Water surface target detection method based on grey prediction auxiliary area suggestions
The invention discloses a water surface target detection method based on grey prediction auxiliary area suggestions, and belongs to the field of intelligent unmanned ships. The method comprises the following steps: firstly, detecting a water surface target by using a trained model; then, carrying ou...
Saved in:
Main Authors | , , , , , , , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
10.09.2019
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention discloses a water surface target detection method based on grey prediction auxiliary area suggestions, and belongs to the field of intelligent unmanned ships. The method comprises the following steps: firstly, detecting a water surface target by using a trained model; then, carrying out gray prediction by using the water surface target position information in the continuous video frames; thirdly, feeding back and guiding regional suggestions of the neural network by using a grey prediction result; and finally, identifying the water surface target newly entering the video frame through the more accurate area suggestion result obtained in the previous step. According to the method, the water surface target identification accuracy and speed can be improved by simplifying the neural network structure and providing more accurate regional suggestions.
本发明公开一种灰色预测辅助区域建议的水面目标检测方法,属于智能无人船舶领域。本发明包括:首先,利用已训练好的模型检测到水面目标;然后,利用连续视频帧中的水面目标位置信息进行灰色预测;接着,运用灰色预测的结果对神经网络的区域建议进行反馈和引导;最后,通过上一步得到的更加准确的区域建议结果识别新入视频帧中的 |
---|---|
Bibliography: | Application Number: CN201910480821 |