We can take the smallest nadir picture, run some segmentation pipeline. Hmm, collects, sounds not so challenging now. Dataset had 27 collects pictures of the same place with nadirs ranging from 7 to 54 degrees. The goal of SpaceNet 7 1, 2 was to track building construction over time in moderate resolution Planet imagery over a deep time stack and dozens of areas of interest (AOIs) across six continents. See GitHub repository and blog post (in Japanese) for details. A key challenge of all previous Spacenet competitions was to find objects at the satellite image.
Tracking urbanization and construction at the unique building level is critical for improving upon existing course population estimates. Implemented “Gaussian YOLOv3: An Accurate and Fast Object Detector Using Localization Uncertainty for Autonomous Driving” in PyTorch and reproduced the performance of the original Darknet implementation.ĭeveloped a convolutional neural network to count cars in aerial images. SpaceNet 7 has demonstrated that identifying, tracking, and detecting change in precise building footprints using moderate resolution (4m) Planet imagery is possible. See GitHub repository.īy integrating it with a monocular deep learning-based distance estimation method, I have constructed a monocular 3D human recoginition system.
Implemented “A simple yet effective baseline for 3d human pose estimation” in PyTorch and reproduced the performance of the original Tensorflow implementation. See winners announcement, slides, video, and GitHub repository for details. I have developed an effective pipeline combining deep learning-based instance segmentation and post-filtering using LightGBM. In SpaceNet-6 competition, participants were asked to extract building footprints from SAR (Synthetic Aperature Radar) remote sensing imageries. Won the 4th place ( top 1% amoung 400+ teams) in SpaceNet-6 competition which was held as a part of CVPR’20 EarthVision workshop. I have developed a pipeline that combined deep learning-based instance segmentation and a simple yet effective rule-based multi-object tracking method.Ĥth Place Solution for SpaceNet-6 Challenge
#SPACENET 7 SERIES#
In SpaceNet-7 competition, participants were asked to identify and track buildings in satellite imagery time series collected over rapidly urbanizing areas. Won the 4th place ( top 1.3 % amoung 300+ teams) in SpaceNet-7 which was featured as a competition at NeurIPS’20 conference.
Portfolio 4th Place Solution for SpaceNet-7 Challenge
0 kommentar(er)
