Binocular stereoscopic vision target detection equipment and method

Abstract

The invention discloses binocular stereoscopic vision target detection equipment and method. A binocular camera in the equipment collects a left image and a right image; a programmable logic processoradopts a preset similarity matching algorithm to obtain a parallax value of a pixel pair and a pixel pair which are matched with each other, at least one pixel point in the left image to be detectedand at least one pixel point in the right image to be detected; a preset neural network target object detection algorithm is adopted to obtain a target object in the to-be-detected left side image with preset resolution and attribute information of the target object; a distance value between the target object and the vehicle is obtained according to a parallax value corresponding to a target pixelpoint of the target object in the parallax values of the pixel point pairs and a preset camera parameter; and the application processor outputs the attribute information of the target object and thedistance value between the target object and the vehicle. Therefore, the device obtains the parallax value and the distance value of the target object through the embedded programmable logic processor, thereby reducing the power consumption and the occupied space volume.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a binocular stereo vision target detection device and method. Background technique [0002] At present, the vision system has been widely used in the fields of intelligent detection system and intelligent equipment, and has become an important standard to measure the performance of intelligent equipment. Target detection technology is one of the key technologies of the vision system of intelligent equipment. Traditional target detection technology relies on monocular vision and has less information, while binocular stereo vision (Binocular Stereo Vision) positioning technology can obtain the distance value from a pair of corresponding images, but due to its complex calculation, it is not suitable for real-time calculation . Binocular stereo vision is an important form of machine vision. It is based on the principle of parallax and uses imaging equipment to obtain two...

AI Technical Summary

Problems solved by technology

[0004] However, the inventors found that the GPU has the disadvantages of high power consumption and high heat generation during parallel data processing, and the host computer matched with the GPU is generally large in size, and there is a space limitation in the application on the aircraft or vehicle, which reduces the user experience

Images