The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
A detecção de caminhão de sujeira descoberto visa detectar o caminhão de sujeira e distinguir se ele está coberto ou não por uma rede à prova de poeira para rastrear a fonte de poluição. Ao contrário do problema de detecção tradicional, o recall de todos os caminhões descobertos é mais importante do que a localização precisa para rastreabilidade da poluição. Quando dois objetos estão muito próximos em uma imagem, o objeto ocluído pode não ser recuperado porque o algoritmo de supressão não máxima (NMS) pode remover a proposta sobreposta. Para resolver esse problema, propomos um método Location First NMS para combinar as caixas de verdade e as caixas previstas por posição, em vez de identificador de classe (ID) no estágio de treinamento. Em primeiro lugar, um método de correspondência de caixa é introduzido para reatribuir o ID da caixa prevista usando o mais próximo da verdade, o que pode evitar a falta de objetos quando o IoU de duas propostas for maior que o limite. Em segundo lugar, projetamos uma função de perda para adaptar o algoritmo proposto. Em terceiro lugar, um sistema de detecção de caminhão de lixo descoberto é projetado usando o método em uma cena real. Os resultados do experimento mostram a eficácia do método proposto.
Yuxiang ZHANG
Qingdao University of Technology
Dehua LIU
the Qingdao Xizheng Technology Co., Ltd.
Chuanpeng SU
the Qingdao Xizheng Technology Co., Ltd.
Juncheng LIU
the Qingdao Xizheng Technology Co., Ltd.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Yuxiang ZHANG, Dehua LIU, Chuanpeng SU, Juncheng LIU, "Location First Non-Maximum Suppression for Uncovered Muck Truck Detection" in IEICE TRANSACTIONS on Fundamentals,
vol. E106-A, no. 6, pp. 924-931, June 2023, doi: 10.1587/transfun.2022EAP1100.
Abstract: Uncovered muck truck detection aims to detect the muck truck and distinguish whether it is covered or not by dust-proof net to trace the source of pollution. Unlike traditional detection problem, recalling all uncovered trucks is more important than accurate locating for pollution traceability. When two objects are very close in an image, the occluded object may not be recalled because the non-maximum suppression (NMS) algorithm can remove the overlapped proposal. To address this issue, we propose a Location First NMS method to match the ground truth boxes and predicted boxes by position rather than class identifier (ID) in the training stage. Firstly, a box matching method is introduced to re-assign the predicted box ID using the closest ground truth one, which can avoid object missing when the IoU of two proposals is greater than the threshold. Secondly, we design a loss function to adapt the proposed algorithm. Thirdly, a uncovered muck truck detection system is designed using the method in a real scene. Experiment results show the effectiveness of the proposed method.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2022EAP1100/_p
Copiar
@ARTICLE{e106-a_6_924,
author={Yuxiang ZHANG, Dehua LIU, Chuanpeng SU, Juncheng LIU, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Location First Non-Maximum Suppression for Uncovered Muck Truck Detection},
year={2023},
volume={E106-A},
number={6},
pages={924-931},
abstract={Uncovered muck truck detection aims to detect the muck truck and distinguish whether it is covered or not by dust-proof net to trace the source of pollution. Unlike traditional detection problem, recalling all uncovered trucks is more important than accurate locating for pollution traceability. When two objects are very close in an image, the occluded object may not be recalled because the non-maximum suppression (NMS) algorithm can remove the overlapped proposal. To address this issue, we propose a Location First NMS method to match the ground truth boxes and predicted boxes by position rather than class identifier (ID) in the training stage. Firstly, a box matching method is introduced to re-assign the predicted box ID using the closest ground truth one, which can avoid object missing when the IoU of two proposals is greater than the threshold. Secondly, we design a loss function to adapt the proposed algorithm. Thirdly, a uncovered muck truck detection system is designed using the method in a real scene. Experiment results show the effectiveness of the proposed method.},
keywords={},
doi={10.1587/transfun.2022EAP1100},
ISSN={1745-1337},
month={June},}
Copiar
TY - JOUR
TI - Location First Non-Maximum Suppression for Uncovered Muck Truck Detection
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 924
EP - 931
AU - Yuxiang ZHANG
AU - Dehua LIU
AU - Chuanpeng SU
AU - Juncheng LIU
PY - 2023
DO - 10.1587/transfun.2022EAP1100
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E106-A
IS - 6
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - June 2023
AB - Uncovered muck truck detection aims to detect the muck truck and distinguish whether it is covered or not by dust-proof net to trace the source of pollution. Unlike traditional detection problem, recalling all uncovered trucks is more important than accurate locating for pollution traceability. When two objects are very close in an image, the occluded object may not be recalled because the non-maximum suppression (NMS) algorithm can remove the overlapped proposal. To address this issue, we propose a Location First NMS method to match the ground truth boxes and predicted boxes by position rather than class identifier (ID) in the training stage. Firstly, a box matching method is introduced to re-assign the predicted box ID using the closest ground truth one, which can avoid object missing when the IoU of two proposals is greater than the threshold. Secondly, we design a loss function to adapt the proposed algorithm. Thirdly, a uncovered muck truck detection system is designed using the method in a real scene. Experiment results show the effectiveness of the proposed method.
ER -