• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>International Journal of Pattern Recognition and Artificial Intelligence >Target Identification and Location Algorithm Based on SURF-BRISK Operator
【24h】

Target Identification and Location Algorithm Based on SURF-BRISK Operator

机译：基于SURF-BRISK算子的目标识别与定位算法

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Accurate and fast target image recognition is an important function of applications such as remote sensing imaging and medical imaging. However, an operator such as speeded up robust feature (SURF) cannot be accurately matched in the recognition process of a target image. This led us to propose the use of a method capable of matching identification, i.e. binary robust invariant scalable keypoints (BRISK) operators, in combination with SURF operators. The proposed algorithm combines the accuracy of SURF operators and the rapidity of BRISK operators to obtain a quick and accurate way of matching. The initial matching of image feature extraction for targets is performed using the SURF-BRISK algorithm, and similarity measurements of feature matching are performed for the feature points of initial matching using the Hamming distance. Then, secondary fine matching is performed using the M-estimator Sample and Consensus (MSAC) algorithm to eliminate mismatched point pairs in order to achieve recognition of target images. Then, the three-dimensional coordinates of the work piece are obtained by using a binocular stereo vision system to provide location coordinates for the robots to grasp the work pieces accurately. In the experiment, stereo vision matching is conducted for targets obtained using the SURF-BRISK algorithm, and the location coordinates of targets are passed to the robot controller. The experimental results show that if the special geometric distortion is neglected, this method can be adapted for accurate positioning of the target; hence, it can identify the target in complex environments, access the location coordinates of the target, and achieve accurate robotic grasping of the work piece in real time.
机译：准确，快速的目标图像识别是诸如遥感成像和医学成像等应用程序的重要功能。然而，诸如加速健壮特征（SURF）之类的运算符不能在目标图像的识别过程中准确地匹配。这导致我们提出结合SURF运算符使用能够匹配标识的方法，即二进制鲁棒不变可缩放关键点（BRISK）运算符。该算法结合了SURF算子的准确性和BRISK算子的快速性，从而获得了快速准确的匹配方法。使用SURF-BRISK算法执行针对目标的图像特征提取的初始匹配，并使用汉明距离对初始匹配的特征点执行特征匹配的相似性测量。然后，使用M估计器样本和共识（MSAC）算法执行次级精细匹配，以消除不匹配的点对，从而实现目标图像的识别。然后，通过使用双目立体视觉系统获得工件的三维坐标，以提供位置坐标供机器人准确地抓取工件。在实验中，对使用SURF-BRISK算法获得的目标进行立体视觉匹配，并将目标的位置坐标传递给机器人控制器。实验结果表明，如果忽略了特殊的几何畸变，则该方法可以适用于目标的精确定位。因此，它可以在复杂的环境中识别目标，访问目标的位置坐标，并实时准确地自动抓取工件。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号