Õª Òª
¡¡¡¡°²È«Ã±×÷Ϊʩ¹¤ÈËÔ±·À»¤Í·²¿µÄ¹¤¾ß£¬¶Ô¹¤È˵ÄÈËÉí°²È«ÓÐ׿« ÆäÖØÒªµÄ×÷Óá£ÏÖ´æµÄ°²È«Ã±Åå´÷״̬¼ì²âÖ÷ÒªÊÇ»ùÓÚÈ˹¤¼à¿Ø£¬ÈË ¹¤¼à¿ØÒªÇó¼à²âÈËԱʱ¿Ì±£³ÖÔÚ¸Ú״̬£¬µ«ÊÇÈ˹¤¼à¿ØÈÝÒ׳öÏÖÊè ©£¬´æÔÚ©¼ìÇé¿ö¡£»ùÓÚ¼ÆËã»úÊÓ¾õµÄ°²È«Ã±Åå´÷״̬¼ì²â¿ÉÒÔ´úÌæ È˹¤¼à¿Ø¹¤×÷¡£ÔÚ²»Í¬Ê©¹¤³¡¾°Ï¼ì²â°²È«Ã±µÄÅå´÷Çé¿öÒѳÉΪÈ˹¤ ÖÇÄÜ¡¢Ä£Ê½Ê¶±ðµÈÁìÓòµÄÑо¿Èȵ㣬ÓÈÆäÊÇÔÚ¼ÆËã»úÊÓ¾õÁìÓò¡£Õë¶Ô °²È«Ã±Åå´÷״̬¼ì²âµÄ¸´ÔÓÐÔ£¬±¾ÎÄ´ÓͼÏñÌØÕ÷ÃèÊö¡¢ÌØÕ÷ÌáÈ¡¡¢Ä¿ ±ê¼ì²âµÈ»·½ÚÌá³öÁËÁ½ÖÖ¼ì²âÑо¿·½·¨¡£±¾ÎľßÌåÑо¿¹¤×÷ºÍ´´Ðµ㠰üÀ¨ÒÔÏÂÁ½µã£º
¡¡¡¡Ò»¡¢Ìá³öÒ»ÖָĽøÔöÇ¿Ëæ»úÞ§Ëã·¨£¨Improved Boosted Random Ferns Algorithm, IBRFs£©£¬²¢Ó¦ÓÃÔÚ°²È«Ã±Åå´÷״̬¼ì²âÖС£Õë¶Ô°² ȫñÅå´÷״̬¼ì²âÔÚ¸´ÔӵĽ¨Öþ³¡¾°ÏÂÈÝÒ×ʧЧµÄÎÊÌ⣬±¾ÎÄÒÔËæ»ú Þ§Ëã·¨×÷Ϊ»ù´¡£¬Ê×ÏÈ£¬ÀûÓ÷½ÏòÌݶÈÖ±·½Í¼£¨Histogram of Oriented Gradient, HOG£©¶ÔͼÏñ½øÐÐÌØÕ÷ÌáÈ¡£¬ÐγÉͼÏñµÄÌØÕ÷Óò¿Õ¼ä£»Æä´Î£¬ ÔÚÌØÕ÷Óò¿Õ¼ä²ÉÓÃËæ»ú¶þÔª²âÊԵķ½·¨¹¹½¨Ëæ»úÞ§£»ÔٴΣ¬ÔÚËæ»úÞ§ ÉÏÑ¡ÔñÄ¿±êÓë±³¾°Ö®¼ä´æÔÚÏÔÖø²îÒìµÄλÖõãºÍ²ÎÊý£¬¹¹½¨¾ßÓÐÒ»¶¨ ʶ±ðÄÜÁ¦µÄÈõ·ÖÀàÆ÷£»×îºó£¬²ÉÓà Real AdaBoost Ëã·¨£¬¹¹½¨Ç¿·ÖÀà Æ÷¶ÔͼÏñ½øÐзÖÀàÔ¤²â¡£ÓÉÓÚ Real AdaBoost Ëã·¨ÖзÖÀàÆ÷µÄ¹¹½¨Ö» Êǽ«¶à¸öÈõ·ÖÀàÆ÷½øÐмòµ¥µÄ¼ÓȨ×éºÏ£¬Ã»Óкܺõķ¢»ÓÈõ·ÖÀàÆ÷µÄ ×÷Óã¬Òò´Ë±¾ÎÄÌá³öÁËÒ»ÖָĽøÔöÇ¿Ëæ»úÞ§Ëã·¨£¬ÀûÓüÓȨϵÊý·¨Ñ¡ Ôñ¶à¸öÅбðÐÔ½ÏÇ¿µÄÈõ·ÖÀàÆ÷¹¹½¨Ç¿·ÖÀàÆ÷¡£ÔÚʵʱµÄ°²È«Ã±¼ì²âÈÎÎñÖУ¬±¾ÎÄÌá³öµÄ IBRFs Ëã·¨¼ì²â¾«È·¶È´ïµ½ 92.74%,Ïà¶ÔÓÚ BRFs Ë㷨׼ȷÂÊÌá¸ßÁË 9.5%×óÓÒ£¬¼«´óµØÌá¸ßÁË·ÖÀà׼ȷÂÊ¡£ÔÚÆÕͨ»· ¾³Ï£¬¿ÉÒÔʵÏÖ 8 f / sµÄ¼ì²âËٶȣ¬Âú×ãʵʱ¼ì²âµÄÒªÇó¡£
¡¡¡¡¶þ¡¢¹¹½¨ResNet50-SSD£¨ResNet50-Single Shot Multi-box Detector£© Ä£ÐÍ£¬²¢Ó¦ÓÃÔÚ°²È«Ã±Åå´÷״̬¼ì²âÖС£ÎªÁ˽â¾öʵʱ³¡¾°ÏÂС³ß´ç Ä¿±ê¼ì²â׼ȷÐԽϵ͵ÄÎÊÌ⣬±¾ÎÄÌá³öÁËÒ»ÖÖ»ùÓÚ ResNet50-SSD µÄ °²È«Ã±Åå´÷״̬¼ì²âÄ£ÐÍ¡£¸ÃÄ£ÐÍÒÔ SSD£¨Single Shot Multi-box Detector£©ÍøÂç×÷Ϊ»ù´¡£¬Ê×ÏÈ£¬½«»ù´¡ÍøÂçVGG-16Ì滻ΪResNet-50, ²¢¶Ô ResNet-50 ÖÐ conv4_x µÄµÚÒ»¸ö²Ð²îÍøÂç½á¹¹½øÐÐÐ޸ģ¬Ê¹µÃµÚ Ò»¸öÊä³öµÄÌØÕ÷¾ØÕó´óС²»±ä£»Æä´Î£¬ÔÚ conv4_x Ö®ºóÌí¼ÓһϵÁÐµÄ ¸½¼Ó²ã£¬²¢ÒýÈëÅúÁ¿¹éÒ»»¯£¨Batch Normalization, BN£©²ã£¬¼Ó¿ìÍø ÂçÊÕÁ²Ëٶȣ»ÔٴΣ¬·Ö±ðÔÚ²»Í¬µÄÌØÕ÷²ãÉ϶Բ»Í¬³ß´çµÄÄ¿±ê½øÐÐÔ¤ ²â£»×îºó£¬Í¨¹ý·Ç¼«´óÖµÒÖÖÆËã·¨£¬Â˳ýµôС¸ÅÂʵÄÄ¿±êµÃµ½×îÖÕµÄ ¼ì²â½á¹û¡£ÔÚ¹ú¼Ê¹«¿ªµÄ°²È«Ã±Êý¾Ý¼¯ SHWD ÉÏ£¬±¾ÎÄÌá³öµÄ ResNet50-SSD Ä£Ðͼì²â¾«È·¶ÈΪ 80.4%,Ïà¶ÔÓÚ VGG16-SSD Ä£ÐÍÌá¸ß ÁË 2.23%×óÓÒ£¬¼ì²âËÙ¶ÈΪÿÃë 35 Ö¡£¬Âú×ãʵʱ¼ì²âµÄÒªÇó¡£ ±¾ÎĽ«Ìá³öµÄÁ½ÖÖËã·¨·Ö±ðÔÚʵʱ³¡¾°ÏµÄÊý¾Ý¼¯ºÍ¹ú¼Ê¹«¿ª µÄÊý¾Ý¼¯É϶԰²È«Ã±Åå´÷״̬½øÐÐʵÑé¼ì²â£¬²¢ÓëÏÖ´æµÄÆäËüÏȽøÄ¿ ±ê¼ì²âËã·¨½øÐбȽϲ¢·ÖÎö£¬ÑéÖ¤Á˱¾ÎÄËã·¨ÔÚ¼ì²â¾«È·¶ÈºÍʵʱÐÔ ÉϵÄÓÅÔ½ÐÔ¡£Í¨¹ýʵÑéÑéÖ¤£¬±¾ÎÄÌá³öµÄËã·¨ÔÚÈ·±£½Ï¸ß¼ì²â¾«È·¶È µÄͬʱÒàÄÜÂú×ãʵʱ¼ì²âµÄÒªÇ󣬾ßÓнÏÇ¿µÄ·º»¯ÄÜÁ¦ºÍ³°ôÐÔ£¬ÄÜ ¹»ÊÊÓ¦¸´ÔÓ»·¾³Ïµİ²È«Ã±Åå´÷״̬ʵʱ¼ì²â¡£
¡¡¡¡¹Ø¼ü´Ê£ºÄ¿±ê¼ì²â°²È«£¬Ã±Åå´÷״̬¼ì²â£¬ÔöÇ¿Ëæ»úÞ§¼ÓȨϵÊý£¬ResNet50-SSD
¡¡¡¡Abstract:
¡¡¡¡Safety helmet, as a tool to protect the head of constructor, plays an extremely important role in the personal safety protection of workers. The existing helmet wearing status detection is mainly based on manual monitoring. Manual monitoring requires monitoring personnel to keep on duty at all times, however, manual monitoring is prone to oversight, there is a situation of missing inspection. Safety helmet wearing status detection based on computer vision can replace manual monitoring. Detecting the wearing of safety helmet in different construction scenarios has become a research hot spot in artificial intelligence, pattern recognition and other fields, especially in computer vision field. Aiming at the complexity of detecting the safety helmet wearing status, this paper proposes two detection methods in the process of image feature description, feature extraction, and object detection. The specific research work and innovations of this paper include the following two point
¡¡¡¡1. Propose an Improved Boosted Random Ferns Algorithm £¨IBRFs£© and apply it to the detection of safety helmet wearing status. Aiming at the problem that helmet wearing status detection is easy to fail in complex architectural scenes, this paper takes random ferns algorithm as the basis. Firstly, the image features are extracted by Histogram of Oriented Gradient £¨HOG£© to form the feature domain space of the image. Secondly, random ferns were constructed by random binary test in feature domain space. Then, the position points and parameters with significant difference between object and background were selected on random ferns to construct weak classifiers with certain recognition ability. Finally, the Real AdaBoost algorithm is used to construct a strong classifier for image classification and prediction. Since the construction of classifier in Real AdaBoost algorithm is only a simple weighted combination of multiple weak classifiers and does not play the role of weak classifiers well, an improved enhanced random ferns algorithm is proposed in this paper. The weighted coefficient method is used to select multiple weak classifiers with strong discriminant characteristics to build a strong classifier. In the real-time helmet detection task, the detection accuracy of the IBRFS model proposed in this paper reaches 92.74%, which is about 9.5% higher than that of the BRFS model, and greatly improves the classification accuracy of the classifier. In a normal environment, the detection speed of 8 f / s can be achieved to meet the requirements of real-time detection.
¡¡¡¡2. The ResNet50-SSD £¨ResNet50-Single Shot Multi-box Detector£© network is constructed and applied in the safety helmet wearing status detection. In order to solve the problem of low accuracy of small-scale object detection in real-time scenes, this paper proposes a method of helmet wearing status detection based on ResNet50-SSD network. The model is based on the SSD £¨Single Shot Multi-box Detector£© network. Firstly, the basic network VGG-16 is replaced with ResNet-50, and the structure of the first residual network of conv4_x in ResNet-50 is modified to make the size of the eigenmatrix of the first output constant; Secondly, a series of additional layers are added after conv4_x and the Batch Normalization £¨BN£© layer is introduced to speed up the convergence of the network. Thirdly, the objects of different sizes are predicted in different feature layers. Finally, the non-maximum suppression algorithm is used to filter out the objects with low probability to get the final detection result. On the international pubilc safety helmet dataset SHWD, the detection accuracy of the ResNet50-SSD model proposed in this paper reaches 80.4%, Compared with vgg16-ssd model, it improves by 2.23%, and the detection speed is 35 frames per second, which meets the requirements of real-time detection.
¡¡¡¡The two algorithms proposed in this paper are used to test the wearing status of safety helmets in real-time scenes datasets and international public datasets respectively, and compared with other existing advanced object detection algorithms, it verifies the superiority of the proposed algorithm in detection accuracy and real-time performance. Through experimental verification, the proposed algorithm can not only ensure high detection accuracy, but also meet the requirements of real-time detection. It has strong generalization ability and robustness, and can adapt to real-time detection of safety helmet wearing status in complex environment.
¡¡¡¡Keywords: Object detection Safety helmet wearing status detection Boosted Random Ferns Weighting coefficient ResNet50-SSD
¡¡¡¡Ä¿ ¼
¡¡¡¡1 Ð÷ ÂÛ
¡¡¡¡ÔÚ±¾ÕÂÖУ¬Ê×ÏÈ£¬¶Ô±¾ÎĵÄÑо¿±³¾°¼°ÒâÒå½øÐмòÒª²ûÊö£»Æä´Î£¬¶Ô¹úÄÚÍⰲȫ ñÅå´÷״̬¼ì²âÑо¿µÄ×îÐÂÑо¿ÏÖ×´½øÐвûÊö£¬·ÖÎöÏÖ´æµÄÖ÷ÒªÎÊÌ⣻×îºó£¬½éÉܱ¾ ÎĵÄÖ÷ÒªÑо¿ÄÚÈÝ£¬²¢¶Ô±¾ÎĵÄ×éÖ¯½á¹¹½øÐмòÒª¸ÅÀ¨¡£
¡¡¡¡1.1 Ñо¿±³¾°¼°ÒâÒå
¡¡¡¡1.1.1 Ñо¿±³¾°
¡¡¡¡½üÄêÀ´£¬Ëæ×ųÇÕò»¯½¨Éè¡¢ÐÂÅ©´å½¨ÉèµÄ¼ÓËÙ·¢Õ¹£¬¹ú¼Ò´óÁ¦·¢Õ¹»ù´¡ÉèÊ©½¨É裬 ¸÷ÀàÊ©¹¤×÷Òµ±é²¼´óС³ÇÊС£Éú²ú°²È«Ò²³ÉΪÁ˹ú¼Ò¼°¸÷¼¶Õþ¸®×îΪ¹Ø×¢ºÍÖØÊÓµÄÎÊ Ì⣬Ȼ¶øÐí¶àÊ©¹¤³¡µØÈÔ´æÔÚ×ÅÑÏÖصݲȫÒþ»¼£¬¸ù¾Ýס·¿ºÍ³ÇÏ罨É貿ͳ¼Æ£¬2019 ÄêÈ«¹ú¹²·¢Éú·¿ÎÝÊÐÕþ¹¤³ÌÉú²ú°²È«Ê¹Ê773Æð¡¢ËÀÍö904ÈË£¬¸øÉç»áºÍ¼ÒÍ¥´øÀ´ÁË¾Þ ´óµÄËðʧ£¬½¨ÖþÊ©¹¤ÐÐÒµµÄ°²È«ÐÎÊÆÊ®·ÖÑϾþ[1].Òò´Ë£¬×î´óÏ޶ȵؽµµÍ°²È«Òþ»¼£¬ Ìá¸ßÊ©¹¤ÈËÔ±×ÔÉíµÄ°²È«ÒâʶÆÈÔÚü½Þ¡£ËäȻͨ¹ý°²È«½ÌÓýÅàѵ£¬¿ÉÒÔÔöÇ¿×÷Òµ¹¤ÈË µÄ°²È«Òâʶ£¬Ò»¶¨³Ì¶ÈÉϼõÉÙ°²È«Ê¹ʷ¢ÉúµÄ¸ÅÂÊ£¬µ«ÊÇ£¬Òª´Ó¸ù±¾ÉϽâ¾öÎÊÌ⣬±Ü ÃâÒòΪδÅå´÷°²È«Ã±¶øÔì³ÉµÄ°²È«Ê¹ʣ¬»¹ÊÇÐèÒª¿Æѧ¡¢ºÏÀí¡¢¸ßЧµÄ¼à¿ØºÍ¹ÜÀí[2]. Ëæ×ſƼ¼µÄ·¢Õ¹£¬´ó¶àÊýµÄÊ©¹¤×÷Òµ³¡µØ¶¼°²×°Á˼à¿ØÉ豸£¬¸ºÔð¼à¿Ø¹ÜÀíµÄÈË Ô±»á¶¨Ê±½øÐÐÊÓƵ¼à¿Ø´¦Àí£¬ÊµÏÖ°²È«Ã±Åå´÷״̬µÄʵʱ¼à²â£¬µ«ÊÇÄ¿Ç°µÄ°²È«Ã±¼ì ²âϵͳÒÀ¾É´æÔÚ×ÅÒÔÏÂÎÊÌ⣺
¡¡¡¡£¨1£©¼ì²â¾«È·¶ÈµÍ£ºÄ¿Ç°µÄ°²È«Ã±¼ì²âËã·¨·ÖΪ´«Í³µÄÄ¿±ê¼ì²âËã·¨ºÍÉî¶Èѧ Ï°¼ì²âËã·¨¡£µ«ÊÇÓÉÓÚÊ©¹¤³¡¾°¸´ÔÓ¡¢¼ì²âÄ¿±êÖ®¼ä´æÔÚ²¿·ÖÕÚµ²¡¢ÏÖ³¡ÌìÆø±ä»¯¡¢ ¹âÕÕµÈÓ°Ï죬µ¼Ö°²È«Ã±Åå´÷×´¿öµÄ¼ì²â¾«È·¶È½ÏµÍ¡£
¡¡¡¡£¨2£©¼ì²â³É±¾½Ï¸ß£º¶ÔÕû¸öÊ©¹¤ÏÖ³¡²¿ÊðÉãÏñÍ·£¬¹æÄ£½Ï´óÇÒ¾¼Ã³É±¾½Ï¸ß¡£ ¶øÇÒ£¬Ê©¹¤³¡µØ¹¤×÷·±Ã¦ÇÒ¸´ÔÓ£¬ÈôרÃÅ°²ÅÅÈËÔ±½øÐмà¿Ø£¬¼«´óµØÏûºÄÁËÈËÁ¦¡¢Îï Á¦¡¢²ÆÁ¦¡£¶øÇÒ¼ì²â¹¤È˲»ÔÚ¹¤×÷¸Úλʱ£¬ÈÝÒ׳öÏÖ©¼ìµÄÇé¿ö£¬Ò»µ©³öÏÖÊè©£¬½« »áÔì³É²»¿É¹ÀÁ¿µÄºó¹û¡£
¡¡¡¡1.1.2 Ñо¿ÒâÒå
¡¡¡¡°²È«Ã±×÷Ϊʩ¹¤ÈËÔ±×î»ù±¾µÄ°²È«·À»¤ÓÃÆ·£¬Ö÷Òª×÷ÓÃÊDZ£»¤Ê©¹¤ÏÖ³¡ÈËÔ±µÄÍ· ²¿°²È«£¬·ÀÖ¹±»¼âÈñÎïÌå´ò»÷¡¢Åöײ£¬¶ÔÊ©¹¤ÈËÔ±µÄ¸öÈËÈËÉí°²È«¼°²Æ²ú°²È«ÓÐןÜÖØÒªµÄÒâÒå[3-4].ÔÚÊ©¹¤ÏÖ³¡£¬ÓÉÓÚÊ©¹¤ÈËÔ±²»ÖØÊÓ¡¢È±·¦°²È«Òâʶ¡¢ÕÆÎÕ°²È«ÖªÊ¶ ǷȱµÈÔì³ÉµÄһϵÁÐÎ¥¹æÎ¥¼Í²Ù×÷£¬°²È«Ê¹ʷ¢ÉúµÄÏÖÏóʱÓз¢Éú¡£ÓÉÓÚÊ©¹¤×÷Òµ³¡ µØ»·¾³½ÏΪ¸´ÔÓ¡¢·¶Î§½Ï¹ã¡¢Ê©¹¤×÷ÒµÈËÔ±ÊýÁ¿½Ï¶à¡¢¸ºÔð¼à¶½¹ÜÀíµÄÈËÔ±ÓÐÏÞµÈÔ Òò£¬Òª¶ÔÊ©¹¤ÏÖ³¡¹¤ÈË°²È«Ã±Åå´÷״̬½øÐÐÈ«·½Î»¡¢È«¹ý³ÌµÄ¼ì²âÊǺÜÄÑʵÏֵġ£ ΪÁ˱£Ö¤Ê©¹¤×÷ÒµÈËÔ±µÄÈËÉí²Æ²ú°²È«£¬ÓбØÒª²ÉÓÃÈ˹¤ÖÇÄܵļ¼Êõ·½·¨£¬¼õÉÙ Ê©¹¤ÏÖ³¡ÒòΪûÓÐÅå´÷°²È«Ã±¶øµ¼Öµİ²È«Ê¹ʷ¢Éú¡£±¾¿ÎÌâ»ùÓÚ¼ÆËã»úÊÓ¾õ£¬´ÓÀí ÂÛÉÏ£¬¶ÔÖ÷Á÷µÄÄ¿±ê¼ì²âËã·¨½øÐÐÏêϸ²ûÊö£¬²¢×öÁËÏà¹ØµÄ¸Ä½ø£¬½áºÏͼÏñ´¦Àí¼¼Êõ Ìá³öÁË°²È«Ã±Åå´÷״̬µÄ¼ì²â·½·¨¡£ÊµÏÖÁ˸´ÔÓ×÷Òµ³¡¾°ÏÂÊ©¹¤×÷ÒµÈËÔ±ÊÇ·ñÅå´÷°² ȫñµÄ¾«È·¼ì²â£¬½µµÍÁËÊ©¹¤ÈËÔ±ÓÉÓÚδÅå´÷°²È«Ã±¶øÒý·¢µÄ°²È«Ê¹ʷ¢ÉúÂÊ£¬¶ÔÊ© ¹¤ÏÖ³¡µÄ°²È«¹æ·¶¹ÜÀí¾ßÓÐÒ»¶¨µÄÖ¸µ¼ÒâÒåºÍʵ¼ÊÓ¦ÓüÛÖµ¡£
¡¡¡¡1.2 ¹úÄÚÍâÑо¿ÏÖ×´
¡¡¡¡°²È«Ã±Åå´÷״̬¼ì²âʵ¼ÊΪ¶þ·ÖÀà¶ÔÏó¼ì²âÎÊÌ⣬ΪÁËÌá¸ß°²È«Ã±Åå´÷×´¿ö¼ì²â µÄ¾«È·¶ÈºÍʵʱÐÔ£¬¹úÄÚÍâѧÕß¿ªÕ¹Á˺ܶà¾ßÓд´ÐÂÐÔµÄÑо¿¹¤×÷¡£ÔÚģʽʶ±ðÁìÓò ÖУ¬°²È«Ã±Åå´÷״̬¼ì²â·½·¨Ö÷Òª·ÖΪ´«Í³µÄÄ¿±ê¼ì²âËã·¨ºÍ»ùÓÚÉî¶ÈѧϰµÄ¼ì²âËã ·¨£¬ÎªÊ©¹¤ÏÖ³¡°²È«Ã±Åå´÷״̬¼ì²âÑо¿ÌṩÁ˽ÏÓмÛÖµµÄ½è¼ø¡£
¡¡¡¡1.2.1 »ùÓÚ´«Í³Ä£Ê½Ê¶±ðµÄ°²È«Ã±Åå´÷״̬¼ì²â
¡¡¡¡»ùÓÚ´«Í³Ä¿±ê¼ì²âËã·¨µÄ°²È«Ã±Åå´÷״̬¼ì²â£¬Ê×ÏÈ£¬²ÉÓû¬¶¯´°¿Ú¶ÔͼÏñÉÏÿ ¸öÏñËصã½øÐбéÀú£»Æä´Î£¬¶Ô¼ì²âÄ¿±êµÄijÖÖÌØÐÔ½øÐÐͳ¼ÆÃèÊö£»×îºó£¬Í¨¹ý·ÖÀàµÄ ·½·¨»òÕ߶ÔÌáÈ¡µ½µÄÌØÕ÷½øÐн¨Ä££¬Åж¨¼ì²âÄ¿±êÊÇ·ñÅå´÷°²È«Ã±¡£
¡¡¡¡Rubaiyat µÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚÌØÕ÷ÌáÈ¡µÄ°²È«Ã±Åå´÷¼ì²âËã·¨[5].Ê×ÏÈ£¬½áºÏ·½ ÏòÌݶÈÖ±·½Í¼ HOG ºÍͼÏñµÄƵÓòÐÅÏ¢£¬¶ÔÊ©¹¤ÈËÔ±½øÐмì²â£»Æä´Î£¬½áºÏÑÕÉ«ºÍÔ² »·»ô·ò±ä»»£¨Circle Hough Transform,CHT£©ÌØÕ÷£¬¶Ô°²È«Ã±½øÐмì²â¡£¸Ã·½·¨ÔÚ°² ȫñÅå´÷¼ì²âÈÎÎñÖУ¬Äܹ»ÊµÏÖ°²È«Ã±µÄ¼ì²â£¬µ«ÊǼì²â¾«È·¶È½ÏµÍÇÒÖ»Äܼì²âÌض¨ ÑÕÉ«µÄ°²È«Ã±£¬·º»¯ÄÜÁ¦½Ï²î¡£ÀîäìµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚÖ§³ÖÏòÁ¿»ú£¨Support Vector Machine,SVM£©·ÖÀàÆ÷ÓëÌØÕ÷ÌáÈ¡µÄ°²È«Ã±¼ì²âÄ£ÐÍ[6].Ê×ÏÈ£¬²ÉÓÃÖ¡¼ä²î·Ö·¨¶Ô ÊÓƵͼÏñ½øÐзָÆä´Î£¬ÌáȡͼÏñµÄ HOG ÌØÕ÷£¬²¢Í¨¹ý SVM ·ÖÀàÆ÷ÅжÏͼÏñÖÐ ÊÇ·ñ´æÔÚÐÐÈË£»×îºó£¬¶ÔͼÏñÖеÄÐÐÈËÌáÈ¡ Haar-like ÌØÕ÷£¬²¢²ÉÓà AdaBoost ·ÖÀàÆ÷ ÅжÏÐÐÈËÊÇ·ñÅå´÷°²È«Ã±¡£¸Ã·½·¨ÌáȡͼÏñµÄË«ÖØÌØÕ÷£¬ºÜ´ó³Ì¶ÈÉÏÌá¸ßÁ˼ì²â¾«È· ¶È£¬µ«ÊÇÌØÕ÷ÌáÈ¡µÄ¹ý³Ì½Ï¸´ÔÓ¡¢·º»¯ÄÜÁ¦²î£¬²»ÄÜÂú×ãʵʱ¼ì²â¡£·ë¹ú³¼µÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ»ìºÏ¸ß˹ģÐ͵ݲȫñ×Ô¶¯¼ì²âÄ£ÐÍ[7].Ê×ÏÈ£¬²ÉÓûìºÏ¸ß˹ģÐÍÔÚͼÏñ Ô¤´¦Àí½×¶Î½øÐÐÇ°¾°¼ì²â£»Æä´Î£¬¶ÔÁ¬Í¨Óò½øÐд¦Àí·ÖÎö£¬ÅжϸÃÁ¬Í¨ÓòÊÇ·ñÊôÓÚÈË Ì壬ʵÏÖ¼ì²âÄ¿±êµÄ×Ô¶¯¸ú×Ù£»×îºó£¬ÔÚ¶¨Î»µ½µÄÈËÍ·ÇøÓò½øÐа²È«Ã±Åå´÷״̬µÄ×Ô ¶¯¼ì²âʶ±ð¡£¸Ã·½·¨¼ì²â¾«È·¶È¸ß£¬ÊÜ»·¾³±ä»¯Ó°Ïì½ÏС£¬µ«ÊÇÔÚͼÏñ´¦Àí½×¶ÎºÄʱ ½Ï³¤£¬ÔÚʵ¼ÊÓ¦ÓÃÖкÜÄÑÂú×ãʵʱ¼ì²â¡£ÁõÏþ»ÛµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ SVM ·ÖÀàÆ÷µÄ °²È«Ã±¼ì²âËã·¨[8].
¡¡¡¡Ê×ÏÈ£¬¶ÔͼÏñ½øÐÐÔ¤´¦Àí£¬½¨Á¢°²È«Ã±Êý¾Ý¼¯£»Æä´Î£¬²ÉÓ÷ôÉ« ¼ì²âµÄ·½·¨¶ÔÈËÁ³ÇøÓò½øÐж¨Î»£¬»ñµÃ°²È«Ã±ËùÔÚλÖõĴÖÂÔ¶¨Î»£¬²¢ÊÖ¶¯¼ô²Ã°²È« ñËùÔÚÇøÓòͼÏñ£»½Ó×Å£¬ÌáÈ¡°²È«Ã±µÄ Hu ¾ØÌØÕ÷ÏòÁ¿£»×îºó£¬½áºÏ Hu ¾ØÌØÕ÷ÏòÁ¿ ºÍ SVM ·ÖÀàÄ£Ð͸ø³ö×îÖÕ·ÖÀà½á¹û¡£¸Ã·½·¨Ïà¶ÔÓÚÉñ¾ÍøÂçËõ¶ÌÁËʱ¼ä¡¢Ê¶±ðÂÊÒ² ÓÐÒ»¶¨µÄÌáÉý¡¢¾ßÓнϺõķº»¯ÄÜÁ¦¡£µ«ÊÇ£¬¸Ã·½·¨Ó¦Óó¡¾°µ¥Ò»£¬½öÓÃÓÚ¹¤µØ½ø³ö ¿ÚµÈ¹Ì¶¨³¡¾°£¬¶øÇÒÊÜÊÓ½ÇÓ°Ïì½Ï´ó¡£ÖÜÑÞÇàµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ¾Ö²¿¶þֵģʽ £¨Local Binary Pattern, LBP£©ÌØÕ÷µÄ°²È«Ã±¼ì²âËã·¨[9].Ê×ÏÈ£¬¶ÔÊÓƵͼÏñ½øÐÐÔ¤´¦ Àí£¬µÃµ½´óСΪ22 ? 22µÄÈËÍ·ÇøÓòͼÏñ£»Æä´Î£¬ÌáÈ¡¸Ã²¿·ÖµÄ¾Ö²¿¶þ½øÖÆͳ¼ÆÌØÕ÷£» ×îºó£¬ÀûÓ÷´Ïò´«²¥ÍøÂ磨Back Propagation, BP£©Éñ¾ÍøÂç¶Ô¼ì²âÄ¿±ê½øÐзÖÀàÔ¤²â£¬ µÃµ½×îÖյļì²â½á¹û¡£¸Ã·½·¨¼òµ¥Ò×ÐУ¬µ«ÊÇÓÉÓÚͼÏñ·Ö±æÂʵ͡¢Êµ¼ÊÓ¦Óû·¾³¸´ÔÓ£¬ ËùÒÔ¼ì²â¾«È·¶È»¹ÓдýÌá¸ß¡£
¡¡¡¡ÍõÓêÉúµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ×Ë̬¹À¼ÆµÄ°²È«Ã±Åå´÷״̬ ¼ì²âËã·¨[10].Ê×ÏÈ£¬ÔÚ OpenRose Ä£ÐÍÉÏÒýÈëÁ˲вîÍøÂç¶Ô¼ì²âÄ¿±ê½øÐÐ×Ë̬¹À¼Æ£» Æä´Î£¬Í¨¹ýÈýµã¶¨Î»·¨È·¶¨Í·²¿µÄλÖã¬ËõСĿ±ê¼ì²â·¶Î§£»×îºó£¬Ê¹Óà RetinaNet ÍøÂç¶ÔÍ·²¿ÇøÓò½øÐмì²âÅжϰ²È«Ã±µÄÅå´÷Çé¿ö¡£¸Ã·½·¨ºÜºÃµØ½â¾ö°²È«Ã±ÓëÊ©¹¤±³ ¾°Ö®¼äÀà±ð²»Æ½ºâµÄÎÊÌâ¡£ÓÉÓÚ¼ì²âµÄÍ·²¿ÇøÓòÊÇÈ·¶¨µÄ£¬ÑµÁ·Ê±¼ì²âЧ¹û½ÏºÃ£¬µ« ÊÇÔÚʵ¼ÊÓ¦ÓÃÖзº»¯ÄÜÁ¦½ÏÈõ£¬¶Ô½ÏС³ß´çµÄÄ¿±ê¼ì²âЧ¹û²»ÀíÏë¡£Park µÈÈËÌá³öÁË Ò»ÖÖ»ùÓÚ±³¾°²î·Ö·¨µÄ°²È«Ã±¼ì²âËã·¨[11].
¡¡¡¡Ê×ÏÈ£¬ÀûÓñ³¾°²î·Ö·¨ÌáÈ¡Ç°¾°£¬»ùÓÚͼ ÏñµÄ HOG ÓòÌáÈ¡Ä¿±ê±ßÔµÌØÕ÷£»Æä´Î£¬Í¨¹ýÖ§³ÖÏòÁ¿»úʵÏÖÈËÌåºÍ°²È«Ã±µÄ¼ì²â£» ×îºó£¬¸ù¾ÝÈËÌåºÍ°²È«Ã±µÄ¿Õ¼ä¼¸ºÎ¹Øϵȷ¶¨¹¤×÷ÈËÔ±ÊÇ·ñÅå´÷°²È«Ã±¡£µ«ÊÇ£¬µ±¼ì ²âÄ¿±ê±»²¿·ÖÕÚµ²£¬»òÕß´¦ÓÚ·ÇÕ¾Á¢¡¢¾²Ö¹²»¶¯µÄʱºò£¬¸Ã·½·¨´æÔÚÒ»¶¨µÄ¾ÖÏÞÐÔ¡£ ×ÛÉÏËùÊö£¬ÏÖ´æµÄ»ùÓÚ´«Í³Ä£Ê½Ê¶±ðµÄ°²È«Ã±Åå´÷״̬¼ì²â£¬ÔÚʵ¼ÊÓ¦ÓÃÖдæÔÚ ¼ì²âЧ¹û²î¡¢¼ì²âËÙ¶ÈÏà¶Ô½ÏÂý¡¢Ä£ÐÍ·º»¯ÄÜÁ¦²îµÈÎÊÌâ¡£ÔÚʵ¼ÊÓ¦ÓÃÖв»ÄÜÂú×ãʵ ʱ¼à²âµÄÒªÇó¡£
¡¡¡¡1.2.2 »ùÓÚÉî¶ÈѧϰµÄ°²È«Ã±Åå´÷״̬¼ì²â
¡¡¡¡»ùÓÚÉî¶ÈѧϰµÄÄ¿±ê¼ì²âËã·¨ÊÇÏÖ´æµÄÖ÷Á÷·½·¨£¬Í¨¹ý¶à´ÎµÄ¾í»ý¡¢³Ø»¯´¦Àí£¬ ½«µÃµ½µÄÌØÕ÷¾ØÕóÊäÈë·ÖÀàÆ÷½øÐмì²âÅж¨[12-17].
¡¡¡¡Ê×ÏÈ£¬Í¨¹ý¾í»ý²ãʵÏÖÄ¿±êµÄÌØÕ÷ ÌáÈ¡£»Æä´Î£¬½«ÌáÈ¡µ½µÄÌØÕ÷ÊäÈë·ÖÀàÆ÷ÖУ¬µÃµ½×îÖյķÖÀà½á¹û¡£ ÐìÏÈ·åµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ MobileNet-SSD µÄ°²È«Ã±Åå´÷Çé¿ö¼ì²â·½·¨[18].Ê× ÏÈ£¬¸ÃÍøÂçÒýÈë MobileNet ÍøÂçÌæ»»µ¥µã¶à¿ò¼ì²â£¨Single Shot Multi-box Detector, SSD£©ÍøÂçÖÐµÄ backbone,¹¹½¨¸ü¼ÓÇáÁ¿»¯µÄÍøÂ磬Ìá¸ßÄ£Ð͵ļì²âËٶȣ»Æä´Î£¬Òý ÈëǨÒÆѧϰ²ßÂÔ½â¾öÄ£ÐÍѵÁ·À§ÄѵÄÎÊÌâ¡£ÔÚ°²È«Ã±¼ì²âÈÎÎñÖУ¬MobileNet-SSD Ïà ±È½ÏÓÚ SSD,ÔÚ CPU ƽ̨ϵļì²âËٶȵõ½Á˺ܴóµÄÌáÉý£¬µ«ÊǼì²â¾«È·¶ÈÓÐÒ»¶¨ µÄËðʧ¡£³ÂÁøµÈÈËÌá³öÒ»ÖÖ»ùÓÚÊÓ¾õ¸ÐÊÜÒ°µÄÇáÁ¿»¯¾í»ýÉñ¾ÍøÂç¶Ô°²È«Ã±Åå´÷Çé¿ö ½øÐмì²âÑо¿[19].Ê×ÏÈ£¬¸ÃÍøÂçÔÚ RFNet ÍøÂçµÄ»ù´¡ÉÏÌí¼ÓÁËÌØÕ÷½ð×ÖËþÍøÂçÄ£¿é£» Æä´Î£¬ÔÚ±£Ö¤ÌØÕ÷ÌáÈ¡ÄÜÁ¦²»±äµÄÇé¿öÏ£¬²ÉÓà SE-Ghost Ä£¿é¶ÔÖ÷¸ÉÍøÂç½øÐÐÇáÁ¿ »¯´¦Àí£»×îºó£¬ÔÚ¸´ÔÓµÄÊ©¹¤»·¾³Ï¶Բ»Í¬ÐÎ̬²»Í¬³ß¶È´óСµÄ°²È«Ã±½øÐÐʶ±ð¼ì²â¡£
¡¡¡¡Óë¾í»ýÉñ¾ÍøÂçÏà±È£¬¸Ã·½·¨ÔÚÈ·±£¼ì²â¾«È·¶ÈµÄͬʱ£¬²ÉÓÃÁËÇáÁ¿»¯µÄÄ£¿é£¬¼«´ó µØËõСÁËÍøÂç´¢´æ¿Õ¼ä£¬Ìá¸ßÁ˼ì²âËٶȡ£ÅÓÊâÑîµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ¶àÈÎÎñ¼¶Áª¾í »ýÉñ¾ÍøÂ磨Multitask Cascaded Convolutional Neural Networks, MTCNN£©ÍøÂçµÄ°²È« ñʶ±ð·½·¨[20].Ê×ÏÈ£¬¶Ô MTCNN ÍøÂç½øÐмò»¯£»Æä´Î£¬²ÉÓÃÆÕͨ¾í»ý²ã¹¹½¨È«¾í»ý ÍøÂ磬Ìá¸ßÄ£Ðͼì²â¾«È·¶È£»ÔٴΣ¬ÒýÈëÇáÁ¿»¯ÍøÂç½á¹¹ MobileNet ¶ÔÍøÂç½øÐÐÓÅ»¯ ´¦Àí£¬¼õСÍøÂç´¢´æ²ÎÊý£¬¼Ó¿ìÔËÐÐËٶȣ»×îºó£¬µ÷ÕûÍøÂç²ÎÊý¶Ô²»Í¬³ß´çµÄ°²È«Ã± ½øÐÐʶ±ð¼ì²â¡£ÁÖ¿¡µÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ YOLO£¨You Only Look Once, YOLO£©µÄ°² ȫñÅå´÷¼ì²âËã·¨[21].
¡¡¡¡Ê×ÏÈ£¬½« YOLO Ó¦Óõ½Î´Åå´÷°²È«Ã±µÄµ¥Àà¼ì²âÎÊÌâÖУ¬¶Ô YOLO ÍøÂç½øÐÐÐ޸ģ¬½«Êä³ö¸ÄΪ 18 άµÄÕÅÁ¿£»Æä´Î£¬¸ù¾ÝËðʧº¯ÊýºÍ IOU ÇúÏß¶Ô Ä£ÐͽøÐе÷Õû£¬µÃµ½×îÖյݲȫñ¼ì²âÄ£ÐÍ¡£µ«ÊÇ£¬¸ÃËã·¨ÔÚÄ¿±êÖصþ¡¢Ãܼ¯Ä¿±ê³¡ ¾°ÏÂÒÀÈ»»á³öÏÖ©¼ìÇé¿ö¡£·½Ã÷µÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ YOLOv2 µÄ¿ìËÙ°²È«Ã±¼ì²âËã ·¨[22].Ê×ÏÈ£¬ÔÚ YOLOv2 ÖмÓÈëÃܼ¯¿é£¬Ìá¸ßÁ˶ÔСĿ±ê¼ì²âµÄÃô¸Ð¶È£»Æä´Î£¬²É ÓÃÇáÁ¿»¯ÍøÂç½á¹¹¶ÔÍøÂç½øÐÐѹËõ£¬¼õСÁ˲ÎÊýÁ¿£¬Ôö¼ÓÁËÄ£Ð͵ĿÉÓÃÐÔ¡£¸Ä½øºóµÄ YOLOv2 Ä£ÐÍÔÚ±£Ö¤Ò»¶¨µÄ¼ì²â׼ȷÂʵÄͬʱ£¬ºÜ´ó³Ì¶ÈÉÏÌáÉýÁ˼ì²âËٶȣ¬µ«ÊÇ£¬ µ±¼ì²âÄ¿±êºÍ±³¾°ÑÕÉ«Ïà²î²»´óʱ£¬¸Ä½øºóµÄ YOLOv2 Ä£ÐͶÔÑÕÉ«µÄÊÊÓ¦ÐԽϵͣ¬ »á´æÔÚ©¼ìÏÖÏó¡£ÐìÊØÀ¤µÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚÓïÒå¹æÔòºÍÓïÒåÄ£¿éµÄ YOLOv3 ¼ì²âËã·¨¶Ô°²È«Ã±Åå´÷Çé¿ö½øÐмì²âÑо¿[23].Ê×ÏÈ£¬¶Ô²É¼¯µÄÊý¾Ý½øÐÐÔ¤´¦Àí£¬µÃµ½°²È« ñÅå´÷¼ì²âÊý¾Ý¼¯ºÍͼÏñ×ÖÄ»Êý¾Ý¼¯£»Æä´Î£¬Ê¹Óà K-means Ë㷨ȷ¶¨Êý¾Ý¼¯µÄê¿ò²Î ÊýÖµ£¬²¢Ê¹Óà YOLOv3 ÍøÂç½øÐÐѵÁ·¼ì²â£»ÔٴΣ¬¶¨ÒåÓïÒå¹æÔò²¢½áºÏÄ¿±ê¼ì²â½á ¹û£¬ÌáÈ¡³öÊÓ¾õ¸ÅÄ×îºó£¬¶ÔͼÏñ×ÖÄ»½øÐÐÌî³äµÃµ½¶ÔÓ¦µÄÓï¾äÄ£°å£¬²¢Éú³É°²È« ñÅå´÷µÄͼÏñÃèÊöÓï¾ä¡£Ê©»ÔµÈÈËÌá³öÁËÒ»ÖÖ»ùÓÚ YOLOv3 Ëã·¨µÄ°²È«Ã±Åå´÷Çé¿ö ¼ì²â·½·¨[24].Ê×ÏÈ£¬Í¨¹ýͼÏñ½ð×ÖËþ½á¹¹»ñµÃ²»Í¬³ß¶ÈµÄÌØÕ÷ͼ£»Æä´Î£¬¶ÔÄ¿±ê¼ì²â ¿ò½øÐоÛÀà´¦Àí£¬²¢È·¶¨Ä¿±ê¼ì²â¿ò²ÎÊý£»×îºó£¬½«¸Ä½øµÄ YOLOv3 Ä£ÐÍÓ¦Óõ½Êµ ¼Ê³¡¾°ÖУ¬²¢ÔÚѵÁ·¹ý³ÌÖв»¶Ïµ÷ÕûÊäÈëͼÏñµÄ³ß´ç£¬Ôö¼ÓÄ£Ð͵ķº»¯ÄÜÁ¦¡£
¡¡¡¡½á¹ûÏÔ Ê¾¸Ä½øºóµÄ YOLOv3 Ëã·¨ÔÚ¼ì²â׼ȷÂÊÓë¼ì²âËÙÂÊÉÏÏà½ÏÓÚ YOLOv3 ¾ùÂÔÓÐÌá¸ß£¬ ÔÚ±£Ö¤½Ï¸ß¾«È·¶ÈµÄͬʱÂú×ãÁË°²È«Ã±Åå´÷¼ì²âµÄʵʱÐÔ¡£ÐìÊØÀ¤µÈÈËÕë¶Ô°²È«Ã±¼ì ²âÈÎÎñÖдæÔڵIJ¿·ÖÕÚµ²¡¢¼ì²âÄ¿±ê³ß´ç²»Ò»ÒÔ¼°Ð¡Ä¿±ê¼ì²âÀ§ÄѵÈÎÊÌ⣬Ìá³öÁËÒ» ÖָĽøµÄ Faster RCNN ºÍ¶à²¿¼þ½áºÏµÄ°²È«Ã±Åå´÷¼ì²âËã·¨[25].Ê×ÏÈ£¬ÒÔ Faster RCNN ÍøÂç×÷Ϊ»ù´¡£¬Ê¹Óöà³ß¶ÈѵÁ·¡¢Ôö¼ÓêµãÊýÁ¿¶Ô Faster RCNN ÍøÂçÄ£ÐͽøÐÐÓÅ»¯£¬ ÔöÇ¿ÍøÂç¶Ô²»Í¬³ß´çÄ¿±ê¼ì²âµÄ³°ôÐÔ£»Æä´Î£¬ÒýÈëÔÚÏßÀ§ÄÑÑù±¾ÍÚ¾ò²ßÂÔ£¬½â¾öÕý ¸ºÑù±¾²»¾ùºâÎÊÌâ¡£¸Ã·½·¨½â¾öÁËÄ¿±ê³ß´ç²îÒì´ó¡¢ÕÚµ²µÈÎÊÌ⣬¿Ë·þÁËѵÁ·Ê±¸ºÑù ±¾¿Õ¼ä¹ý´óµÄȱµã£¬½øÒ»²½Ìá¸ßÁ˼ì²âµÄ׼ȷÂÊ¡£µ«ÊÇ£¬ÒòΪÈËÔ±µÄ×Ë̬²»Ò»£¬Faster RCNN ÍøÂçÖ»ÄÜ´ÖÂÔµÄÑ¡È¡°²È«Ã±µÄÏà¶ÔλÖá£Íõ±øµÈÈËÌá³öÁËÒ»ÖÖ»ùÓڸĽø YOLOv3 µÄËã·¨¶Ô°²È«Ã±Åå´÷״̬½øÐмì²âÑо¿[26].
¡¡¡¡Ê×ÏÈ£¬Õë¶Ô YOLOv3 Ä¿±ê¼ì²â Ëã·¨´æÔÚµÄÄ¿±êº¯ÊýÓëÆÀ¼ÛÖ¸±ê²»Í³Ò»µÄÎÊÌ⣬½«¸Ä½øµÄ GIOU Ëã·¨Óë YOLOv3 Ëã ·¨Ä¿±êº¯ÊýÏà½áºÏ£¬½øÒ»²½¼õÉÙÔ¤²â¿òÖиºÑù±¾¶ÔÄ£ÐÍѵÁ·µÄÓ°Ï죻Æä´Î£¬½«¸Ä½øºó µÄ YOLOv3 Ëã·¨Ó¦Óõ½°²È«Ã±Åå´÷¼ì²âÈÎÎñÖС£¸ÄÄ£ÐÍʵÏÖÁËÄ¿±êº¯Êý¾Ö²¿×îÓÅ£¬ ¾ßÓкܺõij°ôÐԺͷº»¯ÄÜÁ¦¡£Íõ³ÉÁúµÈÈËÌá³öÁË»ùÓÚÃ沿ÌØÕ÷Óë¾í»ýÉñ¾ÍøÂçÏà½á ºÏµÄËã·¨[27].Ê×ÏÈ£¬ÒÔ¶àÈÎÎñ¼¶Áª¾í»ýÉñ¾ÍøÂ磨MTCNN£©¶Ô¼ì²âÄ¿±ê½øÐÐÁ³²¿ÌØ Õ÷ÌáÈ¡£¬È»ºó£¬½«¾í»ýÉñ¾ÍøÂçÓë VGG Éî¶È¾í»ýÉñ¾ÍøÂçÏà½áºÏʵÏÖ°²È«Ã±¼ì²â¡£
¡¡¡¡¸Ã¼ì²âÄ£ÐÍÕ¼ÓÃÄÚ´æС¡¢Ê¶±ð׼ȷÐԸߡ¢Ë㷨ʵÓÃÐÔÇ¿¡£Cheng Rao µÈÈËÌá³öÁËÒ»ÖÖ SAS-YOLOv3-tiny Ëã·¨¶Ô°²È«Ã±½øÐмì²âÑо¿[28].Ê×ÏÈ£¬²ÉÓÃÈý³ß¶ÈÌØÕ÷½øÐÐÌØÕ÷Ô¤ ²â£»È»ºó£¬ÔÚÌØÕ÷ÌáÈ¡ÍøÂçÖмÓÈëÁ˸ĽøµÄ¿Õ¼ä½ð×ÖËþ³Ø£¨SPP£©Ä£¿é£¬ÌáÈ¡¾ßÓÐ·á ¸»ÓïÒåÐÅÏ¢µÄ¾Ö²¿ºÍÈ«¾ÖÌØÕ÷£»×îºó£¬ÒýÈëÁËÍêÈ«½»¹ýÁª£¨CIoU£©Ëðʧ¶ÔËðʧº¯Êý½ø ÐиĽø¡£¸ÃËã·¨ÔÚ¼õÉÙ²ÎÊýºÍ¼ÆËãÁ¿µÄͬʱ£¬»ñµÃ¸ü¶àµÄÐÅÏ¢ÌØÕ÷£¬Ìá¸ß¼ì²âÐÔÄÜ£¬ËٶȱÈÖØÁ¿¼¶Ä£Ð͸ü¿ì¡£ ×ÛÉÏËùÊö£¬»ùÓÚÉî¶ÈѧϰµÄ°²È«Ã±Åå´÷״̬¼ì²â£¬ÔÚ¼ì²âЧ¹ûÉÏÒªÓÅÓÚ´«Í³µÄ»ú Æ÷ѧϰËã·¨£¬µ«ÊÇ£¬ÓÉÓÚÖÁ½ñÉÐδÓй«¿ªµÄ´óÐÍ°²È«Ã±Êý¾Ý¼¯¿É¹©Ì½¾¿£¬¶øÇÒ»ùÓÚÉî ¶ÈѧϰµÄÄ¿±ê¼ì²â·½·¨ÐèÒªºÜ¸ßµÄÓ²¼þÖ§³Ö£¬ËùÒÔ»ùÓÚÉî¶ÈѧϰµÄ°²È«Ã±Åå´÷¼ì²âÑÐ ¾¿ÔÚСÑù±¾ÉϵÄÎÞÆ«²î¹À¼ÆЧ¹ûÇ·¼Ñ£¬ÊµÊ±ÐÔÊܵ½ÁËÒ»¶¨µÄÏÞÖÆ¡£
¡¡¡¡1.2.3 ÄѵãÓëÌôÕ½
¡¡¡¡°²È«Ã±Åå´÷״̬¼ì²â£¬Í¨¹ý°²×°ÔÚ¼à¿ØÏÖ³¡µÄÉãÏñÍ·»òÏà»ú»ñµÃͼÏñ£¬ÀûÓüÆËã »úÊÓ¾õ¶ÔͼÏñ½øÐд¦Àí£¬½«Åå´÷°²È«Ã±Óë²»Åå´÷°²È«Ã±µÄÈË×÷ΪÁ½À಻ͬµÄÄ¿±ê½øÐÐ ¼ì²â¡£ËäÈ»¾¹ý¶àÄêµÄÑо¿·¢Õ¹Óë»ýµí£¬°²È«Ã±Åå´÷¼ì²âÈ¡µÃÖÚ¶à·á˶µÄ³É¹û£¬µ«ÊÇ ÔÚʵ¼Ê³¡¾°ÖеݲȫñÅå´÷¼ì²âÈÔÈ»´æÔÚÒÔϼ¸¸öÎÊÌâؽ´ýÑо¿¹¥¿Ë£º
¡¡¡¡£¨1£©´æÔÚ¼ì²âËٶȵͣ¬ÊµÊ±ÐÔ²îµÄÎÊÌâ¡£ ¶ÔÓÚÊ©¹¤ÏÖ³¡µÄ°²È«Ã±Åå´÷ÐÐΪ¼ì²âÑо¿¶øÑÔ£¬³ýÁËÒªÇó¼ì²â¾«È·¶ÈÖ®Í⣬ÊÇ·ñ ÄܽøÐÐʵʱ¼ì²âÒ²ÊÇÒ»¸ö¾Þ´óµÄÌôÕ½¡£ÓÉÓÚѵÁ·Ä£Ð͵IJÎÊý¹ý¶à£¬µ¼ÖÂÔËÐÐÄÚ´æÕ¼Óà Á¿¹ý¸ß¡£¼ÆËã»úºÜÄÑÔÚ¼æ¹Ë¼ì²â¾«È·¶ÈµÄͬʱÂú×ãʵʱ¼ì²âµÄÒªÇó¡£Òò´Ë£¬ÈçºÎÔÚÈ· ±£¼ì²â¾«È·¶ÈµÄͬʱ½øÐÐʵʱ¼à²âÒ²ÊÇ´ý½â¾öµÄÎÊÌâÖ®Ò»¡£
¡¡¡¡£¨2£©´æÔÚ©¼ì¡¢Îó¼ìÏÖÏ󣬾«È·¶È²»¸ßµÄÎÊÌâ¡£ ÔÚʵ¼Ê³¡¾°ÖУ¬ÓÉÓÚ°²È«Ã±ÇøÓòÕ¼Åå´÷°²È«Ã±¹¤ÈËÇøÓò±ÈÀýÏà¶Ô½ÏС£¬ÍøÂç¿ÉÄÜ »á½«Î´Åå´÷°²È«Ã±µÄÈËÔ±Îó¼ìΪĿ±ê£»ÔÚÈ˶ิÔӵĻ·¾³Ï£¬¹¤ÈËÖ®¼äÏ໥ÕÚµ²¡¢¹¤ È˱»ÎïÌåÕÚµ²£¬µ¼Ö¼ì²âµ½µÄ°²È«Ã±Î»Öò»ÊǺÜÃ÷ÏÔ£¬¿ÉÄÜ»á´æÔÚ©¼ìºÍÎó¼ìÏÖÏó¡£ Òò´Ë£¬ÈçºÎ¾«È·ÓÐЧµØ¶ÔÊÓƵÖеÄÊ©¹¤ÈËÔ±ÊÇ·ñÅå´÷°²È«Ã±½øÐмì²âÑо¿£¬Ò²ÊÇÏÖ´æ µÄ°²È«Ã±Åå´÷״̬¼ì²âµÄÖ÷ÒªÎÊÌâ¡£
¡¡¡¡1.3 Ñо¿ÄÚÈÝ
¡¡¡¡±¾ÎÄÀûÓô«Í³Ä¿±ê¼ì²â¼¼ÊõºÍÉî¶Èѧϰ¼¼Êõ¶Ô°²È«Ã±Åå´÷״̬½øÐмì²âÑо¿£¬Ê× ÏÈ£¬½«Ëæ»úÞ§Ëã·¨£¨Boosted Random Ferns Algorithm, BRFs£©Ó¦Óõ½°²È«Ã±Åå´÷״̬ ¼ì²âÑо¿ÉÏ¡£ÊµÑé·ÖÎö·¢ÏÖËæ»úÞ§Ëã·¨ÔÚ°²È«Ã±Åå´÷״̬¼ì²âÈÎÎñÖУ¬µ±¼ì²â»·¾³½Ï Ϊ¸´ÔÓʱ£¬´æÔÚ¾«È·¶È²»¸ß¡¢¶ÔС³ß´çÄ¿±ê¼ì²â²»Ãô¸ÐµÈÎÊÌâ¡£Õë¶Ô¸´ÔÓ»·¾³Ï¾«È· ¶È²»¸ßÕâÒ»ÎÊÌ⣬²ÉÓÃÔöÇ¿²ßÂÔ¶ÔËæ»úÞ§Ëã·¨½øÐиĽø£¬¸Ä½øºóµÄËã·¨ÔÚ±£Ö¤¼ì²âËÙ ¶ÈµÄͬʱÌá¸ßÁ˼ì²âµÄ¾«È·¶È¡£Õë¶ÔС³ß´çÄ¿±ê¼ì²â²»Ãô¸ÐµÄÎÊÌ⣬Ìá³öÁ˸ĽøµÄSSD Ë㷨ʵÏÖÁ˶à³ß¶ÈµÄ¼ì²â£¬ÔÚÒ»¶¨³Ì¶ÈÉÏÌá¸ßÁËС³ß´çÄ¿±êµÄ¼ì²â¾«È·¶È¡£±¾ÎÄ µÄÖ÷ÒªÑо¿ÄÚÈݼ°´´ÐµãÖ÷Òª°üÀ¨£º
¡¡¡¡£¨1£©»ùÓڸĽøÔöÇ¿Ëæ»úÞ§Ëã·¨µÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿ ±¾ÕÂÌá³öÁËÒ»ÖָĽøÔöÇ¿Ëæ»úÞ§Ëã·¨£¨Improved Boosted Random Ferns Algorithm, BRFs£©£¬½«Ëæ»úÞ§Ëã·¨Óë¸Ä½øµÄ Real AdaBoost ²ßÂÔÏà½áºÏ£¬¹¹½¨Ò»¸ö¼ì²â¾«¶È½Ï¸ß µÄ·ÖÀàÆ÷£¬²¢½«ÆäÓ¦Óõ½°²È«Ã±Åå´÷¼ì²âÈÎÎñÖС£Ê×ÏÈ£¬»ùÓÚͼÏñ HOG Óò£¬¶ÔͼÏñ ½øÐÐÌØÕ÷ÌáÈ¡ÐγÉͼÏñÌØÕ÷Óò£»Æä´Î£¬ÔÚͼÏñÌØÕ÷ÓòÉÏͨ¹ýËæ»ú¶þÖµ¼ì²â·½·¨Ìáȡͼ Ïñ±ßÔµ£¬²¢¹¹½¨Ëæ»úÞ§£»½Ó×Å£¬ÔÚËæ»úÞ§ÉÏÑ¡ÔñÄ¿±êÓë±³¾°Ö®¼ä´æÔÚÏÔÖø²îÒìµÄλÖà µãºÍ²ÎÊý£¬¹¹½¨¾ßÓÐÒ»¶¨Ê¶±ðÄÜÁ¦µÄÈõ·ÖÀàÆ÷£»×îºó£¬²ÉÓÃÔöÇ¿²ßÂÔÑ¡ÔñÅбðÐÔ½ÏÇ¿ µÄÈõ·ÖÀàÆ÷¹¹½¨Ç¿·ÖÀàÆ÷£¬ÊµÏÖ°²È«Ã±Åå´÷״̬µÄ¼ì²â¡£»ùÓڸĽøÔöÇ¿Ëæ»úÞ§Ëã·¨£¬ ÔÚ°²È«Ã±Åå´÷״̬¼ì²âÑо¿ÖУ¬Äܹ»´Ó½ÏÉÙµÄѵÁ·Ñù±¾ÉÏÓÐЧµØѧϰĿ±êÌØÕ÷£¬×¼È· µØ¼ì²â³öÄ¿±êÎïÌ壬ÔÚ¸´ÔÓ»·¾³ÏÂÕ¹ÏÖÁ˺ܺõij°ôÐԺͷº»¯ÄÜÁ¦¡£
¡¡¡¡£¨2£©»ùÓڸĽø SSD Ëã·¨µÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿ ±¾ÕÂÌá³öÁËÒ»ÖÖ ResNet50-SSD Ëã·¨£¬ÒÔ SSD ÍøÂç×÷Ϊ»ù´¡Ä£ÐÍ£¬¶Ô SSD ÍøÂç »ù´¡Ä£ÐͽøÐÐÐ޸ģ¬²¢Òý½øÁËÅúÁ¿¹éÒ»»¯²ã£¬¹¹½¨Á˶ÔСĿ±ê¼ì²â¾«¶È½Ï¸ßµÄ·ÖÀàÆ÷£¬ ²¢½«ÆäÓ¦Óõ½°²È«Ã±¼ì²âÈÎÎñÖС£Ê×ÏÈ£¬½«»ù´¡ÍøÂç VGG-16 Ì滻Ϊ²Ð²îÍøÂç ResNet-50,½â¾ö SSD-VGG16 ÔÚÄ£ÐÍѵÁ·ÖгöÏÖµÄÌݶÈÏûʧ¼°ÍË»¯µÈÎÊÌ⣻Æä´Î£¬ÔÚ ResNet-50 Ö®ºóÌí¼ÓһϵÁеĸ½¼Ó²ã£¬²¢ÔÚ¸½¼Ó²ãÖÐÒýÈë BN ²ã£¬½â¾ö¶à³ß¶ÈÔ¤²âµÄ ÎÊÌ⣻×îºó£¬Í¨¹ý·Ç¼«´óÖµÒÖÖÆËã·¨£¬Â˳ýµôС¸ÅÂʵÄÄ¿±êµÃµ½×îÖյļì²â½á¹û¡£Êµ Ñé±íÃ÷£¬»ùÓÚ ResNet50-SSD µÄ°²È«Ã±Åå´÷Çé¿ö¼ì²âËã·¨£¬ÔÚÒ»¶¨³Ì¶ÈÉÏÌá¸ßÁËС³ß ´çÄ¿±êµÄ¼ì²â¾«È·¶È¡£
¡¡¡¡1.4 ÂÛÎÄ×éÖ¯½á¹¹
¡¡¡¡±¾ÎÄÖ÷Òª¶Ô»ùÓÚ¼ÆËã»úÊÓ¾õµÄ°²È«Ã±Åå´÷״̬¼ì²âÈÎÎñ½øÐÐÑо¿£¬Ö÷ÒªÑо¿ÄÚÈÝ ·ÖΪÎå¸öÕ½ڡ£
¡¡¡¡£¨1£©Ð÷ÂÛ£ºÊ×ÏÈ£¬¸ÅÊö±¾ÎĵÄÑо¿±³¾°ÓëÒâÒ壻Æä´Î£¬²ûÊö¹úÄÚÍⰲȫñÅå´÷ ״̬¼ì²âÑо¿µÄ×îÐÂÑо¿ÏÖ×´£¬²¢·ÖÎö°²È«Ã±Åå´÷״̬¼ì²âÑо¿´æÔÚµÄÎÊÌ⣻×îºó£¬ ½éÉܱ¾ÎĵÄÖ÷ÒªÑо¿ÄÚÈÝ£¬²¢¶Ô±¾ÎĵÄ×éÖ¯½á¹¹½øÐмòÒª¸ÅÀ¨¡£
¡¡¡¡£¨2£©Ïà¹ØÀíÂÛ»ù´¡£º¶ÔËæ»úÞ§Ëã·¨ºÍ SSD ÍøÂçµÄÏà¹ØÀíÂÛ½øÐмòÒªÐðÊö¡£
¡¡¡¡£¨3£©»ùÓÚÔöÇ¿Ëæ»úÞ§µÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿£ºÊ×ÏÈ£¬Í¨¹ý·½ÏòÌݶÈÖ±·½Í¼ ¶ÔͼÏñ½øÐÐÌØÕ÷ÌáÈ¡£¬ÐγÉͼÏñµÄÌØÕ÷Óò¿Õ¼ä£»Æä´Î£¬ÔÚÌØÕ÷Óò¿Õ¼äÉϲÉÓÃËæ»ú¶þÔª ²âÊÔ·½·¨¹¹½¨Ëæ»úÞ§£¬²¢Í¨¹ýËæ»úÞ§¹¹½¨Èõ·ÖÀàÆ÷£»ÔٴΣ¬½áºÏ¼ÓȨϵÊý·¨ÀûÓà Real AdaBoost Ë㷨ѡÔñ×î¾ßÅбðÐÔµÄÈõ·ÖÀàÆ÷¹¹½¨Ç¿·ÖÀàÆ÷£»×îºó£¬ÔÚ°²È«Ã±Êý¾Ý¼¯ÖÐ¶Ô IBRFs ½øÐÐѵÁ·ºÍ²âÊÔ¡£
¡¡¡¡£¨4£©»ùÓÚ ResNet50-SSD Ëã·¨µÄ°²È«Ã±Åå´÷¼ì²âÑо¿£ºÊ×ÏÈ£¬½« SSD ÍøÂçµÄ»ù ´¡ÍøÂç VGG-16 Ì滻Ϊ ResNet-50,²¢ÔÚ¸½¼Ó²ãÖÐÒýÈë BN ²ã£»Æä´Î£¬·Ö±ðÔÚ²»Í¬µÄ ÌØÕ÷²ãÉ϶Բ»Í¬³ß´çµÄÄ¿±ê½øÐÐÔ¤²â£»×îºó£¬ÀûÓ÷Ǽ«´óÖµÒÖÖÆËã·¨Çó³ö¼ì²â½á¹û£¬ ²¢¶ÔÆä½øÐзÖÎö±È½Ï¡£
¡¡¡¡£¨5£©×ܽáÓëÕ¹Íû£º¶Ô±¾ÎĵÄÖ÷ÒªÑо¿ÄÚÈݲ¢¶Ô½á¹û½øÐÐ×ܽᣬ²¢¶ÔÏÂÒ»²½Ñо¿ ¹¤×÷½øÐй滮¡£ ±¾Îĸ÷¸öÕ½ڼäµÄ×éÖ¯¹ØϵÈçͼ 1-1 Ëùʾ¡£
¡¡¡¡2 Ïà¹ØÀíÂÛ»ù´¡
¡¡¡¡2.1 Ëæ»úÞ§Ëã·¨
¡¡¡¡2.1.1 ±´Ò¶Ë¹·ÖÀà
¡¡¡¡2.1.2 ÆÓËر´Ò¶Ë¹·ÖÀà
¡¡¡¡2.1.3 °ëÆÓËر´Ò¶Ë¹·ÖÀà
¡¡¡¡2.1.4 Ëæ»úÞ§·ÖÀàÆ÷¹¹½¨
¡¡¡¡2.2 ¾í»ýÉñ¾ÍøÂç
¡¡¡¡2.2.1 ¾í»ý²ã
¡¡¡¡2.2.2 ³Ø»¯²ã
¡¡¡¡2.2.3 È«Á¬½Ó²ã
¡¡¡¡2.3 SSD ÍøÂç½á¹¹
¡¡¡¡2.3.1 ÌØÕ÷ÌáÈ¡
¡¡¡¡2.3.2 Prior box Ñ¡È¡
¡¡¡¡2.3.3 Ԥѡ¿òµÄ´¦Àí
¡¡¡¡3 »ùÓڸĽøÔöÇ¿Ëæ»úÞ§µÄ°²È«Ã±Åå´÷¼ì²âÑо¿
¡¡¡¡3.1 ÔöÇ¿Ëæ»úÞ§Ëã·¨
¡¡¡¡3.1.1 »ùÓÚ HOG ÓòµÄËæ»úÞ§·ÖÀàÆ÷
¡¡¡¡3.1.2 ÔöÇ¿Ëæ»úÞ§
¡¡¡¡3.2 ¸Ä½øÔöÇ¿Ëæ»úÞ§Ëã·¨
¡¡¡¡3.2.1 AdaBoost Ëã·¨
¡¡¡¡3.2.2 Real AdaBoost Ëã·¨
¡¡¡¡3.2.3 ¸Ä½øµÄÔöÇ¿Ëæ»úÞ§Ëã·¨
¡¡¡¡3.3 ʵÑé½á¹û¼°·ÖÎö
¡¡¡¡3.3.1 ʵÑéÉèÖÃ
¡¡¡¡3.3.2 ²»Í¬¼¯³ÉËã·¨¶Ô±È·ÖÎö
¡¡¡¡3.3.3 Ïàͬ³¡¾°Ï¸ĽøÇ°ºóµÄ¼ì²âЧ¹û
¡¡¡¡3.3.4 °²È«Ã±Åå´÷״̬¼ì²âÄ£ÐÍ·ÖÎö
¡¡¡¡3.4 ±¾ÕÂС½á
¡¡¡¡4 »ùÓÚ ResNet50-SSD µÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿
¡¡¡¡4.1 »ù´¡ÍøÂç
¡¡¡¡4.1.1 ResNet-50
¡¡¡¡4.1.2 ÅúÁ¿¹éÒ»»¯
¡¡¡¡4.2 ResNet50-SSD Ä£ÐÍ
¡¡¡¡4.2.1 ÍøÂç½á¹¹
¡¡¡¡4.2.2 Ëðʧº¯Êý
¡¡¡¡4.3 ʵÑé½á¹û¼°·ÖÎö
¡¡¡¡4.3.1 ʵÑé»·¾³ÅäÖÃ
¡¡¡¡4.3.2 loss ÇúÏß¼°µü´ú´ÎÊýʵÑé½á¹û·ÖÎö
¡¡¡¡4.3.3 С³ß´çÄ¿±êµÄ¼ì²âÐÔÄÜ·ÖÎö
¡¡¡¡4.3.4 ¼ì²â׼ȷ¶È·ÖÎö
¡¡¡¡4.4 ±¾ÕÂС½á
¡¡¡¡5 ×ܽáÓëÕ¹Íû
¡¡¡¡±¾¹¤×÷±à³ÌÓïÑÔ»·¾³Îª Matlab ºÍ Python,Õë¶Ô°²È«Ã±Åå´÷״̬¼ì²â¾«È·¶È ½ÏµÍºÍʵʱÐԽϲîµÄÎÊÌ⣬·Ö±ðÌá³öÁË IBRFs Ëã·¨ºÍ ResNet50-SSD Ä£ÐÍ£¬²¢²É Óùú¼Ê¹«¿ªµÄ°²È«Ã±Êý¾Ý¼¯ºÍ»¥ÁªÍøÅÀ³æÊý¾Ý¼¯ÑéÖ¤Ëã·¨ºÍÄ£Ð͵ÄÓÐЧÐÔ¡£ÏÖ½« ±¾Ñо¿µÄ¹¤×÷½øÐÐ×ܽáÓëÕ¹Íû¡£
¡¡¡¡5.1 ×ܽá
¡¡¡¡±¾ÎÄÖ÷Òª¶Ô»ùÓÚ¼ÆËã»úÊÓ¾õµÄ°²È«Ã±Åå´÷Çé¿ö¼ì²âËã·¨Õ¹¿ªÑо¿£¬Ìá³öÁË IBRFs Ëã·¨ºÍ ResNet50-SSD Ä£ÐͶÔÊ©¹¤¹¤ÈËÅå´÷°²È«Ã±µÄÇé¿ö½øÐÐÔ¤²â£¬ÊµÏÖ ÁË¿ìËÙ׼ȷµÄ¼ì²â£¬Îª¹¤µØ°²È«¼à¶½¼°Ö´·¨ÌṩÁ˼¼Êõ±£ÕÏ£¬¶ÔÊ©¹¤ÏÖ³¡µÄ°²È« ¹ÜÀí¾ßÓÐÖØÒªµÄÒâÒå¡£±¾ÎÄÖ÷Òª¹¤×÷×ܽἰ´´ÐÂÈçÏ£º
¡¡¡¡£¨1£©Êý¾Ý¼¯¹¹½¨¡£±¾ÎÄ»ùÓÚ¼ÆËã»úÊÓ¾õ¶Ô°²È«Ã±Åå´÷״̬½øÐмì²âÑо¿£¬ µ«Êǵ±Ç°²¢Ã»Óй«¿ª¿ÉÓõĴóÐÍ°²È«Ã±¼ì²âÊý¾Ý¼¯£¬ÕâÎÞÒɺܴó³Ì¶ÈÉÏÔö¼ÓÁËÑÐ ¾¿µÄÄѶȡ£Õë¶ÔÕâ¸öÎÊÌ⣬±¾ÎÄͨ¹ýÍøÂçÅÀ³æ¼¼Êõ£¬²¢²ÉÓÃÁË LabelImg ¶ÔÊý¾Ý ½øÐбê¼Ç´¦Àí£¬¹¹½¨°²È«Ã±¼ì²âÊý¾Ý¼¯¡£±¾ÎĹ¹½¨µÄ°²È«Ã±¼ì²âÊý¾Ý¼¯ÎªºóÐøµÄ ʵÑéÌṩÁËÊý¾ÝÖ§³Å¡£
¡¡¡¡£¨2£©»ùÓÚÔöÇ¿Ëæ»úÞ§µÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿¡£Ìá³öÁËÒ»ÖָĽøÔöÇ¿Ëæ »úÞ§µÄËã·¨£¨IBRFs£©£¬¸Ã·½·¨½«ÔöÇ¿Ëæ»úÞ§Ëã·¨£¨BRFs£©ºÍÒ»ÖָĽøµÄÔöÇ¿²ß ÂÔÏà½áºÏ£¬¹¹½¨Ò»¸ö¼ì²â¾«¶È½Ï¸ßµÄ·ÖÀàÆ÷¡£Ê×ÏÈ£¬½è¼øÔöÇ¿Ëæ»úÞ§Ëã·¨£¨BRFs£© µÄ˼Ï룬»ùÓÚͼÏñ HOG Óò£¬¶ÔͼÏñ½øÐÐÌØÕ÷ÌáÈ¡£¬ÐγÉͼÏñÌØÕ÷Óò£»È»ºó£¬ÔÚ Í¼ÏñÌØÕ÷ÓòÉϹ¹½¨Í¨¹ý¶þÖµ±È½Ï·½·¨¹¹½¨Ëæ»úÞ§£¬²¢Ñ¡ÔñËæ»úÞ§ÖÐÅбðÐÔ×îÇ¿µÄ λÖõãºÍ²ÎÊý¹¹½¨Èõ·ÖÀàÆ÷£»×îºó£¬²ÉÓÃÔöÇ¿²ßÂÔ£¬Ñ¡Ôñ×î¾ßÅбðÐÔµÄÈõ·ÖÀàÆ÷£¬ ¹¹½¨×îÓŵݲȫñÅå´÷¼ì²â·ÖÀàÆ÷¡£ÊµÑéÖ¤Ã÷£¬ÔÚ½øÐа²È«Ã±Åå´÷ÈÎÎñ¼ì²âʱ£¬ ±¾ÎÄÌá³öµÄ¸Ä½øÔöÇ¿Ëæ»úÞ§µÄËã·¨£¨IBRFs£©¾ßÓнϸߵļì²â¾«¶È¡£
¡¡¡¡£¨3£©»ùÓÚÉî¶ÈѧϰµÄ°²È«Ã±Åå´÷״̬¼ì²âÑо¿¡£Ìá³öÁËÒ»ÖÖ ResNet50-SSD Ä£ÐÍ£¬¸ÃÄ£ÐÍÒÔ SSD ÍøÂç×÷Ϊ»ù´¡£¬Ê×ÏÈ£¬Ê¹Óà ResNet-50 ÍøÂç½øÐÐÌØÕ÷ÌáÈ¡£¬ ²¢¶Ô ResNet-50 ÖÐ conv4_s µÄµÚÒ»¸ö²Ð²îÍøÂç½á¹¹½øÐÐÐ޸ģ¬Ê¹µÃµÚÒ»¸öÊä³öµÄ ÌØÕ÷¾ØÕó´óС²»±ä£¬Æä´Î£¬ÔÚ conv4_s ºóÐøÌí¼ÓµÄһϵÁи½¼Ó²ãÖÐÒýÈëÁË BN ²ã£¬·Ö±ðÔÚ²»Í¬µÄÌØÕ÷²ãÉ϶Բ»Í¬³ß´çµÄÄ¿±ê½øÐÐÔ¤²â£»×îºó£¬Í¨¹ý·Ç¼«´óÖµÒÖÖÆËã ·¨£¬Â˳ýµôС¸ÅÂʵÄÄ¿±êµÃµ½×îÖյļì²â½á¹û¡£ÊµÑéÖ¤Ã÷£¬¸ÃÄ£ÐÍʹÓÃÁ˲вîÍø Âç½øÐжà³ß¶ÈѵÁ·ºó£¬Ä£ÐÍÄܹ»ÊÊÓ¦²»Í¬·Ö±æÂʵݲȫñÅå´÷״̬¼ì²âÈÎÎñ¡£
¡¡¡¡5.2 Õ¹Íû
¡¡¡¡»ù´¡ÉèÊ©µÄ½¨Éè¼Ó¿ìÁ˳ÇÕò»¯½¨ÉèµÄ½Å²½£¬Ê©¹¤³¡ËùµÄ°²È«ÎÊÌâÒ²³ÉΪÁËÈË Ãǵ£ÐĵÄÖصãÎÊÌâ¡£±¾Ñо¿Ö÷ÒªÕë¶Ô°²È«Ã±Åå´÷״̬½øÐÐÏà¹ØµÄ¼ì²âÑо¿£¬´Ó´« ͳµÄ»úÆ÷ѧϰ·½·¨ºÍÉî¶Èѧϰ·½·¨Á½·½Ã棬¶Ô°²È«Ã±Åå´÷״̬¼ì²âÄ£ÐͽøÐÐ¸Ä ½ø£¬µ«ÊÇÑо¿ÖÐÈÔÈ»´æÔÚһЩ²»×㣬Ö÷ÒªÌåÏÖÔÚÒÔÏÂÁ½¸ö·½Ã棺
¡¡¡¡£¨1£©¼ì²â¾«È·¶ÈµÍÎÊÌâ¡£Õë¶Ô°²È«Ã±Åå´÷״̬¼ì²â¾«È·¶È½ÏµÍµÄÎÊÌ⣬±¾ ÎIJ¢Î´¿¼Âǵ½µ±ÅÄÉã½Ç¶È²»Í¬¡¢¹âÕձ仯¡¢ÌìÆø±ä»¯¡¢ÎïÌåÕÚµ²µÈÎÊÌâ¶Ô¼ì²â¾« ¶ÈµÄÓ°Ï죬Òò´ËÏÂÒ»²½Ñо¿µÄÖص㽫¿¼ÂÇÈçºÎ½â¾ö¸ÃÎÊÌâ¡£
¡¡¡¡£¨2£©¼ì²âËٶȵÍÎÊÌâ¡£Õë¶Ô°²È«Ã±Åå´÷״̬¼ì²âËٶȽϵ͵ÄÎÊÌâ¡£±¾ÎIJ¢ δ¿¼Âǵ½ÈçºÎ´Ó¼õСģÐͲÎÊý¡¢ËõСÍøÂç½á¹¹¡¢¶ÔͼÏñ½øÐÐÔ¤´¦ÀíµÈ·½Ãæ¼õСÔË ÐÐÄÚ´æ¡£ÏÂÒ»²½½«Ñо¿ÈçºÎ¼õСģÐ͵ÄÔËÐÐÄڴ棬Ìá¸ßÄ£Ðͼì²âËÙ¶ÈÎÊÌâ¡£ Õë¶ÔÉÏÊöÎÊÌ⣬¶ÔδÀ´µÄ¹¤×÷Õ¹ÍûÈçÏ£º
¡¡¡¡£¨1£©¶àÌØÕ÷Èںϡ£¿¼ÂÇ´Ó¶à¸ö½Ç¶ÈÌáÈ¡¶àÖØÌØÕ÷£¬½«²»Í¬µÄÌØÕ÷½øÐÐÈںϣ¬ Ìá¸ß¼ì²âÄ£ÐͶԼì²âÄ¿±êµÄÁéÃô¶È£¬½«´«Í³µÄÌØÕ÷ÌáÈ¡·½·¨ÓëÉî¶Èѧϰ·½·¨Ï໥ Èںϣ¬½øÒ»²½Ìá¸ßÄ£Ðͼì²â¾«È·¶ÈºÍ¼ì²âËٶȡ£
¡¡¡¡£¨2£©ÆÕÊÊÐÔ¡£¿¼ÂÇÈçºÎ½« IBRFs Ëã·¨ºÍ ResNet50-SSD Ä£ÐÍÔËÓõ½°²È«Ã± Åå´÷¼ì²âµÄ¾ßÌåÓ¦ÓÃÖУ¬³¢ÊÔ·ÖÎö±¾ÎĵĸĽøÖ®´¦ÊÇ·ñ¾ßÓÐÆÕÊÊÐÔ£¬ÄÜ·ñÑÓÉìµ½ ͨÓõÄÄ¿±ê¼ì²âÁìÓòÖС£
¡¡¡¡Ö л
¡¡¡¡ÐÐÎÄÖÁ´Ë£¬Âä±ÊΪÖÕ£¬Ãñ´óÆßÔØÇà´º£¬Ê¼ÓÚ 14 ÄêÊ¢ÏÄ£¬ÖÁÓÚ 21 ÄêÊ¢ÏÄ¡£Ä¿ Ö®ËùÏò£¬½ÔΪ¹ýÍù¡£ÎÒÔøÓйýÀ§»ó¡¢ÃÔã¡¢ÎÞÖú£¬µ«Ò²Ôø±»ÐÅÈΡ¢±»ÒÀÀµ¡¢±»ÈÏ ¿É£¬ÆßÔØÇà´º³ÐÔØÁËÎÒÔÙ¶àµÄ±¯»¶Ï²ÀÖ£¬ÁôϵĶ¼ÊdzÁµéµéµÄÇà´ººÍÊÕ»ñ£¬×Ýʹ Íò°ã²»ÉáÒ²ÒªÐÄ»³¸Ð¼¤¡£ ÌÒÀî²»ÑÔ£¬ÏÂ×Ô³Éõè¡£¸ÐлÎҵĵ¼Ê¦ÕÅǬÀÏʦ£¬Äú´ÓδÏÓÆúÎÒ²ÅѧÊèdz£¬´Ó ÂÛÎĵÄÑ¡Ìâµ½¶¨¸å£¬Ã¿Ò»¸ö½×¶Î¶¼Çã×¢ÁË´óÁ¿µÄÐÄѪºÍ¾«Á¦¡£Ñ§¹óµÃʦ£¬Òà¹óµÃ ÓÑ£¬ÓÚÎÒ¶øÑÔÊÇָ·ÈË¡¢µ«¸üÏñÊÇÅóÓÑ£¬ÄúµÄ¹ÄÀø¸øÎÒ´øÀ´ÁËĪ´óµÄÓÂÆø£¬ÔÚÎÒ ÃÔãºÍ²»°²µÄʱºò¸øÁËÎÒÒ»¸ö¼á¶¨µÄ·½Ïò£¬ÄܳÉΪ"ÕÅ×Ü"µÄѧÉúÊÇÈÃÎÒ×îÒýÒÔ Îª°Á¡¢×î°²ÐĵÄÊÂÇé¡£³ÐÃɽ̻壬ѧÉúÐÄ»³¸Ð¼¤£¬Ô¸ÄúÌÒÀîÂúÌìÏ£¬ÊÂÊÂ˳Ëì¡£ ¸¸Ä¸Ö®°®×Ó£¬ÔòΪ֮¼ÆÉîÔ¶¡£¸Ðл°Ö°ÖÂèÂè´ÍÎÒÉúÃü£¬½ÚÒÂËõʳ£¬ÓýÎҳɳ¤¡£
¡¡¡¡·ÖÏíÎÒµÄϲۧÀÖ£¬ñöÌýÎÒµÄÏë·¨£¬¸øÓèÎÒ×î´óµÄÖ§³ÖÓë¿Ï¶¨¡£¸Ðл½ã½ãºÍµÜµÜ ÃǵÄÅã°é£¬¼ûÖ¤×ű˴˵ijɳ¤£¬Ã¿´Îµ±ÎÒÇéÐ÷±ÀÀ£µÄʱºò£¬ÄãÃÇÊÇÎÒ×î¼áʵµÄºó ¶Ü£¬ÈÃÎÒÐÄ°²¡¢¸øÎÒÎÂů£¬ËùÓеÄίÇü¡¢²»¿ªÐĵ½ÄãÃÇÕâÀï¾ÍÏûʧÁË¡£ÎÞÒԻر¨£¬ Ö»Ïë¸ü¼ÓŬÁ¦£¬³ÉΪÄãÃǵĽ¾°Á¡£
¡¡¡¡ËùÓö½ÔÁ¼ÈË£¬¸ÐлÎÒµÄʦĸ½ðÉý¾Õ£¬Ïñ½ã½ãÒ»Ñù·ÖÏí×ÅËýµÄÉú»î£¬½Ì»áÎÒÃÇ ÎªÈË´¦ÊÀÖ®µÀ£¬ËäÈ»ÎÒÃÇÕæÕýÊìϤµÄʱ¼ä²»³¤£¬µ«ÊǽðÀÏʦ×Ü»á¸øÈËÒ»ÖÖ°²ÐÄ̤ ʵµÄ¸Ð¾õ¡£Ò²Òª¸ÐлÎÒµÄС»ï°éÉÛ¶¨ÇÙ£¬ÆßÄêÈçÒ»ÈÕÕÕ¹ËÎÒ£¬Ò»Æð̽ÌÖ½â¾öÎÊÌ⣬ Ï໥°ïÖú£¬·ÖÏí×ű˴ËÉú»îµÄµãµãµÎµÎ£¬¼ûÖ¤±Ë´ËµÄ³É³¤¡£»¹Òª¸ÐлʵÑéÊÒС»ï °é¶ÔÎÒ±ÏÒµÂÛÎĵİïÖú£¬¸øÎÒµÄÉú»î´øÀ´ÁË»¶ÉùЦÓï¡£
¡¡¡¡Æ½Éú¸ÐÖª¼º£¬·½´çÆñÓÆÓÆ¡£¸ÐлÎÒµÄСæ¢ÃÃÎâÙ»¡¢Ð¤ÁÕ¡¢°×Óî¡¢ÁºÃ÷Ð㣬ÔÚ ÎÒµÍÂäµÄʱºò¸øÎÒ¹ÄÀø£¬°üÈÝÎÒµÄСƢÆø£¬²»ÒÅÓàÁ¦µØ°ïÖúºÍÖ§³ÖÎÒ£¬¸ÐлÄãÃÇ ³öÏÖÔÚÎÒµÄÉúÃüÖУ¬²¢¸øÎÒ´øÀ´ÁËÎÂů¡£Ô¸ÓÐÇ°³Ì¿É±¼¸°£¬ÒàÓÐËêÔ¿ɻØÊס£
¡¡¡¡ÒÔÃÎΪÂí£¬²»¸ºÉØ»ª£¬¸Ðл²»ÔøÑÔÆúµÄ×Ô¼º£¬×ÝÈ»¹ý³Ì¿²¿²¿À¿À£¬µ«ÒÀ¾ÉÈÈ °®Éú»î£¬Ã¿Ò»´Î¾Àú¶¼ÊÇÒ»Öֳɳ¤£¬¶¼»áÔÚ²»¾Òâ¼äΪÎÒÃǵÄÉú»îÀÛ»ý×ʱ¾£¬Ô¸ Äã×ö¸ö˵»°×öÊÂÈÃÈËÊæ·þµÄÅ®º¢×Ó¡£ Ô¸Îáʦ£¬ÊÂÊÂ˳Ë죡Ը¼ÒÈË£¬ÉíÌ彡¿µ£¡Ô¸ÎáÓÑ£¬Ç°³ÌËƽõ£¡Ô¸Ãñ´ó£¬ÔÐÓý Ó¢²Å£¡
£¨ÈçÄúÐèÒª²é¿´±¾Æª±ÏÒµÉè¼ÆÈ«ÎÄ£¬ÇëÄúÁªÏµ¿Í·þË÷È¡£©