Blindman assisting device based on object identification

Abstract

A life assistance device for the blind based on object recognition, wherein: the object information buffer unit stores the type information of each object detected in the environment, the position coordinates in the image, and the three-dimensional position information, feature information and attention level of the relative blind person ; The object detection unit detects whether there is an object that can be detected in the detection area of ​​the image sensor, compares whether the detected object is an object that has been recorded, and compares the position coordinates of the newly detected object in the image and the category of the object The information is stored in the object information buffer unit, and the highest attention level is given; the object ranging unit measures the three-dimensional position information of each object in the object information buffer unit relative to the blind person; the sensory replacement unit compares the objects in the object information buffer unit to the blind person The requirements of the system are converted into sound signals or tactile signals, so that blind people can perceive the existence of objects in the environment. The invention improves the efficiency of blind people's cognitive environment.

Description

Technical field [0001] The present invention relates to a life assistance device for the blind in the field of information technology, in particular to a life assistance device for the blind based on object recognition. Background technique [0002] With the development of computer vision, pattern recognition technology and multimedia technology, people have begun to try to use informationized electronic devices to assist the blind in their lives. In the field of life assistance for the blind, hearing or touch is generally used instead of sight to convey information to the blind (the term is called "sensation substitution"). However, because human information acquisition relies heavily on vision, more than 80% of human information is acquired through visual channels, and hearing and touch convey very little information. Therefore, how to use sound or touch to properly represent the information in the environmental image? Making blind people get the most important information in t...

AI Technical Summary

Problems solved by technology

But its main problem is: first, when there are multiple objects in the environment, the blind person will hear the sound of all objects in the environment at the same time, making it difficult for the blind person to distinguish the sound of a single object

Second, if in practice the sounds heard by the blind person are made to have a certain interval before and after so that the blind person has time to hear each sound, the time taken to represent multiple objects in the entire environment will be extended, which undoubtedly reduces the blind person's The speed of obtaining environmental information

Third, when the position of the blind person changes or the objects in the environment move, new objects appear in the environment image. However, in the above invention, since the methods of representing each object after recognition are consistent, the newly recognized objects are very easy to recognize. Difficult to distinguish from objects that have already been identified

Finally, when the object detection unit of the above invention detects multiple objects of the same type in the environment image, a false negative error occurs, blind people may regard the recognized objects in the environment as the same object, causing the original object to move. illusion

[0004] The difficulty of the existing technology is that although there are ways to obtain the types and positions of objects in the environment, since the amount of information transmitted by hearing is much smaller than that of vision, it is necessary to consider how to use a more efficient way to inform blind people that they can know the obstacles in the surrounding environment Information

Images