Multiple object detection of workpieces based on fusion of deep learning and image processing

Yi  Lei; Xifan Yao; Wocheng  Chen; Junming  Zhang; Jorn Mehnen; Erfu Yang

Multiple object detection of workpieces based on fusion of deep learning and image processing

Yi Lei, Xifan Yao, Wocheng Chen, Junming Zhang, Jorn Mehnen, Erfu Yang

Design, Manufacturing And Engineering Management

Research output: Contribution to conference › Paper › peer-review

11 Downloads (Pure)

Abstract

A workpiece detection method based on fusion of deep learning and image processing is proposed. Firstly, the workpiece bounding boxes are located in the workpiece images by YOLOv3, whose parameters are compressed by an improved convolutional neural network residual structure pruning strategy. Then, the workpiece images are cropped based on the bounding boxes with cropping biases. Finally, the contours and suitable gripping points of the workpieces are obtained through image processing. The experimental results show that mean Average Precision (mAP) is 98.60% for YOLOv3, and 99.38% for that one by pruning 50.89% of its parameters, and the inference time is shortened by 31.13%. Image processing effectively corrects the bounding boxes obtained by deep learning, and obtains workpiece contour and gripping point information.

Original language	English
Number of pages	6
Publication status	Accepted/In press - 19 Apr 2020
Event	IEEE World Congress on Computational Intelligence 2020 - Glasgow, United Kingdom Duration: 19 Jul 2020 → 24 Jul 2020 https://wcci2020.org/

Conference

Conference	IEEE World Congress on Computational Intelligence 2020
Abbreviated title	WCCI
Country	United Kingdom
City	Glasgow
Period	19/07/20 → 24/07/20
Internet address	https://wcci2020.org/

Keywords

manufacturing and industrial applications
applications of deep networks
workpieces detection
deep learning
pruning filters
image processing

Access to Document

Lei-etal-WCCI-2020-Multiple-object-detection-of-workpieces-based-on-fusionAccepted author manuscript, 555 KB

https://wcci2020.org/

Fingerprint Dive into the research topics of 'Multiple object detection of workpieces based on fusion of deep learning and image processing'. Together they form a unique fingerprint.

Cite this

@conference{38b1b272aa4847d19d39b6ffff72cbbb,

title = "Multiple object detection of workpieces based on fusion of deep learning and image processing",

abstract = "A workpiece detection method based on fusion of deep learning and image processing is proposed. Firstly, the workpiece bounding boxes are located in the workpiece images by YOLOv3, whose parameters are compressed by an improved convolutional neural network residual structure pruning strategy. Then, the workpiece images are cropped based on the bounding boxes with cropping biases. Finally, the contours and suitable gripping points of the workpieces are obtained through image processing. The experimental results show that mean Average Precision (mAP) is 98.60% for YOLOv3, and 99.38% for that one by pruning 50.89% of its parameters, and the inference time is shortened by 31.13%. Image processing effectively corrects the bounding boxes obtained by deep learning, and obtains workpiece contour and gripping point information.",

keywords = "manufacturing and industrial applications, applications of deep networks, workpieces detection, deep learning, pruning filters, image processing",

author = "Yi Lei and Xifan Yao and Wocheng Chen and Junming Zhang and Jorn Mehnen and Erfu Yang",

year = "2020",

month = apr,

day = "19",

language = "English",

note = "IEEE World Congress on Computational Intelligence 2020, WCCI ; Conference date: 19-07-2020 Through 24-07-2020",

url = "https://wcci2020.org/",

}

TY - CONF

T1 - Multiple object detection of workpieces based on fusion of deep learning and image processing

AU - Lei, Yi

AU - Yao, Xifan

AU - Chen, Wocheng

AU - Zhang, Junming

AU - Mehnen, Jorn

AU - Yang, Erfu

PY - 2020/4/19

Y1 - 2020/4/19

N2 - A workpiece detection method based on fusion of deep learning and image processing is proposed. Firstly, the workpiece bounding boxes are located in the workpiece images by YOLOv3, whose parameters are compressed by an improved convolutional neural network residual structure pruning strategy. Then, the workpiece images are cropped based on the bounding boxes with cropping biases. Finally, the contours and suitable gripping points of the workpieces are obtained through image processing. The experimental results show that mean Average Precision (mAP) is 98.60% for YOLOv3, and 99.38% for that one by pruning 50.89% of its parameters, and the inference time is shortened by 31.13%. Image processing effectively corrects the bounding boxes obtained by deep learning, and obtains workpiece contour and gripping point information.

AB - A workpiece detection method based on fusion of deep learning and image processing is proposed. Firstly, the workpiece bounding boxes are located in the workpiece images by YOLOv3, whose parameters are compressed by an improved convolutional neural network residual structure pruning strategy. Then, the workpiece images are cropped based on the bounding boxes with cropping biases. Finally, the contours and suitable gripping points of the workpieces are obtained through image processing. The experimental results show that mean Average Precision (mAP) is 98.60% for YOLOv3, and 99.38% for that one by pruning 50.89% of its parameters, and the inference time is shortened by 31.13%. Image processing effectively corrects the bounding boxes obtained by deep learning, and obtains workpiece contour and gripping point information.

KW - manufacturing and industrial applications

KW - applications of deep networks

KW - workpieces detection

KW - deep learning

KW - pruning filters

KW - image processing

UR - https://wcci2020.org/

M3 - Paper

T2 - IEEE World Congress on Computational Intelligence 2020

Y2 - 19 July 2020 through 24 July 2020

ER -