• Korean Journal of Agricultural Science
• /
• v.50 no.3
• /
• pp.357-364
• /
• 2023

In this study, we propose a ripe tomato bunch localization method based on convolutional neural networks, to be applied in robotic harvesting systems. Tomato images were obtained from a smart greenhouse at the Rural Development Administration (RDA). The sample images for training were extracted based on tomato maturity and resized to 128 × 128 pixels for use in the classification model. The model was constructed based on four-layer convolutional neural networks, and the classes were determined based on stage of maturity, using a Softmax classifier. The localization of the ripe tomato bunch region was indicated on a class activation map. The class activation map could show the approximate location of the tomato bunch but tends to present a local part or a large part of the ripe tomato bunch region, which could lead to poor performance. Therefore, we suggest a recursive method to improve the performance of the model. The classification results indicated that the accuracy, precision, recall, and F1-score were 0.98, 0.87, 0.98, and 0.92, respectively. The localization performance was 0.52, estimated by the Intersection over Union (IoU), and through input recursion, the IoU was improved by 13%. Based on the results, the proposed localization of the ripe tomato bunch area can be incorporated in robotic harvesting systems to establish the optimal harvesting paths.

• Journal of Convergence for Information Technology
• /
• v.11 no.10
• /
• pp.45-52
• /
• 2021

In this paper, we proposed an Atrous Residual U-Net (AR-UNet) to improve the segmentation accuracy of semantic segmentation method based on U-Net. The U-Net is mainly used in fields such as medical image analysis, autonomous vehicles, and remote sensing images. The conventional U-Net lacks extracted features due to the small number of convolution layers in the encoder part. The extracted features are essential for classifying object categories, and if they are insufficient, it causes a problem of lowering the segmentation accuracy. Therefore, to improve this problem, we proposed the AR-UNet using residual learning and ASPP in the encoder. Residual learning improves feature extraction ability and is effective in preventing feature loss and vanishing gradient problems caused by continuous convolutions. In addition, ASPP enables additional feature extraction without reducing the resolution of the feature map. Experiments verified the effectiveness of the AR-UNet with Cityscapes dataset. The experimental results showed that the AR-UNet showed improved segmentation results compared to the conventional U-Net. In this way, AR-UNet can contribute to the advancement of many applications where accuracy is important.

• Food Science and Preservation
• /
• v.6 no.3
• /
• pp.264-269
• /
• 1999

The effects of film thickness and moisture absorbing materials(MAM) on the ginger qualities were investigated during MA storage for 150 days. LDPE films of 0.04, 0.06 and 0.08mm thicknesses were applied to select the adequate film in gas permiability. MAMs of sheet and sachet type were applied to 0.06mm-LDPE film bag to prevent moisture condensation during storage. Thickness of film used affected neither weight loss nor firmness of gingers during U storage. But the thinner in thickness showed the less changes in sprouting, spoilage and reducing sugar. Moreover, compared with MAP alone, MAPs with MAM of both sachet and sheet styles led to the weight loss of gingers increasing by more than 3-4 times, but sprouting rate, spoilage rate and reducing sugar decreasing by 3.3, 1.4 and 2.3 times, respectively. These results showed that MAP with UM was significantly effective to prolong the freshness of ginger by longer than 2 times, compared with MAP alone.

• Korean Journal of Materials Research
• /
• v.3 no.4
• /
• pp.336-343
• /
• 1993

The standard map for polyerystalline $YBa_2Cu_3O_7$-X was construeted and indexed from pressed tape sintered bulk and compressed bulk samples were drawn by the [001-110-110] stereographie quadrant map and by the ECPs from individual grains of the speeimens. The relationship between the (001) texture development and critical current density (Jc) for various samples was discussed. An (001) orintation was preferred in the pressed tape thickness was decreased. and in the compressed bulk as the compression strain was increaesed.The (001) orientation made a signifieant in crease on Jc. We conclded that an ECP technique was very useful for the analysis of the orientation distribution in small-sized $YBa_2Cu_3O_7$-X specimens such as superonducting tapes or thin films.

• Journal of the Korean Geophysical Society
• /
• v.9 no.3
• /
• pp.209-217
• /
• 2006

The major continental blocks in NE-Asia are the North China Block and the South China Blo, which have collided, starting from the Korean peninsula. The suture zone in NE China between two blocks is well defined from the QinIing-Dabie-Orogenic Belt to the Jiaodong (Sulu) Belt by the geological and geophysical interpretation. The discovery of high pressure metamorphic rocks in the Hongsung area of the Korean peninsula can be used to estimate the suture zone. This indicates that the suture zone in the Jiaodong Belt might be extended to Hongsung area. However, due to the lack of geological and geophysical data over the Yellow sea, the extension of the suture zone to the Korean peninsula across the Yellow Sea is obscure. To find out the tectonic relationship between NE China and the Korean peninsula it is necessary to complete U-ie homogeneous geophysical dataset of NE Asia, which can be provided by satellite observations. The CHAMP lithospheric magnetic field (MF3) and CHAMP-GRACE gravity field, combined with surface measured data, allow a much more accurate in-ference of tectonic structures than previously available. The CHAMP magnetic anomaly map reveals significant magnetic lows in the Yellow Sea near Nanjing and Hongsung, where are characterized by gravity highs on U-ie CHAMP-GRACE gravity anomaly map. To evaluate the depth and location of poten-tial field anomaly causative bodies, the Euler Deconvolution method is implemented. After comparing the two potential field solutions with the simplified geological map containing tectonic lines and the distribution of earthquakes epicenters, it is found that the derived structure boundaries of both are well coincident with the seismic activities as well as with the tectonic lineaments. The interpretation of the CHAMP satellite magnetic and GRACE satellite gravity datasets reveal two tectonic boundaries in U-ie Yellow Sea and the Korean peninsula, indicating U-ie norttiern and southern margins of the suture zone between the North China Block and the South China Block. The former is extended from the Jiaodong Belt in East China to the Imjingang Belt on the Korean peninsula, the later from Nanjing, East China, to Hongsung, the Korean peninsula. The tectonic movement in or near the suture zone might be responsible for the seismic activities in the western region of the Korean Peninsula and the development of the Yellow Sea sedimentary basin.

• The Journal of the Acoustical Society of Korea
• /
• v.42 no.6
• /
• pp.544-551
• /
• 2023

Speech enhancement used to improve the perceptual quality and intelligibility of noise speech has been studied as a method using a complex-valued spectrum that can improve both magnitude and phase in a method using a magnitude spectrum. In this paper, a study was conducted on how to apply attention mechanism to complex-valued spectrum-based speech enhancement systems to further improve the intelligibility and quality of noise speech. The attention is performed based on additive attention and allows the attention weight to be calculated in consideration of the complex-valued spectrum. In addition, the global average pooling was used to consider the importance of the feature map. Complex-valued spectrum-based speech enhancement was performed based on the Deep Complex U-Net (DCUNET) model, and additive attention was conducted based on the proposed method in the Attention U-Net model. The results of the experiments on noise speech in a living room environment showed that the proposed method is improved performance over the baseline model according to evaluation metrics such as Source to Distortion Ratio (SDR), Perceptual Evaluation of Speech Quality (PESQ), and Short Time Object Intelligence (STOI), and consistently improved performance across various background noise environments and low Signal-to-Noise Ratio (SNR) conditions. Through this, the proposed speech enhancement system demonstrated its effectiveness in improving the intelligibility and quality of noisy speech.

• Proceedings of the KSPS conference
• /
• 2002.11a
• /
• pp.129-132
• /
• 2002

Aerodynamic analysis study was performed on 14 normal subjects(2 male, 12 female) by nonsense syllables composed of Korean bilabial stop(/p, p', $p^{h}$) and their preceding and/or following vowel /i, a, u/. That is [pi, p'i, phi, pa, p'a, pha, pu, p'u, $p^{h}u$]. All measures were analysed using Aerophone II voice function analyzer and included peak air pressure, mean air pressure, maximum flow rate, volume, mean SPL. As results, first, MSPL and MAP of /p, p', $p^{h}$/ themselves were significantly different. In addition, different vowel enviroment also produced significantliy different aerodynamic chracteristics those consonants.

• Bulletin of the Korean Mathematical Society
• /
• v.57 no.3
• /
• pp.583-595
• /
• 2020

Let HE_I, HE_II, HE_III and HE_IV be the first, second, third and fourth type Loo-Keng Hua domain respectively, 𝜑 a holomorphic self-map of HE_I, HE_II, HE_III, or HE_IV and u ∈ H(𝓜) the space of all holomorphic functions on 𝓜 ∈ {HE_I, HE_II, HE_III, HE_IV}. In this paper, motivated by the well known Hua's matrix inequality, first some inequalities for the points in the Bers-type spaces of the Loo-Keng Hua domains are obtained, and then the boundedness and compactness of the weighted composition operators W_𝜑,u : f ↦ u · f ◦ 𝜑 on Bers-type spaces of these domains are characterized.

• Communications of the Korean Mathematical Society
• /
• v.18 no.4
• /
• pp.653-659
• /
• 2003

Let $1{\;}\leq{\;}p{\;}\infty{\;}and{\;}{\alpha}{\;}>{\;}-1$. If f is a holomorphic self-map of the open unit disc U of C with f(0) = 0, then the quantity $\int_U\;\{\frac{$\mid$f'(z)$\mid$}{1\;-\;$\mid$f(z)$\mid$^2}\}^p\;(1\;-\;$\mid$z$\mid$)^{\alpha+p}dxdy$ is equivalent to the operator norm of the composition operator $C_f{\;}:{\;}B{\;}\rightarrow{\;}A^{p,{\alpha}$ defined by $C_fh{\;}={\;}h{\;}\circ{\;}f{\;}-{\;}h(0)$, where B and $A^{p,{\alpha}$ are the Bloch space and the weighted Bergman space on U respectively.

• The Journal of Asian Finance, Economics and Business
• /
• v.8 no.3
• /
• pp.1417-1425
• /
• 2021

The purpose of this study is to map mineral and coal (Minerba) policies by analyzing the law and producing research novelty on mineral and coal governance policies in Indonesia with the Theory U approach. This study uses a qualitative research with the U theory approach. The results of the SNA analysis indicate that good mineral and coal governance must be carried out at various levels, from micro to macro. First, related to regulations on mineral and coal governance at the macro level. Regulations regarding the management of mineral and coal need to be carried out with a deeper evaluation related to the tax system, licensing system, increased value added and downstreaming and mineral and coal funds. The second is related to the regulation of mining management in the meso level. Third, related to the management of mineral and coal at the micro level, it is necessary to conduct more stringent supervision of the impacts caused by the mining sector of the Minerba. In addition, surveillance is also carried out as per law in order to avoid harmful behavior for both the company and the state. The originality of this research is the theory of U in the Mining Law research.