شناسايي خودروهاي اضطراري مبتني بر يادگيري عميق به منظور استفاده در خودروهاي بدون راننده

Emergency vehicles recognition based on deep learning for driver-less cars

ﻫﺪف از ﻃﺮاﺣﯽ و ﺳﺎﺧﺖ ﺧﻮدروﻫﺎي ﺑﺪون راﻧﻨﺪه ﺣﺬف ﻋﺎﻣﻞ اﻧﺴﺎﻧﯽ ﺑﻪ ﻣﻨﻈﻮر ﮐﺎﻫﺶ ﺗﻠﻔﺎت، ﻫﺰﻳﻨﻪﻫﺎ و ﻧﻴﺰ اﻓﺰاﯾﺶ اﯾﻤﻨﯽ ﺧﻮدرو ﺑﺎ ﺟﺎﯾﮕﺰﯾﻨﯽ ﺗﺠﻬﯿﺰات ﻫﻮﺷﻤﻨﺪ اﺳﺖ. اﻣﺮوزه ﺑﺎ ﺑﻬﺮهﻣﻨﺪي از ﻓﻨﺎوريﻫﺎي ﻫﻮش ﻣﺼﻨﻮﻋﯽ و ﯾﺎدﮔﯿﺮي ﻣﺎﺷﯿﻦ ﺷﺎﻫﺪ ﭘﯿﺸﺮﻓﺖﻫﺎي ﭼﺸﻢ- ﮔﯿﺮي در ﺻﻨﻌﺖ ﺣﻤﻞ و ﻧﻘﻞ ﻫﻮﺷﻤﻨﺪ ﺑﻪ وﯾﮋه ﺧﻮدروﻫﺎي ﺗﻤﺎم ﺧﻮدﮐﺎر ﻫﺴﺘﯿﻢ ﮐﻪ ﺑﺎ اﺳﺘﻔﺎده از ﺣﺴﮕﺮﻫﺎي ﭘﯿﺸﺮﻓﺘﻪ و ﺗﮑﻨﯿﮏ ﺑﯿﻨﺎﯾﯽ ﻣﺎﺷﯿﻦ ﻗﺎدر ﺑﻪ ﺗﺠﺰﯾﻪ و ﺗﺤﻠﯿﻞ اﻃﻼﻋﺎت ﻣﺤﯿﻂ ﭘﯿﺮاﻣﻮن ﺧﻮد ﻫﺴﺘﻨﺪ. از ﭼﺎﻟﺶﻫﺎي ﻣﻄﺮح در ﻃﺮاﺣﯽ ﺳﯿﺴﺘﻢ اﯾﻦ ﻧﻮع از ﺧﻮدروﻫﺎ، ﺷﻨﺎﺳﺎﯾﯽ درﺳﺖ ﺳﺎﯾﺮ وﺳﺎﯾﻞ ﻧﻘﻠﯿﻪي اﻃﺮاف ﻣﺴﯿﺮ ﺣﺮﮐﺖ ﺧﻮدرو اﺳﺖ. در اﯾﻦ ﻣﻘﺎﻟﻪ، ﺑﺮاي ﺷﻨﺎﺳﺎﯾﯽ ﺧﻮدروﻫﺎي اﺿﻄﺮاري ﻳﻚ روش ﻣﺒﺘﻨﯽ ﺑﺮ ﻳﺎدﮔﻴﺮي ﻋﻤﻴﻖ اراﺋﻪ ﺷﺪه اﺳﺖ ﮐﻪ ﻓﺮاﯾﻨﺪﻫﺎي اﺳﺘﺨﺮاج وﻳﮋﮔﯽ و ﻃﺒﻘﻪﺑﻨﺪي درآن ﺑﻪ ﺻﻮرت ﻫﻤﺰﻣﺎن اﻧﺠﺎم ﻣﯽﺷﻮد. ﺷﺒﮑﻪ ﻋﻤﯿﻖ ﻣﻮرد اﺳﺘﻔﺎده در اﯾﻦ ﭘﮋوﻫﺶ ﺷﺒﮑﻪ ﭘﯿﭽﺸﯽ ﻣﯽﺑﺎﺷﺪ. در ﺷﺒﮑﻪﻫﺎي ﻋﺼﺒﯽ ﭘﯿﭽﺸﯽ دﺳﺘﻴﺎﺑﯽ ﺑﻪ ﻧﺘﺎﻳﺞ ﻗﺎﺑﻞ ﻗﺒﻮل وﻋﻤﻠﮑﺮد ﻣﻨﺎﺳﺐ، ﻣﺴﺘﻠﺰم در اﺧﺘﻴﺎر داﺷﺘﻦ ﺣﺠﻢ ﻋﻈﻴﻤﯽ از دادهﻫﺎ ﺑﺮاي آﻣﻮزش ﺷﺒﮑﻪ ﻣﯽﺑﺎﺷﺪ. ﺑﺎ ﺗﻮﺟﻪ ﺑﻪ ﻣﺤﺪود ﺑﻮدن ﺗﻌﺪاد ﺗﺼﺎوﻳﺮ ﻣﻮﺟﻮد در ﻣﺠﻤﻮﻋﻪ داده ﻣﻮرد اﺳﺘﻔﺎده در اﻳﻦ ﭘﮋوﻫﺶ و ﺑﻪ ﻣﻨﻈﻮر اﻓﺰاﻳﺶ دﻗﺖ ﺷﻨﺎﺳﺎﯾﯽ، از ﻓﺮاﻳﻨﺪ ﻳﺎدﮔﻴﺮي اﻧﺘﻘﺎﻟﯽ و ﺷﺒﮑﻪ ﭘﻴﺶآﻣﻮزشدﻳﺪه VGG16 ﻧﻴﺰ اﺳﺘﻔﺎده ﺷﺪه اﺳﺖ. ﺑﺮاي اﻳﻦ ﺗﺤﻘﻴﻖ دو ﻣﺠﻤﻮع داده ﺟﺪﻳﺪ اﻳﺠﺎد و در ﻛﻨﺎر دو ﻣﺠﻤﻮﻋﻪ داده دﻳﮕﺮ ﻣﻮرد آزﻣﺎﻳﺶ ﻗﺮار ﮔﺮﻓﺖ. روش ﭘﻴﺸﻨﻬﺎدي ﺑﺎ ﭼﻬﺎر روش دﻳﮕﺮ ﻧﻴﺰ ﻣﻮرد ﻣﻘﺎﻳﺴﻪ ﻗﺮار ﮔﺮﻓﺖ و ﻧﺘﺎﻳﺞ ﺑﻪ دﺳﺖ آﻣﺪه ﻧﻤﺎﻳﺎﻧﮕﺮ ﻛﺎراﻳﯽ ﺑﺴﻴﺎر ﺧﻮب روش ﭘﻴﺸﻨﻬﺎدي اﺳﺖ.

The purpose of design and building autonomous cars is to eliminate the human factor in order to reduce losses and costs and also increase safety by replacing smart equipment. Todays, using artificial intelligence and machine learning, we are witnessing significant advances in the intelligent transportations, especially fully automated vehicles, which are able to analyze environmental information using advanced sensors and machine vision techniques. One of the challenges in designing such systems is a correct identification of other vehicles around the route of the vehicle. In this paper, a deep learning-based method for identifying emergency vehicles is presented in which feature extraction and classification processes are performed simultaneously. The deep network used in this research is a convolutional network. In Convolutional Neural Networks (CNN), achieving acceptable results and proper performance requires having a huge amount of data for network training. Due to the limited number of images in the data set used in this study and in order to increase the identification accuracy, transfer learning process and VGG16 pre-trained network have been used. Two new datasets were created for this study and furthermore two other known datasets were also examined. The proposed method was compared with four other known methods from the literature, where the final results showed supremacy of the proposed approach.