EfficientDet for fabric defect detection based on edge computing

Song, Shaojun; Jing, Junfeng; Huang, Yanqing; Shi, Mingyang

doi:10.1177/15589250211008346

Cited by 29 publications

(18 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we present a brief overview of the four state-of-the-art object detectors that we tested, as well as the mathematical background of the technique that we applied for geolocating the detected objects. The selected models are representative of the recent trends in the families of one-stage and two-stage object detectors and have been proven successful in terms of average precision and inference speed in a wide variety of applications [20][21][22][23][24][25].…”

Section: Theoretical Overviewmentioning

confidence: 99%

Deep-Learning-Based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images

2021

View full text Add to dashboard Cite

In this paper, we propose an original deep learning framework for the automated counting and geolocation of palm trees from aerial images using convolutional neural networks. For this purpose, we collected aerial images from two different regions in Saudi Arabia, using two DJI drones, and we built a dataset of around 11,000 instances of palm trees. Then, we applied several recent convolutional neural network models (Faster R-CNN, YOLOv3, YOLOv4, and EfficientDet) to detect palms and other trees, and we conducted a complete comparative evaluation in terms of average precision and inference speed. YOLOv4 and EfficientDet-D5 yielded the best trade-off between accuracy and speed (up to 99% mean average precision and 7.4 FPS). Furthermore, using the geotagged metadata of aerial images, we used photogrammetry concepts and distance corrections to automatically detect the geographical location of detected palm trees. This geolocation technique was tested on two different types of drones (DJI Mavic Pro and Phantom 4 pro) and was assessed to provide an average geolocation accuracy that attains 1.6 m. This GPS tagging allows us to uniquely identify palm trees and count their number from a series of drone images, while correctly dealing with the issue of image overlapping. Moreover, this innovative combination between deep learning object detection and geolocalization can be generalized to any other objects in UAV images.

show abstract

Section: Theoretical Overviewmentioning

confidence: 99%

Deep-Learning-Based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images

2021

View full text Add to dashboard Cite

show abstract

“…Song et al 39 ported the EfficientDet network to TX2 and used a lightweight network to achieve real-time defect detection. Haut et al 40 used the low-power TX2 to classify hyperspectral images with a deep learning algorithm. The deep learning algorithm classifies the hyperspectral images and achieves promising results.…”

Section: Edge Computingmentioning

confidence: 99%

YOLOV4-TinyS: a new convolutional neural architecture for real-time detection of fabric defects in edge devices

Yang,

Jing,

Wang

et al. 2023

Textile Research Journal

Self Cite

View full text Add to dashboard Cite

Real-time detection of fabric defects is a fairly critical part of industrial production. However, there are still some key issues to be solved in practical detection production, such as low detection speed and delays in traditional cloud detection. To address these issues, in this paper, a new detection network architecture, called YOLOV4-TinyS, is proposed. Firstly, the k-medoids clustering algorithm is used to improve the matching of anchor points and ground truths for datasets with great differences. Secondly, the residual structure is changed to reduce the complexity of the network structure, and a depth-separable convolution is used instead of partial convolution to improve the detection speed. Thirdly, the output feature layer is designed with shallow feature fusion to improve the location information extraction capability and use spatial attention and channel attention to improve the network efficiency. Finally, the whole network is trained and tested on four different datasets and extensive experiments show that the network has higher detection accuracy and faster detection speed compared to existing methods. Compared to the original network, YOLOV4-Tiny, the model complexity is reduced by 67.86% and the highest detection accuracy of 99.91% is achieved. Furthermore, the establishment of an efficient fabric inspection system and the validation of the method allows for the fast detection of fabric defects on conveyor belts. Thus, the proposed method has the potential to lay the foundation for the real-time detection of fabric defects and their application in industry.

show abstract

“…The average detection accuracy of this method for hydroponic lettuce seedlings is 86.2%. Song et al (2021) use EfficientDet‐D0, a faster detection algorithm, and deploy the trained model to the edge device, the NVIDIA Jetson TX2.…”

Section: Introductionmentioning

confidence: 99%

Real‐time detection of Angelica dahurica tablet using YOLOX_am

Qin,

Li,

Yan

et al. 2023

J Food Process Engineering

View full text Add to dashboard Cite

In the production of Angelica dahurica tablet (ADT), the manual sorting approach often leads to inefficiency, inconsistent standards, and subjective grading results. The traditional machine vision‐based sorting method, while helping to reduce the demand for labor in factories, suffers from problems such as incomplete contour detection and poor classification of dahurica tablets. To address the above problems, this paper proposes YOLOX_am, a novel deep learning‐based network that combines the fast detection ability of YOLOX and the feature weighting ability of the attention mechanism. In addition, a real‐time sorting system for dahurica tablets is also built and YOLOX_am is deployed in it. The experimental results show that the mAP of ADT's detection using the proposed model reaches 83.53%, which outperforms the original network YOLOX by 4.88%. Moreover, the detection speed of YOLOX_am reaches 1390 ms per image, which meets the requirement of real‐time sorting. Therefore, the combination of YOLOX and the attention mechanism is feasible and effective. YOLOX_am is both fast and accurate for ADT's detection and can be deployed in the sorting system to meet actual production needs.Practical applicationsThe traditional classification of defects in herbal tablets mainly relies on manual detection. This paper introduces a deep learning method into the sorting of herbal tablets. To solve the problem of subtle differentiation of tablets' defects, a neural network structure that can quickly and accurately detect small defective samples is proposed. This network has been applied to practical production and achieved high effectiveness.

show abstract

EfficientDet for fabric defect detection based on edge computing

Cited by 29 publications

References 30 publications

Deep-Learning-Based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images

Deep-Learning-Based Automated Palm Tree Counting and Geolocation in Large Farms from Aerial Geotagged Images

YOLOV4-TinyS: a new convolutional neural architecture for real-time detection of fabric defects in edge devices

Real‐time detection of Angelica dahurica tablet using YOLOX_am

Contact Info

Product

Resources

About