In S502 Video Frames Are Acquired
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 sphere of computer know-how, and particularly, to a goal tracking method and machine, a target choice method, a pc-readable storage medium, and ItagPro electronic tools. 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。
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 broadly utilized in robot navigation, clever video surveillance, industrial inspection, iTagPro smart tracker aerospace and iTagPro smart tracker plenty of other fields. It is a vital branch of image processing and computer vision disciplines, and can be the core part of intelligent surveillance methods. At the identical time, target detection can also be a primary algorithm in the sphere of pan-identification, which performs an important position in subsequent duties comparable to face recognition, iTagPro support gait recognition, crowd counting, and instance segmentation.
The deep learning object detection algorithm primarily based on the neural community as the main model has been developed relatively rapidly. 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 must be famous that the information disclosed within the above background expertise part is only used to reinforce the understanding of the background of the embodiments of the current disclosure, iTagPro official and subsequently might include information that does not represent prior artwork known to those of bizarre talent 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 aim of the embodiments of the current disclosure is to provide a goal monitoring technique and device, a goal selection technique, a pc-readable storage medium, and an electronic gadget, iTagPro smart tracker no less than to a sure extent, to overcome the inability to concurrently optimize the detection accuracy and time-consuming detection in target detection in associated applied sciences 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 options and advantages of the embodiments of the current disclosure can be obvious from the following detailed description, or partly discovered by practicing 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 accordance with the first aspect of the embodiments of the present disclosure, iTagPro smart tracker there's provided a goal tracking technique, iTagPro tracker the above-mentioned method includes: buying a video frame, performing goal detection processing on the above-talked about video frame by a first detection module, iTagPro smart tracker and iTagPro obtaining N detected targets in the above-mentioned video body , and the primary coordinate data of every of the above-mentioned detection targets, whereby, N is a positive integer; according to the above-mentioned first coordinate data, iTagPro smart tracker the partial picture in the above-mentioned video body is obtained, and the i-th image corresponding to the i-th detection goal is obtained; through the second detection The module performs target detection processing on the above-mentioned i-th picture to acquire the second coordinate data of the above-talked about i-th detection target, wherein the above-mentioned i-th detection target is any one of many above-mentioned N detection targets, and the worth of i isn't better than N A constructive integer; the monitoring of the i-th detection goal is realized via the second coordinate information of the i-th detection goal.