Convolutional neural network-based image geographic positioning system and method

Abstract

The invention provides a convolutional neural network-based image geographic positioning system and method, and relates to the field of satellite positioning. The system is characterized by comprising a video uploading module, a picture uploading module, an image frame extraction module, a geographic mark detection module, a decomposition module, a feature extraction module, an activation function, a pooling module, a full connection module, a classifier analysis and judgement module, a learning module, a regional database, a pixel database and a result display module. The system and method provided by the invention have the advantages of rapid in image analysis and correct in image recognition, positioning video information and having learning ability.

Description

technical field [0001] The invention relates to the field of satellite positioning, in particular to an image geographic positioning system and method based on a convolutional neural network. Background technique [0002] For image positioning, if the scene being pushed is a well-known scenic spot or a landmark building, then we can know it at a glance through the special scene mark. For the location of photos of common locations, people generally mark them by adding geotags to photos. For example, when many Android phones are taking pictures, after activating the camera and entering the shooting menu, there will be a "location" function. We only need to enable this function before shooting. [0003] The photos taken in this way will automatically add the local geographic location information. When we view these photos on the computer, switch to "Detailed Information", and under the GPS item, we can see the actual geographic location of the photo. Here Use GPS latitude and...

AI Technical Summary

Problems solved by technology

[0005] 1. The accuracy of image recognition is not enough: Most of the existing image positioning systems store a large number of samples directly in the database, and different images have various characteristics

Therefore, directly comparing the samples will lead to very poor comparison effect and very low accuracy.

[0006] 2. Lack of learning ability: the existing image positioning lacks the learning ability in the actual use process, no matter what algorithm and analysis and judgment method is used, it will always lead to deviations in image positioning. If you cannot continue to learn and improve during use , will cause the entire image positioning system to stagnate

[0007] 3. Image analysis is slow: the existing image positioning system adopts traditional image analysis algorithms, which follow the commonly used modes of image decomposition and in-depth analysis.

The positioning speed of a picture is very slow, and due to the limitations of the algorithm, the positioning results are often inaccurate

[0008] 4: Does not have the positioning of video information: the existing image positioning system basically does not have the function and means of positioning video information

Images