نوع مقاله : مقاله پژوهشی
نویسندگان
مجتمع مواد و فناوریهای ساخت، دانشگاه صنعتی مالک اشتر، تهران، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
It is worth mentioning that the exact volume of damages in parts that have been repaired and repaired has an important role in improving the quality of repair. In this paper, a method for identification of corrosion geometry of parts using intelligent tools is presented. At first, by using photogrammetry, point cloud has been generated from damaged areas. Then by clustering of points cloud, the design of the damage is determined. to obtain the amount of damages in the damage, a plate is required to coincide with the boundary edge of the damage, with the plate matching, the enclosed space between the interior surface of the damage and the plate accordingly. The clustering of point cloud was performed by k - means and the adaptation of the plate with the points cloud was performed using several characteristics of segment image including damage center and upper boundary line. The segmentation is performed with two methods (mask r - cnn) and (Segment any thing) that training the first model is done using 900 data. At the end, the path of the deposition tool was created in the produced volume and the simulated mechanism for the robot arm was designed to move the laser system. This method can be used for small samples, especially in cases where there is no healthy sample.
کلیدواژهها [English]
)