• 스마트미디어저널
• /
• 제12권9호
• /
• 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.

• 농업생명과학연구
• /
• 제44권5호
• /
• pp.75-79
• /
• 2010

사과는 전 세계적으로 대표적 과수의 하나로서 우량 사과의 생산을 위하여 신속하고 경제적이며 정확한 사과바이러스 진단이 요구되고 있다. RT-PCR은 사과바이러스 진단을 위한 중요한 기술로서 우선 시료조직의 분쇄 및 균질화를 통한 양질의 RNA 추출이 필수적이다. 그러나 분쇄작업은 다량의 시료의 경우 많은 시간과 노동이 요구된다. 본 연구에서는 조직 분쇄과정이 없이 단순 가열에 의한 RNA 추출을 시도하였으며 줄기조직이 잎조직보다 약간 더 적합함을 보여주었다. 그러나 RT-PCR에 의한 사과바이러스 진단에서는 모두 동일한 결과를 나타냈다. 이로써 사과 조직에 대한 단순가열로써 매우 간편하게 양질의 RNA추출이 가능함을 제시하였다.

• The Plant Pathology Journal
• /
• 제37권6호
• /
• pp.543-554
• /
• 2021

To preventively control fire blight in apple trees and determine policies regarding field monitoring, the Maryblyt ver. 7.1 model (MARYBLYT) was evaluated in the cities of Chungju, Jecheon, and Eumseong in Korea from 2015 to 2020. The number of blossom infection alerts was the highest in 2020 and the lowest in 2017 and 2018. And the common feature of MARYBLYT blossom infection risks during the flowering period was that the time of BIR-High or BIR-Infection alerts was the same regardless of location. The flowering periods of the trees required to operate the model varied according to the year and geographic location. The model predicts the risk of "Infection" during the flowering periods, and recommends the appropriate times to control blossom infection. In 2020, when flower blight was severe, the difference between the expected date of blossom blight symptoms presented by MARYBLYT and the date of actual symptom detection was only 1-3 days, implying that MARYBLYT is highly accurate. As the model was originally developed based on data obtained from the eastern region of the United States, which has a climate similar to that of Korea, this model can be used in Korea. To improve field utilization, however, the entire flowering period of multiple apple varieties needs to be considered when the model is applied. MARYBLYT is believed to be a useful tool for determining when to control and monitor apple cultivation areas that suffer from serious fire blight problems.

• The Plant Pathology Journal
• /
• 제39권4호
• /
• pp.319-334
• /
• 2023

Plant disease is an important factor affecting crop yield. With various types and complex conditions, plant diseases cause serious economic losses, as well as modern agriculture constraints. Hence, rapid, accurate, and early identification of crop diseases is of great significance. Recent developments in deep learning, especially convolutional neural network (CNN), have shown impressive performance in plant disease classification. However, most of the existing datasets for plant disease classification are a single background environment rather than a real field environment. In addition, the classification can only obtain the category of a single disease and fail to obtain the location of multiple different diseases, which limits the practical application. Therefore, the object detection method based on CNN can overcome these shortcomings and has broad application prospects. In this study, an annotated apple leaf disease dataset in a real field environment was first constructed to compensate for the lack of existing datasets. Moreover, the Faster R-CNN and YOLOv3 architectures were trained to detect apple leaf diseases in our dataset. Finally, comparative experiments were conducted and a variety of evaluation indicators were analyzed. The experimental results demonstrate that deep learning algorithms represented by YOLOv3 and Faster R-CNN are feasible for plant disease detection and have their own strong points and weaknesses.

• Journal of Plant Biotechnology
• /
• 제46권3호
• /
• pp.228-235
• /
• 2019

사과(Malus pumila)는 국내에서 가장 경제적으로 중요한 과수 중의 하나이다. 하지만 사과 바이러스 감염은 생산량을 감소시키고 수확량 손실과 과일 품질 저하와 같은 심각한 문제를 야기한다. 국내에 감염된 사과 바이러스 및 비로이드 종류는 Apple chlorotic leaf spot virus (ACLSV), Apple stem pitting virus (ASPV), Apple stem grooving virus (ASGV), Apple mosaic virus (ApMV)와 Apple scar skin viroid (ASSVd) 등이 알려져 있다. 사과는 바이러스나 비로이드에 감염되어 있어도 대체로 이상한 징후가 발견되지 않아 바이러스로 인해 피해가 많았다. 본 연구는 사과 왜성대목 M.9 및 M.26의 무독묘 생산을 위하여 고온처리($37^{\circ}C$, 6주), 화학처리(Ribavirin) 및 생장점 배양하여 바이러스 제거 처리를 하였다. 바이러스 검출에 일반적으로 사용되는 방법은 효소면역 측정법(ELlSA)과 중합효소연쇄반응(RT-PCR)을 이용하였는데, RT-PCR은 ELlSA방법보다 10 ~ 30% 더 민감하였다. 사과 왜성대목 바이러스 검정 결과, 바이러스 제거 효율은 생장점 배양이 가장 높았다. 생장점 배양 후 바이러스 무병묘의 획득율은 30 ~ 40%로 높게 나타났다. 생장점 배양에서 사과 왜성대목 M.9은 ACLSV, ASPV 및 ASGV의 비율이 각각 45%, 60%, 50%로 높았고, 사과 왜성대목 M.26에서는 ACLSV, ASPV 및 ASGV의 감염율은 각각 40%, 55%, 55%였다. 이상의 결과, 사과 왜성대목에서 무독묘를 생산할 수 있는 가장 효과적인 방법은 생장점 배양에 의한 것으로 판단되었다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1985년도 하계학술회의논문집
• /
• pp.85-88
• /
• 1985

We have a real-time algorithm which improves some drawbacks in the existed method for detection of the QRS complex waves. This proposed algorithm is programmed with 6502 assembly language based-on Apple II microcomputer.

• 한국균학회지
• /
• 제28권1호
• /
• pp.49-54
• /
• 2000

1. 저장 중 Penicillium에 이병된 사과, 배, 귤, 포도를 TLC 및 HPLC로 주요독소들을 검정한 결과 patulin만 검출되었고, citrinin은 검출되지 않았다. patulin 독소는 사과에서 $5.68\;{\mu}46.81\;{\mu}g/g$, 배에서 $3.48{\sim}84.71\;{\mu}g/g$, 귤에서 $0.16{\sim}0.27\;{\mu}g/g$검출되었으나 포도에서는 전혀 검출되지 않았다. 2. 과실류 Penicillium 저장 중 억제효과를 검정한 결과 사과, 배, 귤에 대하여 sodium hypochloride gas처리나 열처리$(37^{\circ}C)$에 의하여 방제효과가 켰으며 특히 열처리$(37^{\circ}C)$는 100%의 방제효과를 나타냈다. 처리에 의한 장해를 보면 열처리$(37^{\circ}C)$에서는 장해가 발생하지 많았으며 sodium hypochloride gas 처리에는 배, 귤에서 장해가 나타났으나 사과에서는 발생하지 않았다.

• 수산해양기술연구
• /
• 제24권1호
• /
• pp.1-6
• /
• 1988

For the purpose of making the detection range of fish detection system more longer and computerizing the system a parametric sound source, a timer and a digitizing circuit for the Apple II computer have been studied. The parametric sound of 5 KHz generated by passing AND gate two signals from carrier signal generator of 200KHz with modulator of 5KHz. This parametric acoustic source of 5KHz difference frequency had more higher directional resolution of 10 degrees than single frequency sound of 200KHz. Peripheral interface adaptor MC 6821 was adopted for interfacing to the Apple II personal computer. The timer consisted of six decade binary coded decimal counters (74 LS 190), and the digitizing circuit consisted of a sample and hold (LF 398) and an A/D converter(ADC 0808). The timer with 10KHz clock pulse had the measuring time from 0.1msec to 100sec. This time measuring range was satisfactory for the aim of the fish finding acoustic system.

• The Plant Pathology Journal
• /
• 제27권1호
• /
• pp.44-52
• /
• 2011

Field surveys were conducted in 45 stone fruit orchards in seven districts of Isparta Province located in western Mediterranean region of Turkey important for stone fruit production. Leaf samples were collected from 175 trees showing virus-like symptoms. These samples were first tested by ELISA for five different RNA viruses including Apple mosaic ilarvirus (ApMV), Prunus necrotic ringspot ilarvirus (PNRSV), Prune dwarf ilarvirus (PDV), Plum pox potyvirus (PPV), Apple chlorotic leafspot trichovirus (ACLSV). While no ApMV and PPV infection was found, 46, 24 and 16 samples were tested positive for PDV, ACLSV and PNRSV, respectively, in ELISA showing about 45% of symptomatic trees in the region were infected with at least one of these viruses. In addition, it was found that nine sweet cherry trees were mixed infected with two or three of these viruses and PDV with an infection rate of 26.3% was the most widespread virus in symptomatic trees in western Mediterranean region. Thirty samples were selected and tested by a multiplex RT-PCR (mRT-PCR) for simultaneous detection of these viruses. While PPV was not detected, more than half of the tested 20 samples were individually or mixed infected with ApMV, ACLSV, PNRSV and PDV. The mRT-PCR results were confirmed by detection of these viruses individually in some of the field samples using RT-PCR with primes specific to each virus. Comparison of ELSA and mRT-PCR results of 30 samples showed that numbers of infected and mixed infected samples as well as infection and mixed infection rates were significantly higher in RT-PCR (20 and 66.7%) than in ELISA (14 and 46.7%). The results confirm that mRT-PCR is more sensitive than ELISA.

• 한국잡초학회지
• /
• 제30권4호
• /
• pp.454-459
• /
• 2010

비선택성 제초제인 글루포시네이트암모늄액제를 농작업자의 오판에 의한 오용으로 사과나무에 직접 살포된 경우 품종별, 수령별로 나타나는 약해반응과 수확에 미치는 영향을 조사하였다. 사과나무 품종에 따라 약해정도가 다르게 나타났다. 약제살포 당해연도에는 산사 > 쓰가루 > 홍로 > 후지 순으로, 약제살포 다음해에는 후지 > 쓰가루 > 홍로 > 산사 순으로 피해가 나타났으나, 그 피해정도는 당년에 비해 경미하였다. 이듬해, 후지에서는 꽃과 착과정도에서만 30%정도 감소하였으나 쓰가루, 홍로, 산사에서는 개화상태, 착과율 등에서 전혀 피해증상이 관찰되지 않았다. 약해를 받은 당년에 조생종 쓰가루와 중만생종후지 대상으로 각각 수확기 10일전에 사과 중 잔류량을 검사한 결과, 잔류량은 검출한계(0.04ppm) 미만이었다. 그리고 오용 2년 후(2009년)에는 후지, 쓰가루, 홍로, 산사 모두에서 약해증상이 나타나지 않았다.