Research on Image Detection and Application of Security Dangerous Goods Based on SSD

SSD (Single Shot MultiBox Detector) is a widely used target detection framework. Based on the Object Detection API based on the TensorFlow framework and the SSD algorithm, this paper uses SSD to realize the target detection of security dangerous goods images. The experimental results show that the SSD-based image neural network detection model can fully utilize the flexibility of TensorFlow and SSD, effectively improve the stability, speed effectiveness and accuracy of image recognition, and the average detection accuracy rate is over 90%.