416 / 2019-02-28 22:02:39

The automatic aiming technology based on the machine learning

automatic aiming, machine learning

全体主题 > Laser and Particle Beam Fusion

In a large laser device, the diagnostic device is mounted on a diagnostic carrying platform. The aiming accuracy of the platform is greatly influenced by the accuracy of the diagnostic data. At present, the platform adopts aiming method of optical imaging binocular CCD intersection, which imaging the target on the CCD. Aiming is performed by manually interpreting on the image.At this stage, the target room lighting environment is poor, which also increases the difficulty of aiming and has an impacts on the aiming accuracy and efficiency. Therefore, an image recognition method based on machine learning is proposed to achieve automatic target aiming. The image is pre-processed, marked, and then trained in the model to find the parameter configuration of the optimal model, and then the target point of the image is automatically recognized by the model. In the application, the aim of automatic aiming is achieved by narrowing the error between the target point and the offline calibration point.