• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.6
• /
• pp.635-644
• /
• 2020

Object recognition, detection and instance segmentation based on DL (Deep Learning) have being used in various practices, and mainly optical images are used as training data for DL models. The major objective of this paper is object segmentation and building detection by utilizing multimodal datasets as well as optical images for training Detectron2 model that is one of the improved R-CNN (Region-based Convolutional Neural Network). For the implementation, infrared aerial images, LiDAR data, and edges from the images, and Haralick features, that are representing statistical texture information, from LiDAR (Light Detection And Ranging) data were generated. The performance of the DL models depends on not only on the amount and characteristics of the training data, but also on the fusion method especially for the multimodal data. The results of segmenting objects and detecting buildings by applying hybrid fusion - which is a mixed method of early fusion and late fusion - results in a 32.65% improvement in building detection rate compared to training by optical image only. The experiments demonstrated complementary effect of the training multimodal data having unique characteristics and fusion strategy.

• The Journal of Korea Robotics Society
• /
• v.16 no.4
• /
• pp.336-344
• /
• 2021

This paper presents a robust deep learning-based human tracking framework in crowded environments. For practical human tracking applications, a target must be robustly tracked even in undetected or overcrowded situations. The proposed framework consists of two parts: robust deep learning-based human detection and tracking while recognizing the aforementioned situations. In the former part, target candidates are detected using Detectron2, which is one of the powerful deep learning tools, and their weights are computed and assigned. Subsequently, a candidate with the highest weight is extracted and is utilized to track the target human using a Kalman filter. If the bounding boxes of the extracted candidate and another candidate are overlapped, it is regarded as a crowded situation. In this situation, the center information of the extracted candidate is compensated using the state estimated prior to the crowded situation. When candidates are not detected from Detectron2, it means that the target is completely occluded and the next state of the target is estimated using the Kalman prediction step only. In two experiments, people wearing the same color clothes and having a similar height roam around the given place by overlapping one another. The average error of the proposed framework was measured and compared with one of the conventional approaches. In the error result, the proposed framework showed its robustness in the crowded environments.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.2
• /
• pp.119-125
• /
• 2023

In this study, we photograph driving obstacle objects such as crosswalks, side spheres, manholes, braille blocks, partial ramps, temporary safety barriers, stairs, and inclined curb that hinder or cause inconvenience to the movement of the vulnerable using electric mobility aids. We develop an optimal AI model that classifies photographed objects and automatically recognizes them, and implement an algorithm that can efficiently determine obstacles in front of electric mobility aids. In order to enable object detection to be AI learning with high probability, the labeling form is labeled as a polygon form when building a dataset. It was developed using a Mask R-CNN model in Detectron2 framework that can detect objects labeled in the form of polygons. Image acquisition was conducted by dividing it into two groups: the general public and the transportation weak, and image information obtained in two areas of the test bed was secured. As for the parameter setting of the Mask R-CNN learning result, it was confirmed that the model learned with IMAGES_PER_BATCH: 2, BASE_LEARNING_RATE 0.001, MAX_ITERATION: 10,000 showed the highest performance at 68.532, so that the user can quickly and accurately recognize driving risks and obstacles.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.3
• /
• pp.166-172
• /
• 2022

In this study, we photograph driving obstacle objects such as crosswalks, side spheres, manholes, braille blocks, partial ramps, temporary safety barriers, stairs, and inclined curb that hinder or cause inconvenience to the movement of the vulnerable using electric mobility aids. We develop an optimal AI model that classifies photographed objects and automatically recognizes them, and implement an algorithm that can efficiently determine obstacles in front of electric mobility aids. In order to enable object detection to be AI learning with high probability, the labeling form is labeled as a polygon form when building a dataset. It was developed using a Mask R-CNN model in Detectron2 framework that can detect objects labeled in the form of polygons. Image acquisition was conducted by dividing it into two groups: the general public and the transportation weak, and image information obtained in two areas of the test bed was secured. As for the parameter setting of the Mask R-CNN learning result, it was confirmed that the model learned with IMAGES_PER_BATCH: 2, BASE_LEARNING_RATE 0.001, MAX_ITERATION: 10,000 showed the highest performance at 68.532, so that the user can quickly and accurately recognize driving risks and obstacles.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.11a
• /
• pp.116-118
• /
• 2022

MPEG-VCM(Video Coding for Machine)은 입력된 이미지/비디오의 특징(feature)를 압축하는 Track 1 과 입력 이미지/비디오를 직접 압축하는 Track 2 로 나뉘어 표준화가 진행 중이다. 본 논문은 Track 1 의 비전임무 네트워크로 사용하는 Detectron2 의 FPN(Feature Pyramid Network)에서 추출한 멀티-스케일 특징을 효율적으로 압축하는 MSFC 기반의 압축 모델의 개선 기법을 제시한다. 제안기법은 해상도를 줄여서 단일-스케일 압축맵을 압축하는 기존의 압축 모델에서 저해상도 특징맵을 고해상도 특징맵에 바텀-업(Bottom-Up) 구조로 합성하여 단일-스케일 특징맵을 구성하는 바텀-업 MSFF 를 가지는 압축 모델을 제시한다. 제안방법은 기존의 모델 보다 BPP-mAP 성능에서 1 ~ 2.7%의 개선된 BD-rate 성능을 보이며 VCM 의 이미지 앵커(image anchor) 대비 최대 -85.94%의 BD-rate 성능향상을 보인다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.143-145
• /
• 2022

MPEG-VCM(Video Coding for Machine)은 입력된 비디오 특징(feature)를 압축하는 Track1 과 입력 영상을 직접 압축하는 Track2 로 나뉘어 표준화가 진행중이다. 본 논문은 VCM Track 1 에 해당하는 Detectron2 FPN(Feature Pyramid Network)에서 추출한 다중 스케일 특징맵을 VVC 로 압축하는 MSFC(Multi-Scale Feature Compression)을 구조를 제안한다. 본 논문의 MSFC 에서는 다중 스케일 특징을 결합하여 부호화/복호화하는 기존의 구조에서 특징맵의 해상도를 줄여 압축하는 개선된 MSFC 를 제시한다. 제안 방법은 VCM 의 Track2 의 영상 앵커(image anchor) 보다 우수한 BPP-mAP 성능을 보이고 최대 -84.98%의 BD-rate 성능향상을 보인다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.7-9
• /
• 2021

본 논문은 Detectron2 [1]에서 지원하는 객체 검출 임무 수행 네트워크의 과정 중에서 추출한 피쳐 맵을 신경망 기반으로 압축하는 방법을 제안한다. 이를 위해, 신경 망 기반 영상 압축을 지원하는 공개 소프트웨어인 CompressAI [2] 모델 중 하나인 bmshj2018-hyperprior 의 압축 네트워크를 활용하여 임무 수행 네트워크의 과정 중 스탬 레이어(stem layer)에서 추출된 피쳐 맵을 압축하도록 학습시켰다. 또한, 압축 네트워크의 입력 피쳐 맵의 너비와 높이 크기가 64 의 배수가 되도록 객체 검출 네트워크의 입력 영상 보간 값을 조정하는 방법도 제안한다. 제안하는 신경망 기반 피쳐 맵 압축 방법은 피쳐 맵을 최근 표준이 완료된 차세대 압축 표준 방법인 VVC(Versatile Video Coding, [3])로 압축한 결과에 비해 큰 성능 향상을 보이고, VCM 앵커와 유사한 성능을 보인다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.146-149
• /
• 2022

기존 RDO(Rate Distortion Optimization) 기반 압축 방식은 압축 성능에 초점을 두기 때문에 영상 내 인지 특성이 무시될 수 있다. 따라서 RoI(Region of Interest)을 기반으로 압축률을 조절하는 연구가 고안[1, 2, 3, 4] 되었으며, HVS(Human Visual System) 관점에서 영상 내 중요한 부분에 대해 더 높은 품질로 영상을 압축하는 연구가 대부분이다. 최근 인공지능 기술이 발전함에 따라 지능형 영상 분석에 대한 수요가 증가하고 있으며, 이에 따라 머신 비전을 위한 영상 부호화 및 효율적인 전송에 대한 필요성이 대두되고 있다. 본 논문에서는 VVC(Versatile Video Coding)의 dQP(delta Quantization Parameter)를 활용하여 RoI(Region of Interest) 기반압축 방법을 제안하고, 두가지의 RoI 추출 방식을 소개한다. Detectron2 Faster R-CNN X101-FPN [5]의 첫번째 탐지기를 통해 후보 영역 기반 RoI 을 추출하고, 두번째 탐지기를 통해 객체 기반 RoI 을 추출하여, 영상 내 객체 부분과 비객체 부분으로 나누어 서로 다른 압축률로 압축을 수행하였으며, 이에 따른 성능을 비교하고자 한다.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.23 no.3
• /
• pp.156-162
• /
• 2021

The number of smart farms has increased to save labor in agricultural production as the subsidy become available from central and local governments. The number of illegal greenhouses has also increased, which causes serious issues for the local governments. In the present study, we developed Mask-RCNN model to detect greenhouses based on satellite images. Greenhouses in the satellite images were labeled for training and validation of the model. The Mask-RC NN model had the average precision (AP) of 75.6%. The average precision values for 50% and 75% of overlapping area were 91.1% and 81.8%, respectively. This results indicated that the Mask-RC NN model would be useful to detect the greenhouses recently built without proper permission using a periodical screening procedure based on satellite images. Furthermore, the model can be connected with GIS to establish unified management system for greenhouses. It can also be applied to the statistical analysis of the number and total area of greenhouses.