Binocular vision positioning method for target grabbing of underwater robot
Binocular vision positioning method for target grabbing of underwater robot
- Neue Suche nach: LUO WENBO
- Neue Suche nach: CAO YANCHAO
- Neue Suche nach: CHEN YUXING
- Neue Suche nach: ZHAO CHAO
- Neue Suche nach: LIN BIN
- Neue Suche nach: LI RUXIN
- Neue Suche nach: PAN KEDA
- Neue Suche nach: LUO WENBO
- Neue Suche nach: CAO YANCHAO
- Neue Suche nach: CHEN YUXING
- Neue Suche nach: ZHAO CHAO
- Neue Suche nach: LIN BIN
- Neue Suche nach: LI RUXIN
- Neue Suche nach: PAN KEDA
The invention discloses a binocular vision positioning method for target grabbing of an underwater robot, and relates to the technical field of binocular vision localization, and the technical key points of the scheme are as follows: generating a stability coefficient according to the stability under a plurality of continuous time sequences in a time window, and when the stability coefficient exceeds a stability threshold value, marking the continuous time sequence in the current time window; extracting images of the marked continuous time sequence frame by frame from the collected video, obtaining the content richness of the images by combining the information entropies and depth information values of the images, and calculating the maximum evaluation value of all video frame images through the content richness of the images and the time complexity required by calculation; the binocular image corresponding to the maximum evaluation value in all the video frame images is used as the positioning image, the spatial position of the target object is calculated, the underwater robot is controlled to capture the target, the accuracy of binocular vision positioning is improved, and good balance is achieved between image content details and calculation complexity.
本发明公开了一种用于水下机器人目标抓取的双目视觉定位方法，涉及双目视觉定位技术领域，该方案的技术要点为：由时间窗口内若干个连续时序下的稳定度生成稳定系数，当稳定系数超过稳定阈值时，标记当前时间窗口内的连续时序；从采集到的视频中对标记的连续时序逐帧提取图像，结合图像的信息熵和深度信息值，获得图像的内容丰富度，通过图像的内容丰富度和计算所需的时间复杂度，计算获得所有视频帧图像的最大评估值；将所有视频帧图像中的最大评估值对应的双目图像作为定位图像，计算目标物体的空间位置，控制水下机器人进行目标抓取，提高了双目视觉定位的准确性，在图像内容细节和计算复杂度之间达到良好的均衡。
Dokumentinformationen
Binocular vision positioning method for target grabbing of underwater robot
一种用于水下机器人目标抓取的双目视觉定位方法
2024-04-09
Patent
Elektronische Ressource
Chinesisch
| IPC: | B25J Manipulatoren , MANIPULATORS / B63C LAUNCHING, HAULING-OUT, OR DRY-DOCKING OF VESSELS , Zuwasserlassen, Trockenstellen, Docken oder Bergen von Schiffen / G06T Bilddatenverarbeitung oder Bilddatenerzeugung allgemein , IMAGE DATA PROCESSING OR GENERATION, IN GENERAL / G06V / H04N PICTORIAL COMMUNICATION, e.g. TELEVISION , Bildübertragung, z.B. Fernsehen |
Ähnliche Titel
Underwater target grabbing method and device and underwater robot
| Europäisches Patentamt | 2024
Stable swimming parent-child robot for grabbing underwater target
| Europäisches Patentamt | 2020
Searching and rescuing underwater robot apparatus with binocular vision
| Europäisches Patentamt | 2015