Target detection method and device

Abstract

The invention discloses a target detection method and device, and relates to the technical field of computers. A specific embodiment of the method comprises the steps of performing image detection ona collected original image to obtain category information of a target and a bounding box in the original image; clustering pixels belonging to the same target in the bounding box based on the depth information corresponding to the original image so as to perform region division on the target in the bounding box; and counting the depth value distribution of the pixels in the bounding box to obtainan area corresponding to the pixels of which the depth values are in a set depth interval, and determining the position information of the target according to the area. The bounding box is obtained byperforming image detection on the original image, the bounding box is subjected to area division by adopting a clustering mode, and then the non-target area in the bounding box is filtered and removed and the target area is reserved in combination with the counted depth value of the pixels in the bounding box, so that the precision of the bounding box of the bounding target is improved so that target detection precision is improved.

Description

technical field [0001] The invention relates to the field of computers, in particular to a target detection method and device. Background technique [0002] Unmanned vehicles face the problem of how to avoid obstacles while driving on the road. Target detection is an important part of the interaction between unmanned vehicles and the surrounding environment, and it is also the main basis for behavior prediction and decision planning. Target detection The effect will directly affect the overall performance level of unmanned vehicles. Currently widely used target detection methods include: target detection based on lidar, target detection based on millimeter-wave radar, target detection based on monocular, etc. [0003] In the course of realizing the present invention, the inventor finds that there are at least the following problems in the prior art: [0004] For target detection based on laser radar, the cost of the sensor itself is high, it is not easy to obtain the categ...

AI Technical Summary

Problems solved by technology

[0004] For target detection based on laser radar, the cost of the sensor itself is high, it is not easy to obtain the category information of the target, and the recognition ability of long-distance targets is weak; the target detection based on millimeter wave radar has low detection accuracy and cannot detect pedestrians; based on single purpose Target detection, difficult calibration with laser radar, etc., low detection accuracy and large error

Images