Abstract
Object detection plays an important role in computer vision. It has a variety of applications, including security detection, vehicle recognition, and service robots. With the continuous improvement of public databases and the development of deep learning, object detection has witnessed significant breakthroughs. However, the object detection of sweeping robots during operations should consider various factors, including the camera angle, indoor scenery, and identification of object category. To the best of our knowledge, no corresponding database on these conditions has been developed. In this study, we review the development of object detection based on deep learning in computer vision. Then, we propose a large-scale publicly available benchmark dataset called object detection for sweeping robots in home scenes (ODSR-IHS). The dataset has 6,000 images and 16,409 instances of 14 object categories. Finally, we evaluate several state-of-the-art methods on the ODSR-IHS dataset and transplant them to the hardware to establish a benchmark dataset for object recognition research on sweeping robots.
Highlights
Owing to the rapid development of computer vision, sweeping robots based on visual navigation technology have received considerable interest because of their low cost
(2) We evaluated several state-of-the-art object detection methods based on deep learning on our ODSR-IHS dataset
In this study, we propose a novel dataset, called ODSR-IHS, which can be used as a benchmark dataset for object detection of sweeping robots
Summary
Owing to the rapid development of computer vision, sweeping robots based on visual navigation technology have received considerable interest because of their low cost. (2) We evaluated several state-of-the-art object detection methods based on deep learning on our ODSR-IHS dataset. (3) To further verify the feasibility of the model, we transplanted the trained AI model to embedded system, which can provide a meaningful benchmark for the object detection application of sweeping robots in actual environments and promote the development of lightweight (convolutional neural networks) CNNs. The remainder of this paper is organized as follows: In Section II, we summarize past studies on commonly used databases for object detection and their corresponding algorithms, mainly focusing on deep learning. The ODSR-IHS dataset we proposed contains a total of 6,000 images, encompassing 16,409 instances of 14 types of objects These images cover various complex home scenes and different changes in lighting, including single and multiple objects and mutual occlusion between objects. Each line of the text file contains the ID of a specific category and its position and size in the image
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.