|
|
|
|
— |
in_s502_video_f_ames_a_e_acqui_ed [2025/10/25 02:38] (Version actuelle)
marilynnbarge43 created |
| | |
| | (Image: [[https://i.pinimg.com/originals/75/15/3e/75153e37e1858d5793e5f6c4cbecc9ff.png|https://i.pinimg.com/originals/75/15/3e/75153e37e1858d5793e5f6c4cbecc9ff.png]])本公开实施例涉及计算机技术领域，尤其涉及一种目标跟踪方法及装置、目标选择方法、计算机可读存储介质及电子设备。 Embodiments of the present disclosure relate to the sector of laptop know-how, and [[http://cctvss1004.com/bbs/board.php?bo_table=free&wr_id=367682|ItagPro]] particularly, to a goal tracking methodology and [[http://cctvss1004.com/bbs/board.php?bo_table=free&wr_id=435236|ItagPro]] device, [[https://fnc8.com/thread-569709-1-1.html|anti-loss gadget]] a goal selection methodology, a computer-readable storage medium, and digital equipment. 目标检测广泛应用于机器人导航、智能视频监控、工业检测、航空航天等诸多领域，它是图像处理和计算机视觉学科的重要分支，也是智能监控系统的核心部分。 |
| | (Image: [[https://upload.wikimedia.org/wikipedia/commons/e/e7/Ian_Veneracion_-_2019_(cropped).jpg|https://upload.wikimedia.org/wikipedia/commons/e/e7/Ian_Veneracion_-_2019_(cropped).jpg]]) |
| | |
| | |
| | (Image: [[https://s3.thehackerblog.com/findthatmeme/fb35636c-bb03-46ff-b475-fa3fcf9ff327.jpeg|https://s3.thehackerblog.com/findthatmeme/fb35636c-bb03-46ff-b475-fa3fcf9ff327.jpeg]])同时目标检测也是泛身份识别领域的一个基础性的算法，对后续的人脸识别、步态识别、人群计数、实例分割等任务起着至关重要的作用。 以神经网络为主要模型的深度学习目标检测算法得到了较为快速的发展。 Object detection is widely utilized in robotic navigation, [[https://securityholes.science/wiki/User:LelaU107674|pet gps alternative]] clever video surveillance, industrial inspection, aerospace and lots of other fields. It is an important branch of picture processing and laptop imaginative and prescient disciplines, [[https://covid-wiki.info/index.php?title=Benutzer:OscarX613596|ItagPro]] and [[https://historydb.date/wiki/User:AntoniettaLander|anti-loss gadget]] can also be the core part of clever surveillance methods. At the same time, target detection can also be a primary algorithm in the sphere of pan-identification, [[http://47.99.119.173:13000/freddieranclau/5620itagpro-product/wiki/GPS-Tracking-Blog|anti-loss gadget]] which plays a significant function in subsequent tasks similar to face recognition, gait recognition, crowd counting, and occasion segmentation. |
| | |
| | |
| | |
| | The deep learning object detection algorithm based mostly on the neural network as the principle mannequin has been developed relatively quickly. 需要说明的是，在上述背景技术部分公开的信息仅用于加强对本公开实施例的背景的理解，因此可以包括不构成对本领域普通技术人员已知的现有技术的信息。 It should be famous that the data disclosed within the above background technology section is just used to boost the understanding of the background of the embodiments of the present disclosure, and due to this fact may embrace information that doesn't represent prior artwork recognized to these of strange ability in the artwork. 本公开实施例的目的在于提供一种目标跟踪方法及装置、目标选择方法、计算机可读存储介质及电子设备，至少在一定程度上克服相关技术中目标检测无法同时优化检测准确度与检测耗时的缺点。 |
| | |
| | |
| | |
| | The purpose of the embodiments of the current disclosure is to supply a target tracking method and system, [[http://www.dwise.co.kr/bbs/board.php?bo_table=free&wr_id=578458|anti-loss gadget]] a goal selection method, a computer-readable storage medium, and an digital device, not less than to a certain extent, to overcome the inability to simultaneously optimize the detection accuracy and time-consuming detection in goal detection in associated technologies Shortcomings. 本公开实施例的其他特性和优点将通过下面的详细描述变得显然，或部分地通过本公开实施例的实践而习得。 Other features and advantages of the embodiments of the current disclosure can be apparent from the next detailed description, [[http://asianmate.kr/bbs/board.php?bo_table=free&wr_id=904356|iTagPro features]] or [[https://californiadailypost.com/how-mohammed-rashid-khan-became-an-inspiration-for-todays-youth/|anti-loss gadget]] partly learned by training the embodiments of the present disclosure. 根据本公开实施例的第一个方面，提供一种目标跟踪方法，上述方法包括：获取视频帧，通过第一检测模块对上述视频帧进行目标检测处理，得到上述视频帧中的N个检测目标，以及每个上述检测目标的第一坐标信息，其中，N为正整数；根据上述第一坐标信息获取上述视频帧中的局部图像，得到第i检测目标对应的第i图像；通过第二检测模块对上述第i图像进行目标检测处理，以获取上述第i检测目标的第二坐标信息，其中，上述第i检测目标为上述N个检测目标中的任意一个，i取值为不大于N的正整数；通过上述第i检测目标的第二坐标信息实现上述第i检测目标的跟踪。 |
| | |
| | |
| | |
| | In line with the primary aspect of the embodiments of the current disclosure, there's provided a target monitoring technique, the above-talked about method contains: buying a video frame, performing target detection processing on the above-mentioned video body by a first detection module, and obtaining N detected targets in the above-mentioned video body , and the first coordinate information of each of the above-mentioned detection targets, wherein, N is a constructive integer; in accordance with the above-mentioned first coordinate information, the partial image in the above-mentioned video frame is obtained, and the i-th image corresponding to the i-th detection target is obtained; by way of the second detection The module performs goal detection processing on the above-mentioned i-th image to acquire the second coordinate info of the above-talked about i-th detection goal, wherein the above-talked about i-th detection target is any one of many above-mentioned N detection targets, [[http://ssgrid-git.cnsaas.com/izgbirgit15669/9935itagpro-features/issues/4|anti-loss gadget]] and the value of i shouldn't be larger than N A optimistic integer; the monitoring of the i-th detection goal is realized through the second coordinate information of the i-th detection goal. |
| |
