What is the difference between a monocular camera and a binocular camera?

Perceptual camera is an essential hardware device for smart cars. Monocular and binocular are the differences in appearance between single camera and dual camera. Vehicle-mounted monocular vision system is the processing of plane images, and the sensing tasks are lane line detection, traffic sign recognition and obstacle detection (traffic participants such as vehicles and pedestrians). Vehicle-mounted binocular vision system is to construct stereoscopic images of objects by imitating human eyes, and its sensing tasks are target detection, classification ranging, multi-target tracking, traffic space and scene understanding.

It is mainly used for collision warning, adaptive cruise (ACC), lane departure warning (LDW), distance monitoring (HMW), lane keeping (LKA), collision warning (FRW), 360-degree panoramic image, transparent A-pillar, electronic rearview mirror, etc.

Advantages and disadvantages of monocular and binocular vision systems;

The ranging method of binocular vision system has the advantages of high efficiency, high precision, simple system structure and low cost, but the computational ranging means a large amount of calculation and it is difficult to be miniaturized. The disadvantage of vehicle-mounted monocular vision system is that it needs a lot of data, which needs updating, maintenance and continuous optimization.

First, the vehicle monocular vision system workflow:

1. Image acquisition: converting analog images into digital images;

2, image preprocessing: sorting out each character image and giving it to the recognition module for recognition, and performing image denoising, edge enhancement, gray-scale stretching, image segmentation and morphological processing;

3. Feature extraction: use computer to extract image information to determine whether each image point belongs to an image feature, such as shape, area, volume, color, motion state, symmetry, surface roughness, acousto-optic reflectivity, penetrability, etc.

4. Target recognition: It refers to the process of distinguishing a special target (or a class of targets) from other targets (or other types of targets).

Second, the vehicle binocular vision system workflow:

1, camera calibration: establish the geometric model of camera imaging, which is simply understood as three-dimensional coordinates;

2. Stereo matching: matching the corresponding point A 1 (X 1, Y 1, z1) between the left and right image planes;

3. Stereo restoration: the stereoscopic three-dimensional image of the image is restored by using the multi-layer scanning confocal image synthesis technology;

4. Video analysis: a technology that uses computer vision technology to monitor moving objects and analyze, track or identify them.