论文部分内容阅读
利用计算机人工智能与神经网络技术,对于每张金相图像照片中的神经元进行图像识别,从点状非金属夹杂物入手,设计一种新型检测点状非金属夹杂物的方法,克服图像中的干扰项,从中找出真正的非金属夹杂物,构造出一套能自动检测非金属夹杂物的计算机软件模型;对非金属夹杂物检测方法和手段所涉及的关键技术进行分析探讨,并根据具体钢材的金相图运行程序,观察检测结果,最后检验该模型的正确率,得到可行性的解决方案,为非金属夹杂物自动检测与甄别的模块化和规范化作出有益尝试。
Using computer artificial intelligence and neural network technology, the neurons in each photo of metallographic image are identified by image recognition. From the point of non-metallic inclusions, a new method to detect point-like non-metallic inclusions is designed to overcome the interference in the image Item, find the real non-metallic inclusions, construct a set of computer software model that can automatically detect non-metallic inclusions; analyze and discuss the key technologies involved in the detection methods and means of non-metallic inclusions, and according to the specific steel The metallographic diagram of the running program to observe the test results, and finally test the correctness of the model to be feasible solutions to non-metallic inclusions automatic detection and identification of modular and standardized make a useful attempt.