Obstacle detection method and device, storage medium and mobile robot

Abstract

The invention discloses an obstacle detection method and device, a storage medium and a mobile robot. The obstacle detection method specifically comprises the steps that an advancing area image shot by a declination camera device installed on an AGV is acquired; determining the lower part in the advancing area image as a reference area; determining the rest part above the reference area in the advancing area image as a detection area; detecting an object region and a shadow region in the detection region; when it is detected that the object area and the shadow area exist in the detection areaat the same time, whether the position relation between the object area and the shadow area is matched with the light projection direction or not is detected; and when the position relationship between the object area and the shadow area is matched with the light projection direction, marking the object area as a suspected obstacle. According to the embodiment of the invention, the position information of the three-dimensional obstacle is efficiently judged by analyzing the acquired advancing area image, so that a large amount of data is prevented from being used for calculation or training, and relatively high calculation efficiency is realized.

Description

technical field [0001] The present application relates to the field of robot technology, in particular to a method, device, storage medium and mobile robot for detecting obstacles. Background technique [0002] An important symbol of the intelligentization of mobile robots (Automated Guided Vehicle, AGV) is autonomous navigation, and obstacle avoidance is the main research topic in autonomous navigation. Obstacle avoidance generally means that the mobile robot effectively avoids obstacles according to a certain method when it senses static and dynamic objects that hinder its passage through sensors during the walking process, according to the state information of the collected obstacles, and finally reaches the target point. Mobile robots usually use sensors to obtain information about the surrounding environment, including the size, shape, and location of obstacles. At present, the mainstream perception and obstacle avoidance technology solutions of mobile robots are mainl...

AI Technical Summary

Problems solved by technology

Secondly, the amount of calculation is large. Obtaining 3D information of the real world through image or point cloud computing, and processing this information will generate a huge amount of calculation.

Finally, the amount of data required is huge. For example, using deep learning methods requires a large amount of manually marked training data to train the neural network, which makes the manual workload heavy.

Images