Digital instrument character detection method based on SW-FSSD

The invention discloses a digital instrument character detection method based on an SW-FSSD, and the method comprises the steps: constructing an SW-FSSD model as a digital instrument character detection model, and carrying out the model training; acquiring a character image of the digital instrument by using an IPC; inputting the character image into the trained digital instrument character detection model; and performing feature extraction and classification on the character image by using the digital instrument character detection model, performing feature fusion on high and low layers of the model to obtain a character category, and outputting a character detection result. The accuracy and effectiveness of digital instrument character detection can be improved, and the requirement for intelligent detection of digital instrument characters of a traction substation is met. 本发明公开了一种基于SW-FSSD的数字仪表字符检测方法，其包括：构建SW-FSSD模型作为数字仪表字符检测模型，并进行模型训练；利用IPC采集数字仪表的字符图像；将字符图像输入训练好的数字仪表字符检测模型；利用数字仪表字符检测模型对字符图像进行特征提取并分类，将模型的高低