• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1996.06c
• /
• pp.852-862
• /
• 1996

This study was carried out to develop tools to detect defects of apple using machine vision. For the purpose, 6 kinds of frame for color images, R, G, B, h, S, and I frame, and a frame for near infra-red images (NIR frame) were tested first to select one which is useful to segment defect areas from apple images. After then, several methods to classify kind of defect for the segmented defect areas were developed and tested. Five kinds of apple defect -bruise , decay ,fleck worm hole and scar were investigated . The results are as follows: NIR frame was selected as the best one among the 7 kinds of image frame, and R, G and I frames showed favourable result to segment areas of apple defect. Various features of the segmented defect areas were measured to classify the defect areas. Eight kids of feature of the areas-size, roundness, axes length ratio, mean and variance of pixel values, variance of real part of spectrum, mean and variance of power spectrum resulted from spacial ourier transform were observed for the segmented defect areas in the selected 4 frames. then procedures to classify defects using the features were developed for the 4 frames and tested with 75-113 defects on apples. The test resulted that NIR and I frames showed high accuracies to classify the kind of defect as 77% and 76% , respectively.

• Journal of Biosystems Engineering
• /
• v.22 no.2
• /
• pp.217-226
• /
• 1997

This study was to develop a machine vision system to detect and to discriminate 5 kinds of apple surface defectbruise, decay. fleck, worm hole and scar. To detect the defects from an image of apple, thresholding technique was applied to images on various frames (R, G, B, H, S and I) of the color machine vision and an image of near infrared (NIR). To discriminate the detected region of defect, various features of the 5 kind defect regions were extracted from the 4 kinds of images selected above. The features were size of area, roundness, axes length ratio, mean and valiance of pixel values, standard deviation of real part of amplitude spectrum in frequency domain obtained by Fourier transform of pixel data and mean and standard deviation of power spectrum obtained by the same transform of pixel data. Routines to discriminate the defects from the features of image were developed and tested to prove their validity. The test resulted that I-frame and NIR images were the most desirable. Accuracies of the two images to discriminate the defects were noted as 76% and 77%, respectively.

• Journal of Biosystems Engineering
• /
• v.39 no.3
• /
• pp.166-173
• /
• 2014

Purpose: In this study, we focused on the image processing method to determine the external quality of Fuji apples by identifying surface defects such as scabs and bruises. Method: A CCD camera was used to capture filter images with 24 different wavelengths ranging between 530 nm and 1050 nm. Image subtraction and division operations were performed to distinguish the defect area from the normal areas including calyx, stem, and glaring on the apple surface image. All threshold values of the image were examined to reveal the defect area of pretreated filter images. Results: The developed operation methods were [image (720 nm) - image (900 nm)]/image (700 nm) for bruise detection and [image (740 nm) - image (900 nm)]/image (590 nm) for scab detection, which revealed 81% and 90% recognition ratios, respectively. Conclusions: Our results showed several optimal wavelengths and image processing methods to detect Fuji apple surface defects such as bruises and scabs.

• Journal of Biosystems Engineering
• /
• v.36 no.6
• /
• pp.484-490
• /
• 2011

Defects exist underneath the fruit skin are not easily discernable by using conventional color imaging technique in the visible wavelength ranges. Development of sensitive detection methods for the defects is necessary to ensure accurate quality sorting of fruits. Hyperspectral imaging techniques, which combine the features of image and spectroscopy to acquire spatial and spectral information simultaneously, have demonstrated good potentials for identifying and detecting anomalies on biological substances. In this study, a high spatial resolution hyperspectral reflectance technique was presented as a tool for detecting bruises on apple. The two-band ratio (494 nm / 952 nm) and simple threshold methods were applied to investigate the feasibility of discriminating the bruises from sound tissue of apple. The pixel wise accuracy of the discrimination was 74%. The resultant images processed with selected wavebands and morphologic algorithm distinctively showed the early stages of bruises on apple which were not discernable by naked eyes as well as a conventional color camera. Results demonstrated good potential of the hyperspectral reflectance imaging for detection of bruises on apple.

• Smart Media Journal
• /
• v.12 no.9
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.

• Journal of Biosystems Engineering
• /
• v.26 no.2
• /
• pp.169-176
• /
• 2001

Defect of apple was depreciated the product value and causes storage disease seriously. To detect the defect of ‘Fuji’apple with machine vision system, the optical characteristics of defect should be investigated. In this research, absorbance spectra of defect were acquired by spectrophotometer in the range of visible and NIR region(400∼1,100nm) and L*a*b* color values were also acquired by colorimeter. NIR machine vision system was constructed with B&W camera, frame grabber, 16 tungsten-halogen lamps, variable focal length lens and NIR bandpass filter which was mounted to lens outward. Average gray values of defect at 15 NIR wavelength were acquired and the significant NIR wavelength was selected by comparing Mahalanobis distance between sound and defective apple. As the result of Mahalanobis distance analysis, the significant wavelength to discriminate the defectives in ‘Fuji’apple were found to be 720nm for scab and 970nm for bruise and cuts and 920nm was also effective regardless of defective types.

• Korean Journal Plant Pathology
• /
• v.11 no.1
• /
• pp.17-29
• /
• 1995

This study was conducted from 1991 to 1993 to develop a fungicidal spray program for the apples exportable to the United States of America, in which quarantine and pesticide residue in agricultural products are strictly regulated. In 1991, 2 spray schedules were applied to an orchard, in one of which the 7 fungicides registered for apple both in Korea and U.S. were used, and in the other of which the Bordeaux mixture for which the tolerance was exempted in U.S. was used 2 times along with those 7 fungicides. The apple white rot and fruit infection by Alternaria mali were not effectively controlled by the 7 fungicides alone; however, the control efficacy was raised by adding the Bordeaux mixture to the spray schedule. In 1992, 4 spray schedules were applied in which the kinds of fungicides and spray intervals were different one another. The results suggested that an effective spray program can be developed by adopting the Bordeaux mixture for 3 times or adopting the imminoctadine-triacetate which has not tolerance level in U.S. for same times in the vulnerable stage of apple white rot and alternaria blotch. In spite of the high efficacies against major apple diseases, the Bordeaux mixture could not be recommended to the apple growers due to the various defects such as restrictions in compatibility with insecticides and acaricides, troublesomeness in preparation and spray, especially the harmful effect on the finish of Fuji apples. In 1993, a spray program adopting 3 times of imminoctadine-triacetate during the growing season of apple was developed, which not only can effectively control the major apple diseases but also avoid the pesticide residue problems if it was sprayed 2 or 3 times after bagging. On the basis of the 3 years results, a basic fungicide spray program was formulated in which 1~3 times of imminoctadine-triacetate and 1~2 times of bitertanol were adopted in addition of the 7 common fungicides registered in both countries. In the results of application of the spray schedule to the actual farming in the 4 areas of Kyungpook Province in 1994, no noticeable defects were detected at the first year trial. However, this spray program will be continuously evaluated and modified to obtain better control efficacies against major apple diseases.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 1999.12a
• /
• pp.575-580
• /
• 1999

• Korean Journal of Agricultural Science
• /
• v.50 no.3
• /
• pp.373-382
• /
• 2023

The task of sorting and excluding blemished apples and others that lack commercial appeal is currently performed manually by human eye sorting, which not only causes musculoskeletal disorders in workers but also requires a significant amount of time and labor. In this study, an automated apple-sorting machine was developed to prevent musculoskeletal disorders in apple production workers and to streamline the process of sorting blemished and non-marketable apples from the better quality fruit. The apple-sorting machine is composed of an arm-rest, a main body, and a height-adjustable part, and uses object detection through a machine learning technology called 'You Only Look Once (YOLO)' to sort the apples. The machine was initially trained using apple image data, RoboFlow, and Google Colab, and the resulting images were analyzed using Jetson Nano. An algorithm was developed to link the Jetson Nano outputs and the conveyor belt to classify the analyzed apple images. This apple-sorting machine can immediately sort and exclude apples with surface defects, thereby reducing the time needed to sort the fruit and, accordingly, achieving cuts in labor costs. Furthermore, the apple-sorting machine can produce uniform quality sorting with a high level of accuracy compared with the subjective judgment of manual sorting by eye. This is expected to improve the productivity of apple growing operations and increase profitability.

• Journal of Biosystems Engineering
• /
• v.39 no.2
• /
• pp.142-149
• /
• 2014

Purpose: A multispectral algorithm for detection and differentiation of defective (defects on apple skin) and normal Red Delicious apples was developed from analysis of a series of hyperspectral line-scan images. Methods: A fast line-scan hyperspectral imaging system mounted on a conventional apple sorting machine was used to capture hyperspectral images of apples moving approximately 4 apples per second on a conveyor belt. The detection algorithm included an apple segmentation method and a threshold function, and was developed using three wavebands at 676 nm, 714 nm and 779 nm. The algorithm was executed on line-by-line image analysis, simulating online real-time line-scan imaging inspection during fruit processing. Results: The rapid multispectral algorithm detected over 95% of defective apples and 91% of normal apples investigated. Conclusions: The multispectral defect detection algorithm can potentially be used in commercial apple processing lines.