A method and system for object detection based on depthwise separable convolution

A target detection and in-depth technology, applied in the field of neural networks, can solve the problems of insufficient computing power and high computational complexity of the robot platform, and achieve the effect of low computational cost, high efficiency and accuracy, and effective and accurate detection of target objects

CN109472315BActive Publication Date: 2021-09-24JIANGSU MUMENG INTELLIGENT TECH
10 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Publication Date
2021-09-24

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The present invention provides a target detection method and system based on depth-separable convolution, the method comprising: acquiring image data; the image data includes a target object; inputting the image data into a depth-separable neural network, extracting Image features in the image data; fusion detection is performed according to image features of different levels, and a prediction result of the target object is output. The present invention solves the problem of slow speed when using standard convolutional neural network for target object recognition. After using different levels of image features for detection, feature fusion is performed, which ensures the high efficiency and accuracy of target object detection and realizes low calculation cost. , Effectively and accurately detect the purpose of the target object.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the field of neural networks, in particular to an object detection method and system based on depthwise separable convolution. Background technique

[0002] When a robot is running, it needs to consider its safety, efficiency, intelligence, etc. Therefore, it is a development trend to have a certain understanding of its environment, and it is also an inevitable technical requirement for the real application of robots.

[0003] Visual information is of great significance and function to robots. From the perspective of information volume, the amount of visual information is very rich. Just as human beings rely on visual information to a large extent to understand the world, visual information is crucial to understanding the surrounding environment. From the perspective of cost, the current image acquisition equipment achieves high-speed high-definition and low cost. Compared with the expensive lidar, image acquisition equipment ...

Examples

Embodiment Construction

[0069] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the specific implementation manners of the present invention will be described below with reference to the accompanying drawings. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention, and those skilled in the art can obtain other accompanying drawings based on these drawings and obtain other implementations.

[0070] In order to make the drawing concise, each drawing only schematically shows the parts related to the present invention, and they do not represent the actual structure of the product. In addition, to make the drawings concise and easy to understand, in some drawings, only one of the components having the same structure or function is schematically shown, or only one of them is marked. Herein, "a" not only means "only one", but also means "more than one".

[0071] An embodim...