Mountain farmland environment identification method and device, electronic equipment and storage medium

Abstract

The invention provides a mountain farmland environment recognition method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a mountain farmland image collected based on a binocular camera; inputting the mountain farmland image into a preset image segmentation model to obtain object types contained in the mountain farmland image; based on object types contained in the mountain farmland image, determining information of an environment where an operation vehicle is located; the image segmentation model is obtained by training a target neural network based on an image sample marked with a category. According to the mountain farmland environment identification method and device, the electronic equipment and the storage medium provided by the invention, the image segmentation model is obtained through training by collecting the image samples corresponding to various mountain farmland environments, and the image segmentation model obtained through training is deployed in the operation vehicle; therefore, the safe driving area and the dangerous area in the environment where the operation vehicle is located can be determined, and the safety of the operation vehicle can be guaranteed.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a method, device, electronic device and storage medium for recognizing a mountain farmland environment. Background technique [0002] When working in mountain farmland, operation vehicles are generally used for operation. The terrain of mountain farmland is complex, resulting in high management and production costs. Terraced orchards are a typical representative of mountain farmland. The terrain is diverse and the environment is changeable. When operating vehicles are operating in mountain farmlands, they need to be able to accurately identify the environment where the vehicles are located, so as to exclude possible safety problems during normal driving. hidden dangers to ensure safe operation. [0003] At present, when operating vehicles operate in mountain farmland, lidar or 3D cameras are generally used to identify the terrain of mountain farmland. Since the point c...

AI Technical Summary

Problems solved by technology

[0003] At present, when operating vehicles work in mountainous farmland, laser radar or 3D cameras are generally used to identify the terrain of mountainous farmland. Since the point cloud information obtained by laser radar cannot accurately reflect the detailed information of the ground, the operating vehicle is driving on the mountainous farmland ground. , a safety hazard

Recognition of mountainous farmland terrain based on 3D cameras uses threshold method and clustering algorithm to process data, and it is difficult to effectively identify ground obstacles in mountainous farmland, which affects the safe driving of operating vehicles

Images