Motion target detection algorithm based on monocular vision

Feng, Yaokai; Wu, Qizhou; He, Guanghui

doi:10.1145/3056662.3056669

Cited by 3 publications

(2 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Frame differencing is simple to implement, has low computational requirements, and exhibits strong adaptability and robustness in dynamic environments. However, in the presence of large areas of similar grayscale values on the surface of the moving object, frame differencing may result in holes in the image [31,32]. In recent years, deep learning technology has shown its remarkable feature extraction capabilities.…”

Section: Motion Object Extractionmentioning

confidence: 99%

Automatic Detection Method for Black Smoke Vehicles Considering Motion Shadows

Wang,

Chen,

2023

Sensors

View full text Add to dashboard Cite

Various statistical data indicate that mobile source pollutants have become a significant contributor to atmospheric environmental pollution, with vehicle tailpipe emissions being the primary contributor to these mobile source pollutants. The motion shadow generated by motor vehicles bears a visual resemblance to emitted black smoke, making this study primarily focused on the interference of motion shadows in the detection of black smoke vehicles. Initially, the YOLOv5s model is used to locate moving objects, including motor vehicles, motion shadows, and black smoke emissions. The extracted images of these moving objects are then processed using simple linear iterative clustering to obtain superpixel images of the three categories for model training. Finally, these superpixel images are fed into a lightweight MobileNetv3 network to build a black smoke vehicle detection model for recognition and classification. This study breaks away from the traditional approach of “detection first, then removal” to overcome shadow interference and instead employs a “segmentation-classification” approach, ingeniously addressing the coexistence of motion shadows and black smoke emissions. Experimental results show that the Y-MobileNetv3 model, which takes motion shadows into account, achieves an accuracy rate of 95.17%, a 4.73% improvement compared with the N-MobileNetv3 model (which does not consider motion shadows). Moreover, the average single-image inference time is only 7.3 ms. The superpixel segmentation algorithm effectively clusters similar pixels, facilitating the detection of trace amounts of black smoke emissions from motor vehicles. The Y-MobileNetv3 model not only improves the accuracy of black smoke vehicle recognition but also meets the real-time detection requirements.

show abstract

Section: Motion Object Extractionmentioning

confidence: 99%

Automatic Detection Method for Black Smoke Vehicles Considering Motion Shadows

Wang,

Chen,

2023

Sensors

View full text Add to dashboard Cite

show abstract

“…The system has high recognition accuracy for transformations such as illumination and occlusion, which further improves the computing speed of the system. In 2021, Zeng Xiao of Sichuan University [11] proposed a target detection system based on TCP/IP remote communication. The system breaks through the traditional interaction method, but there are problems of high information delay and data loss.…”

Section: Introductionmentioning

confidence: 99%

Assembly training system on HoloLens using embedded algorithm

Qin¹,

Wang²,

Zhang³

et al. 2023

Third International Symposium on Computer Engineering and Intelligent Communications (ISCEIC 2022)

View full text Add to dashboard Cite

In this article, we demonstrate an implementation on Microsoft HoloLens, deep learning supported in the context of object detection. The main aim of the training system is to create the more accurate object detection model for Augmented Reality using deep learning models for image recognition directly on the HoloLens 2. In terms of the object detection approach, a deep learning model called YOLOv5 has been used for the implementation of this system. This article uses the Windows ML API to implement machine learning in augmented reality applications. A simple and easy method of drawing lines between specified 2D coordinates on a canvas is proposed. The module division and development steps of the development of augmented reality training system are given. Our system provides the annotation of augmented object detected and its bounding box via HoloLens. It allows to detect the new object in a few milliseconds. Preliminary results show a great rate of object detection and reasonable detection time.

show abstract

Distributed Video Analysis for Mobile Live Broadcasting Services

Chen

Guan

2020

2020 IEEE Wireless Communications and Networking Conference (WCNC)

View full text Add to dashboard Cite

Motion target detection algorithm based on monocular vision

Cited by 3 publications

References 6 publications

Automatic Detection Method for Black Smoke Vehicles Considering Motion Shadows

Automatic Detection Method for Black Smoke Vehicles Considering Motion Shadows

Assembly training system on HoloLens using embedded algorithm

Distributed Video Analysis for Mobile Live Broadcasting Services

Contact Info

Product

Resources

About