Human and face detection method based on point cloud and camera fusion

Abstract

The invention relates to the technical field of human and face detection, in particular to a human and face detection method based on point cloud and camera fusion. The method comprises the followingsix steps: acquiring three-dimensional point cloud information in real time, clustering specific positions of human bodies, detecting and tracking in real time, estimating positions of human bodies and human heads, capturing pictures of corresponding human and portrait information, extracting human body images and human faces, and identifying the extracted human body images and human faces. Pointcloud data generated by a laser radar, a binocular camera and a ToF depth sensor are adopted to cluster the specific orientation of a human body, then the positions of the human body and the head of the human body are estimated through point cloud recognition, scale estimation and other methods, and dynamic human and portrait information is captured from the corresponding view angle of the camera.Human and face recognition is carried out through the convolutional neural network model, and compared with a pure image mode, the algorithm reduces the risk of false face, can prevent countermeasureattacks, and guarantees the safety and practicality of a detection system.

Description

technical field [0001] The invention relates to the technical field of human and human face detection, in particular to a human and human face detection method based on point cloud and camera fusion. Background technique [0002] Human and face detection is an important problem in the sub-direction of target detection. It is widely used in security monitoring, witness comparison, human-computer interaction, social interaction and entertainment. step. As an important detection device, camera has been widely used in face recognition. For example, people and portraits are captured at key checkpoints such as inside the car (including trains, buses, light rail, smart rail, etc.), stations (including bus stations, subway stations, high-speed rail stations, etc.). But the existing camera-based methods have an important flaw, such as: the use of adversarial attacks can make people invisible. Improper use can easily lead to misjudgment and false alarms. [0003] The combination o...

AI Technical Summary

Problems solved by technology

But the existing camera-based methods have an important flaw, such as: the use of adversarial attacks can make people invisible

Improper use can easily lead to misjudgment and false alarms

[0003] The combination of camera and 3D sensor can effectively solve this problem, but the existing system usually detects and fuses and judges through each channel separately, and does not form an effective discrimination method

Images