Faster-RCNN-based people number detection system

Abstract

The invention relates to the technical field of CAE (Computer Aided Engineering) data, and discloses a Faster-RCNN (Regional Convolutional Neural Network)-based people number detection system which comprises the following steps: S1, acquiring people data samples, and making the data samples into a data set; s2, dividing the data set into a training set and a test set; s3, adopting the training set to train the Faster-RCNN network; and S4, performing detection effect evaluation on the model obtained by training. By detecting the number of people with small mobility in a small area, the detected number of people is fed back to the user, so that the user can quickly and conveniently obtain the number of people in the area; the obtained model is subjected to effect evaluation, the model subjected to character recognition training is subjected to intra-region character detection, characters detected by the recognition model are counted, and the number of people obtained through statistics is fed back to the user, the system can effectively detect and count the characters, the number of people in the region can be accurately fed back to the user, and the user experience is improved. And the user can quickly and conveniently know the number of people in the area and carry out corresponding work.

Description

technical field [0001] The invention relates to the technical field of Faster-RCNN number detection, in particular to a number detection system based on Faster-RCNN. Background technique [0002] In recent years, with the development of disciplines such as computer vision, artificial intelligence, and deep learning, people counting methods have developed rapidly. Compared with the traditional manual counting method, visual people counting has the advantages of quickness and convenience. This paper proposes a population detection system based on the Faster-RCNN network. [0003] The detection of the number of people mainly involves the detection of deep learning targets. In the field of deep learning, the classic detection methods include one-stage one-stage and two-stage two-stage. The single-stage method is mainly the YOLO series, and the two-stage method is the Faster-RCNN series. The advantage of the yolo series in the detection core is that it is fast and suitable for...

AI Technical Summary

Problems solved by technology

The advantage of the yolo series in the detection core is that it is fast and suitable for real-time detection tasks, but compared with the two-stage Faster-RCNN series, the detection effect is generally not good

Although the Faster-RCNN series is not as fast as the YOLO series in terms of detection speed, its detection effect is better.

Images