based volume estimation method for automatic mango grading system

In horticulture, measuring, sorting by shape, and determining the size and volume of fruits are all essential processes for meeting market quality standards and increasing market value. Fruit sorting and grading processes are very laborious and time-consuming task but machine vision-based fruit grading systems have the potential to replace human labour. However, a great challenge in vision-based fruit grading system is the recognition of different features such as shape, size, skin flaw and sometimes even three-dimensional (3D) shape. In this study, a simple and efficient image processing algorithm is proposed for estimating volume and 3D shape of mango fruit. The width and length of mango fruit are obtained from two-dimensional (2D) colour image. Then, fruit thickness is estimated based on light intensity distribution in 2D (top view) of mango fruit and maximum width–thickness correlation. The 3D shape of the mango fruit is then reconstructed. The accuracy of proposed method was compare with two existing volume prediction methods. Estimated volumes were compared with measured measurements using water-displacement method and the reconstructed 3D shapes were compared with measured structures of mango fruits. For a total of 150 mango samples, the results show that the proposed method gave an accuracy of 96.8% whilst the two other methods gave 91.7% and 91.5% respectively. The reconstructed mango shapes were therefore in close agreement with measured shapes.

Fruit sorting robot based on color and size for an agricultural product packaging system

Indonesia's location in the equator gives an ideal condition for agriculture. However, agriculture suffers the issue of old farming due to a lack of youth interest working in this sector. This problem can be overcome by applying digital farming methods, in which one of them is by employing robots. Robotics technology is suitable for handling the harvested product, such as a sorting robot. This paper presents the application of a 4DOF fruit sorting robot based on color and size in a packaging system. The sorting is made possible by image processing where color is recognized by HSV analysis, and the diameter is known in the grayscale image and setting the thresholding. The fruit to be sorted is red and green tomatoes and red and green grapes. The experiments were conducted to show the effectiveness of the proposed method. The time requires for the robot to accomplish the task is 11.91s for red tomatoes, 11.76s for green tomatoes, 12.56s for red grapes, and 12.92s for green grapes. The time difference is due to the position of the boxes for the sorted fruit. The experimental results show that the arm robot manipulator is applicable for a sorting robot using the proposed method.

Estimasi Volume Buah Kiwi Menggunakan Metode Pencitraan dan Aturan Simpson

Volume is one of important quantities that have been applied to fruit sorting based on size. Imaging method or computer vision is a simple non destructive method that has been proposed to measure fruits volume. This study was aimed to estimate the volumes of kiwi fruits using Computer Vision imaging method and compared to a water displacement method. The samples were 20 green kiwi fruits (Actinidia deliciosa). A smartphone camera was used to record the kiwifruit images and Python based program to drive the camera and process the images. Images resulted in Computer Vision are two dimensions (2D) images. The 1/3 rd Simpson rule was employed to determine the volume of kiwi fruits based on the volume integration of a spinning object where surface image of kiwi was divided into 8 parts and then summed. The results show that the 2D imaging method assisted by the Simpson rule was successfully able to determine the kiwi fruit volumes with 4.57 % average difference percentage compared to the water displacement method. This was about 4.97 cm3 of average volume difference of 20 samples. The sample volumes measured using this method ranges from 82,48 cm3 - 126,85 cm3. These results will be one of steps toward the development of machine vision for fruit sorter based on volume

Apple fruit sorting using novel thresholding and area calculation algorithms

The speed and capacity (kg/hr) for sorting apples are two important factors that have to be taken care of post-harvest. To increase the speed and capacity of apples being sorted, a new algorithm for segmenting the region of interest from an apple fruit image and to find the percentage of defective area after sorting process has been proposed in this paper. First, the region of interest from the acquired color image of an apple fruit was segmented using the proposed global thresholding algorithm GTA. Features were extracted from the segmented image using the coefficients obtained by applying wavelet transformation using the Haar filter. Further, the Naive Bayesian classifier was applied to sort apples as defective and sound. Second, the defective area of the defective sample was segmented using the K-means algorithm and median filter. The defective area in percentage was calculated using the proposed diseased area calculation algorithm DACA to decide about the acceptance of that defective sample. The performance of the proposed GTA was compared with Otsu’s and Kapur’s thresholding algorithms. With less segmentation time, it was found that the Naive Bayesian classification using GTA gave a 96.67% accuracy rate than Otsu’s with 65% and Kapur’s with 93.33% accuracy rate. The execution time of the proposed DACA was less when compared to the bwarea function in MATLAB.

Technological Methods of Assessing the Maturity of Tomatoes

Tomato maturity is one of the most important indicators related to determining its quality. The maturity assessment of tomatoes is necessary to determine the timing of fruit harvesting, optimize storage conditions, forecast storage periods, export, and more. Over the past three decades, researchers have developed a number of methods for evaluating the quality and sorting of tomato fruits. (Research purpose) The research purpose is in providing an overview of the technological methods used to measure the maturity of tomatoes. (Materials and methods) The article presents methods (spectral, hyperspectral imaging, luminescent, laser, technical vision, and others) for non-destructive assessment of the quality of tomato fruits. (Results and discussion) It was found that the main drawback of the described devices is that only a small part of the fetal surface is controlled by photodetectors. It was pointed out that most of the flow reflected from the fruit, and therefore information about the color of its surface is lost, bypassing the photodetectors and being absorbed by the walls of the camera. Authors found that the quality of fraction selection can be improved if several spectral sections are filtered out from the rays reflected from the fruit. It was found that with the three or more sections, designs of photometric systems are significantly complicated, since there is a need to have a large number of optical receivers. (Conclusions) The article proposes a technology for measuring the maturity of tomatoes using the luminescent method, which is less time-consuming and more accurate than methods for determining the maturity of tomatoes by their color. Authors found that this technology allows to separate tomatoes with any color according to the maturity. The article shows that the considered method makes it possible to divide tomatoes more efficiently by maturity, taking into account their internal structure.

Manipulation of Fruit Dry Matter via Seasonal Pruning and Its Relationship to d’Anjou Pear Yield and Fruit Quality

Orchard-side optimization of fruit quality is experiencing renewed research focus in the fresh fruit industry as new technologies and quality metrics have emerged to enhance consumer acceptance and satisfaction. Fruit dry matter, one such quality index gaining traction among numerous fresh fruit commodities, was targeted for improvement in d’Anjou pear with the application of seasonal pruning cycles (fall, fall and summer, winter, and winter and summer) across two growing seasons in 2016 and 2017 in a mid-aged, traditionally managed commercial orchard in the Columbia basin, Washington, USA. Dry matter was assessed non-destructively on pears using near-infrared spectroscopy at harvest and fruit categorized in to low (<13%), moderate (13–16%), and high (>16%) dry matter quality categories, revealing that fall pruning positively impacted average predicted fruit dry matter in comparison to winter pruning (15.1 vs. 14.2% in 2016 and 13.7 vs. 13.1% predicted dry matter in 2017 for winter vs. fall pruning, respectively), as well in the abundance of high dry matter fruits. The addition of summer pruning to either fall or winter pruning increased fruit size by up to 13% of proportion of fruits 80 mm or greater in diameter. Further, a tendency for summer pruning to decrease yield (up to nearly 30 kg/tree lower yields), average fruit dry matter (up to 0.5% lower average predicted dry matter), and abundance of high dry matter fruits (up to 11% fewer high predicted dry matter fruits) was observed. Fruit quality classes assembled on predicted dry matter verified the utility of this emerging parameter as a fruit quality metric for pears as demonstrated by more desirable post-harvest eating characteristics such as higher soluble solids content corresponding to greater at-harvest predicted dry matter categories. Targeted seasonal pruning in association with precise at-harvest dry matter fruit sorting may preserve the profitability of pear cultivation through their impact on fruit quality and associated consumer experiences.

Mobile Application for Identification of Coffee Fruit Maturity using Digital Image Processing

Indonesia is an agricultural country that relies on the agricultural sector and is well known in producing various plantation commodities, one of which is coffee. Coffee is a leading export commodity developed in Indonesia. Community coffee plantations play an important role because most of the coffee production comes from community plantations. However, the condition of community coffee plantations can be said to be still hampered, due to the quality of coffee is still relatively low. It is caused by coffee fruit sorting, which is still done conventionally. The conventional sorting process of coffee fruits is still carried out with the help of operator knowledge, so the level of operator knowledge dramatically influences the results of sorting. The ease of sorting coffee ripeness can be done by implementing a mobile application using digital image processing. Techniques used in digital image processing are the HSV color space to get color features of coffee fruit and the K-Nearest Neighbor (KNN) classification method to classify coffee fruit ripeness. The results of the identification are in the form of ripe, half-ripe, or unripe fruits. The mobile application of this research has two main features, namely training data feature and non real-time identification feature. The results of the testing conducted resulted in an accuracy rate of 95.56% with the best membership value (k) of 3.

Modelling the damping response of biomimetic foams based on pomelo fruit

Some sorts of fruits, like pomelo (Citrus maxima or Citrus grandis), also spelled pummelo, are of special interest for researchers and engineers during the development of impact resistant structures. It is assumed that the interaction of structural features on different length-scales of such fruits enables the dissipation of large amounts of energy during their impact against a hard surface. Accordingly, the aim of this paper is to perform an analysis by using a numerical model at different hierarchical levels by means of Finite Element Methods (FEM) in order to identify different structural features that contribute to the damping performance of the pomelo fruit. The considered numerical model was created on the basis of homogeneous aluminum (AlSi7Mg0.3) foam structure, inspired by the pomelo fruit shell structure. In the present approach, the Si- or Fe-rich intermetallic particles on the struts of AlSi7Mg0.3 foam are considered additionally, this enabled more detailed investigations on the structural behavior of the foam during different loading conditions. Comparison of the results from simulation and experimental compression tests showed promising results with respect to the deformation behavior, which are offering support to the design of biomimetic metallic foams. Understanding the principles of combining the structure and material inspired by biological systems enables constructing new lightweight bio-inspired materials of high impact and puncture resistance with a combination of high-energy dissipation, high damping properties and a significant recovery from large deformations.

A Review of Convolutional Neural Network Applied to Fruit Image Processing

Agriculture has always been an important economic and social sector for humans. Fruit production is especially essential, with a great demand from all households. Therefore, the use of innovative technologies is of vital importance for the agri-food sector. Currently artificial intelligence is one very important technological tool widely used in modern society. Particularly, Deep Learning (DL) has several applications due to its ability to learn robust representations from images. Convolutional Neural Networks (CNN) is the main DL architecture for image classification. Based on the great attention that CNNs have had in the last years, we present a review of the use of CNN applied to different automatic processing tasks of fruit images: classification, quality control, and detection. We observe that in the last two years (2019–2020), the use of CNN for fruit recognition has greatly increased obtaining excellent results, either by using new models or with pre-trained networks for transfer learning. It is worth noting that different types of images are used in datasets according to the task performed. Besides, this article presents the fundamentals, tools, and two examples of the use of CNNs for fruit sorting and quality control.

Technical vision system for quality control of objects of the ball-shaped form when sorting on the conveyor

A functional diagram of a robotic complex mock-up for rejecting ball-shaped objects transported on a roller conveyor is presented. The software of the complex is designed to detect an object on the conveyor and determine its coordinates. To detect objects of control, their images obtained in the visible range of the optical spectrum and Viola–Jones object detection algorithms are employed. The developed software is based on a trained cascade classifier, the optimal settings of which are determined. To detect surface defects of objects, their spectrograms obtained in the range of 400… 1000 nm are used. The presented results can be applied to the process automation of sorting fruits and vegetables and other ball-shaped objects.

On line detection of defective apples using computer vision system combined with deep learning methods

A deep-learning architecture based on Convolutional Neural Networks (CNN) and a cost-effective computer vision module were used to detect defective apples on a four-line fruit sorting machine at a speed of 5 fruits/s. A CNN based classification architecture was trained and tested, with the accuracy, recall, and specificity of 96.5%, 100.0%, and 92.9%, respectively, for the testing set. An inferior performance was obtained by a traditional image processing method based on candidate defective regions counting and a support vector machine (SVM) classifier, with the accuracy, recall, and specificity of 87.1%, 90.9%, and 83.3%, respectively. The CNN-based model was loaded into the custom software to validate its performance using independent 200 apples, obtaining an accuracy of 92% with a processing time below 72 ms for six images of an apple fruit. The overall results indicated that the proposed CNN-based classification model had great potential to be implemented in commercial packing line.

Indonesian fruits classification from image using MPEG‐7 descriptors and ensemble of simple classifiers

AbstractFruits classification from image is a very challenging task, particularly for Indonesian indigenous fruits, due to some similarities occurred in several types of the fruits. This study proposes a method to classify Indonesian fruits from image using MPEG‐7 color and texture descriptors. The descriptors were directly extracted from the image without pre‐processing and segmentation steps. Principle component analysis was then applied to reduce the dimension of the descriptors. Four simple classifiers, decision tree, naïve Bayesian, linear discriminant analysis, and k‐nearest neighbor were used to classify the fruit image based on extracted descriptors. An ensemble of simple classifiers trained with some combination of MPEG‐7 descriptors has been constructed to increase the classification accuracy of single simple classifier. To validate the proposed method, an Indonesian fruit images data set consisted of 15 classes was developed in this study. The experiment result showed that the ensemble of simple classifiers achieved the best accuracy of 97.80% by employing linear discriminant analysis, and k‐nearest neighbor as base classifiers trained using CSD, SCD, and the combination of CLD and EHD. Therefore, the proposed method achieved a good classification accuracy and can be applied in vision‐based classification system in industry.Practical ApplicationsThis study proposes a method to classify Indonesian fruits from image using MPEG‐7 descriptors and the ensemble of simple classifiers. The proposed method can be applied in vision‐based fruit sorting system in fruit industry as well as vision‐based fruit pricing system in supermarket.

Hand–eye calibration and grasping pose calculation with motion error compensation and vertical-component correction for 4-R(2-SS) parallel robot

Due to motion constraint of 4-R(2-SS) parallel robot, it is difficult to calculate the translation component of hand–eye calibration based on the existing model solving method accurately. Additionally, the camera calibration error, robot motion error, and invalid calibration motion poses make it difficult to achieve fast and accurate online hand–eye calibration. Therefore, we propose a hand–eye calibration method with motion error compensation and vertical-component correction for 4-R(2-SS) parallel robot by improving the existing eye-to-hand model and solving method. Firstly, the eye-to-hand model of single camera is improved and the robot motion error in the improved model is compensated to reduce the influence of camera calibration error and robot motion error on model accuracy. Secondly, the vertical-component of hand–eye calibration is corrected based on vertical constraint between calibration plate and end effector in parallel robot to calculate the pose and motion error in calibration of 4-R(2-SS) parallel robot accurately. Thirdly, the nontrivial solution constraint of eye-to-hand model is constructed and adopted to remove invalid calibration motion poses and plan calibration motion. Finally, the proposed method was verified by experiments with a fruit sorting system based on 4-R(2-SS) parallel robot. Compared with random motion, the existing model, and solving method, the average time of online calibration based on planned motion decreases by 29.773 s and the average error of calibration based on the improved model and solving method decreases by 151.293. The proposed method can improve the accuracy and efficiency of hand–eye calibration of 4-R(2-SS) parallel robot effectively and further realize accurate and fast grasping.

Design and Construction of the Oil Palm Fruit Sorting Device Based on Color with PC Display

This research discusses the design and construction of an oil palm fruit sorting tool and a counter for the number of oil palm fruits that work automatically using the working principle of a color sensor. This tool can work automatically because it is controlled by an ATmega328 microcontroller. Overall this system consists of designing a color sensor, motor controller, conveyor, and power supply. Item detection consists of 328 microcontroller technology, color sensor, proximity sensor, LCD, IOT, and PC display. The testing stages of this tool include testing the sensor circuit, testing the servo motor circuit, and testing the automatic oil palm fruit sorting and counting tool. When the oil palm fruit is detected, the color sensor will take data. The captured data will be sent to the microcontroller and displayed to the LCD, IOT and PC. There are additional devices in the form of undercooked, ripe and past ripe separation bars by utilizing a DC servo motor.ATmega328

Physical, thermal, and mechanical properties of bael fruit

AbstractThe physical, thermal, and mechanical properties of bael fruit were discussed in this paper. The properties of bael fruit were studied in term of its dimensional, gravimetric, frictional properties, and color values, whereas thermal properties were evaluated by determining the glass transition temperature of bael fruit pulp powder. The effect of the direction of orientation and temperature treatment on the mechanical properties of bael fruit were studied. The sphericity of bael fruit is considered to be spherical. The average length, width, and thickness of bael fruit were 8.92, 8.22, and 7.95 cm, respectively. The bulk density, true density, and porosity were found to be 0.470, 1.067, and 55.96 g/cm3, respectively. The specific heat of bael fruit pulp was 3.2512 kJ/kg °K. The glass transition temperature of the pulp powder was 277.15 °K, with a specific heat of 0.429 J/g °K. The orientation of bael fruit along the thickness required minimum rupture force than orientation along the length and width. The minimum rupture force required to puncture bael fruit along its thickness was 431.14 KN for freezing treatment. The results obtained from this study would help in manufacturing the handling and processing machineries for the commercial production of bael fruit products.Practical ApplicationsBael is an underprocessed indigenous fruit despite its high nutritional and medicinal properties. Knowledge about the physical, thermal, and mechanical properties of bael fruit will encourage researchers to develop and modernize the design of equipment used for sorting, grading, puncturing, pulping, conveying, packaging, transporting, and storing of fruits.

Orientation detection of fruits by means of convolutional neural networks and laser line projection for the automation of fruit packing systems

Sorting and quality inspection of fruits is still a time-consuming task because of the low level of automation. Natural products exhibit a greater variance compared to manufactured products regarding quality and dimensions, which complicates their automated processing. In our paper, we propose a method to determine the orientation of fruits by applying a specially trained Convolutional Neural Network to 2.5-dimensional image data generated by a combination of a color CMOS camera and a red line laser. We achieve an accuracy of the orientation prediction of approximately 5 degrees, which presents very promising results for the automation of fruit packing systems.

SORTING MANALAGI APPLE (Malus sylvetris Mill) USING IMAGE PROCESSING APPLICATION

Manalagi apple (Malus sylvestris Mill) is one type of apple that has been widely known by the public in Indonesia. Fruit sorting in Indonesia, particularly Manalagi apple, is carried out manually so it is still at a disadvantage. This study aimed to arrange image processing applications with criteria for size, width, height, color, and defect area of fruits. The samples of Manalagi apple used in this study were 130 with 50 each in the non reject and reject quality classes and 30 for the validation test. Manalagi apple samples were obtained from one of the collectors located in Batu City. The image of Manalagi apple processed gained several variables, including perimeter, area, height, width, r, g, b, and defect area. Based on statistical analysis, image quality variables that could be used as input for making applications in the form of logical sentences were perimeter, area, and defect area. Based on the application, the validation test obtained an accuracy of 86.65%.