FIELD: invention relates to the field of obstacle warning devices.
SUBSTANCE: method includes obtaining an image of the surrounding space; selecting the best quality image for a given period of time, determining the center and area of the selected image, detecting objects in the image; comparison of images of identified objects with stored reference images of physical objects, to assign identified objects to objects of the first or second type, for each of the identified objects of the first type, the implementation of the definition of a frame that limits the identified object; determination of the frame area; comparing the area of the frame that bounds the detected object with the stored reference values for different physical objects using a heuristic algorithm, if the area of the frame that bounds the detected object is greater than the reference value, determining the center of the frame that bounds the detected object, based on the center of the frame determination on the image of the surrounding space of the location of the identified object, relative to the center of the specified images, generating a warning about the location of the object, if the detected object occupies more than a quarter of the image area; moreover, the detection of objects and the determination of the frame that bounds the object is carried out using a neural network that is trained on a data set that includes images of physical objects.
EFFECT: technical result is to increase the accuracy of timely determination of the location of an object that can interfere with the movement of a visually impaired user, in particular, in conditions of fast walking or running.
9 cl, 1 dwg
Title | Year | Author | Number |
---|---|---|---|
METHOD FOR DETECTING OBJECTS IN THE IMAGE OF THE PLAN-SCHEME OF THE CONSTRUCTION OBJECT | 2022 |
|
RU2785821C1 |
METHOD AND SYSTEM OF ACCURATE LOCALIZATION OF VISUALLY IMPAIRED OR BLIND PERSON | 2017 |
|
RU2681346C2 |
METHOD FOR TRACKING OBJECTS AT THE RECOGNITION STAGE FOR UNMANNED VEHICLES | 2022 |
|
RU2798739C1 |
TRACELESS IMAGE CAPTURE, USING MOBILE DEVICE | 2020 |
|
RU2787136C2 |
METHOD AND ELECTRONIC DEVICE FOR DETECTING THREE-DIMENSIONAL OBJECTS USING NEURAL NETWORKS | 2021 |
|
RU2776814C1 |
METHOD AND SYSTEM FOR DETERMINING THE LOCATION OF THE USER | 2020 |
|
RU2759773C1 |
METHOD FOR DETERMINING DEFECTS IN THE IMAGE OF COMPOSITE PRODUCTS | 2022 |
|
RU2807288C1 |
METHOD, APPARATUS AND SYSTEM FOR MONITORING THE WORKING CONDITION OF HEAVY MACHINERY SUCH AS A MINING EXCAVATOR | 2018 |
|
RU2772929C1 |
DEVICE AND METHOD FOR HIP JOINT DIAGNOSIS | 2022 |
|
RU2795658C1 |
SYSTEM AND METHOD FOR DIAGNOSTICS OF HIP JOINTS | 2022 |
|
RU2801420C1 |
Authors
Dates
2022-12-14—Published
2022-11-14—Filed