基于树莓派的服装识别系统设计

Design of clothing identification system based on Raspberry Pi

针对用户家中服装存放无序、分类管理难度大且效率低下等问题,提出了一种基于树莓派的服装识别系统,该系统使用树莓派作为嵌入式开发板。系统通过树莓派调用摄像头获取待存放的服装图片,将图像传输到已训练好的MobileNet模型为核心的服装识别系统中进行类型识别;最后将服装识别的结果和相对应的储存位置通过语音的方式告知用户,以便用户快速准确地找到服装储存的位置。经实验测试,服装识别的正确率为95.19%。系统能够在准确识别服装的基础上,获取该类服装在衣橱中的存放位置,以帮助用户高效、准确地分类管理不同款式的服装。

In order to solve the problems of disordered storage,difficult classification management and low efficiency of users'clothing at home,a clothing identification system based on Raspberry Pi is proposed.The system uses raspberry pi as an embedded development board.This system uses the camera of Raspberry Pi to get the clothes pictures to be stored,and then the image is transmitted to the clothing identification system based on MobileNet model.Finally,the results of clothing recognition and corresponding storage location are communicated to the user by voice,so that users can quickly and accurately find the storage location of clothing.Through the experimental test,the correct rate of clothing recognition is 95.19%.On the basis of accurately identifying the clothes,the storage location of such clothes in the wardrobe can be obtained.Thus,it can help users to classify and manage different styles of clothes efficiently and accurately.