Drone-Based Object Counting by Spatially Regularized Regional Proposal Network

By M. Hsieh, Yen-Liang Lin, Winston H. Hsu. IEEE International Conference on Computer Vision, 2017

537 citations recorded by the source corpus.

Existing counting methods often adopt regression-based approaches and cannot precisely localize the target objects, which hinders the further analysis (e.g., high-level understanding and fine-grained classification). In addition, most of prior work mainly focus on counting objects in static environments with fixed cameras. Motivated by the advent of unmanned flying vehicles (i.e., drones), we are interested in detecting and counting objects in such dynamic environments. We propose Layout Proposal Networks (LPNs) and spatial kernels to simultaneously count and localize target objects (e.g., cars) in videos recorded by the drone. Different from the conventional region proposal methods, we leverage the spatial layout information (e.g., cars often park regularly) and introduce these spatially regularized constraints into our network to improve the localization accuracy. To evaluate our counting method, we present a new large-scale car parking lot dataset (CARPK) that contains nearly 90,000 cars captured from different parking lots. To the best of our knowledge, it is the first and the largest drone view dataset that supports object counting, and provides the bounding box annotations.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/b43128a3baeb5f815bf8b1dbc9345ca24bc2172e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

1. Yiming Sun
2. Bing Cao
3. Pengfei Zhu
4. Qinghua Hu
5. Guisong Xia
6. Jian Ding
7. Jiebo Luo
8. Shubhra Aich

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. Ming-Hsuan Yang
4. Yi Yang
5. Qi Tian
6. Chunhua Shen
7. Xiaogang Wang
8. Alan L. Yuille

Citations in the corpus

• RemoteCLIP: A Vision Language Foundation Model for Remote Sensing (2023, background): “CARPK [86] 2017 1,568 1 106,690 1280×720 CARPK dataset contains nearly 90,000 cars collected from four different parking lots by drones.”
• Towards Large-Scale Small Object Detection: Survey and Benchmarks (2022, unknown): “CARPK [144] HBB 1448 1 47 512 1378 train/test 1280 × 720 2017 VisDrone [145] HBB 8629 10 128964 90107 65972 train/val/test-dev 1490 × 957 2021 AI-TOD [116] HBB 28036 8 301753 35729 8484 train/val 800 × 800 2021 DOTA [30] OBB 2423 18 181124 52583 45139 train/val 2217 × 2074 2021 DIOR-R [141] OBB 2346”
• Dynamic Refinement Network for Oriented and Densely Packed Object Detection (2020, methodology): “For oriented object detection, some methods [8, 10, 25, 29, 28] adopt the R-CNN [35] framework and use numerous anchors with different angles, scales, and aspect ratios, at the expense of considerably increasing computation complexity.”
• The Unmanned Aerial Vehicle Benchmark: Object Detection and Tracking (2018, methodology): “Hsieh et al. [22] present an annotated dataset aiming at counting vehicles in parking lots.”
• Bayesian Loss for Crowd Count Estimation With Point Supervision (2019, background): “ing estimating the number of vehicles in traffic congestion [29, 30, 56, 14, 28], counting the cells and bacteria from microscopic images [20, 42, 44, 45, 8], and animal crowd estimations for ecological survey [27, 1, 18], to name a few.”
• Drone-Based RGB-Infrared Cross-Modality Vehicle Detection Via Uncertainty-Aware Learning (2020, background): “CARPK [30] contains images172 captured by drones from various parking lots.”
• Improving Object Counting with Heatmap Regulation (2018, methodology): “We evaluate the HR strategy on four different object counting datasets — two for cars (CARPK, PUCPR+) [2], one for crowds [9] and the other for biological cells (VGG-Cells) [10].”
• VisDrone-DET2019: The Vision Meets Drone Object Detection in Image Challenge Results (2019, methodology): “Compared with the previous drone based datasets [29, 15, 9], a larger scale drone based object detection dataset [48] is proposed to evaluate detection algorithms in real scenarios.”
• Guided Attention Network for Object Detection and Counting on Drones (2019, methodology): “org/ YOLO9000 [23], LPN [10], RetinaNet [17], YOLOv3 [24], IoUNet [8], and SA+CF+CRT [15].”
• Counting Vehicles with Deep Learning in Onboard UAV Imagery (2019, methodology): “Finally, comparing our approach with the one proposed by [14] in terms of images processed per second, we can say that the algorithm in [14] is faster than ours (14 FPS instead of 4 FPS), but the error in [14] is about 7 times larger for the CARPK dataset and about 23 times larger for the PUCPR+ dat”
• Object Detection and Image Segmentation with Deep Learning on Earth Observation Data: A Review-Part I: Evolution and Recent Trends (2020, background): “With it, the Region Proposal Network (RPN) module was introduced and finally two stage object detection became end to end trainable models, unified in a single network performing all tasks needed for competitive object detection results.”
• Object Detection in Aerial Images: A Large-Scale Benchmark and Challenges (2021, background): “[36], vehicle datasets [8], [15], [16], [26], [37], [38], [39], ship datasets [4], [40], and plane datasets [17], [41].”
• Vision Meets Drones: A Challenge (2018, methodology): “Some recent efforts [9], [10], [11] have been devoted to construct datasets with drone platform focusing on object detection or tracking.”