CN109977891A - A kind of object detection and recognition method neural network based - Google Patents

Abstract

The invention discloses a kind of object detection and recognition methods neural network based comprising the steps of: A, uses the image information in image acquisition device detection zone；B, identifying processing is carried out to acquired image, neural network model C, is established by computer；D, the main modular of convolutional neural networks is defined；E, collected image information is detected using neural network model calculating, and completes numerical result output；F, it makes mark on the image according to testing result, position and the classification of the mark in each target is marked with rectangle frame；The result of judgement detection and identification；G, the result of identification and detection is verified；The present invention is based on the object detection and recognition methods of neural network to carry out recognition of face using object detection and recognition, and arithmetic speed is fast, and judging result is accurate, effectively increases the precision of detection and the efficiency of identification.

Description

A method of target detection and recognition based on neural network

technical field

The invention relates to the technical field of neural networks, in particular to a target detection and identification method based on neural networks.

Background technique

Nowadays, with the rapid development of related theory and application research of computer vision, the advantages of computer vision technology in daily life are increasingly prominent. Recognition of images by computer is a technology in which the computer extracts the corresponding features from the related video or image sequence, so that the computer "understands" the content of the image and can classify it correctly. The improvement of security awareness has also made people's demand for public and personal security continue to rise, which makes computer neural network technology have a high application value in target detection and recognition.

Object detection is an important research topic in the field of computer vision. It has been widely used in many real-world applications, such as face recognition, traffic safety, crowd monitoring and image retrieval. Real-time target detection based on deep learning refers to marking the location and category of target objects in a natural scene picture or video. In the face of massive image and video data, manual labeling is time-consuming and inefficient, and automated and fast object detection methods are urgently needed.

Since the effect of target detection and recognition is easily affected by many factors, the neural network has less stringent requirements for data and has a higher recognition rate. Therefore, the convolutional neural network has important theoretical and practical significance for the research in the field of face recognition.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a target detection and recognition method based on a neural network to solve the problems raised in the background art.

To achieve the purpose, the present invention provides the following technical solutions:

A method for target detection and recognition based on neural network, comprising the following steps:

A. Use the image acquisition device to collect the picture information in the detection area;

B. Recognize and process the collected images,

C. Establish a neural network model through a computer;

D. Define the main modules of the convolutional neural network;

E. Use the neural network model calculation to detect the collected picture information, and complete the numerical result output;

F. Mark the image according to the detection result, mark the position and category of the mark in each target with a rectangular frame; judge the results of detection and recognition;

G. Verify the results of identification and detection.

As a further technical solution of the present invention: the step A is implemented by a camera.

As a further technical solution of the present invention: the step B is specifically: comprehensively comparing the collected images within a certain period of time, and if the comparison results are consistent, it is determined that the pictures within this period of time are static and no dynamic objects appear. , automatically delete the image information collected in this period of time, so it is determined that the image in this period of time is invalid.

As a further technical solution of the present invention: the step E specifically formats the label information of each image and writes it into a txt file, and also extracts image points different from the original image in the image as the difference image .

As a further technical solution of the present invention, it also includes step F: performing algorithm simulation, and building an experimental platform for the improved identification algorithm for verification, wherein the algorithm simulation is realized by MATLAB software.

Compared with the prior art, the beneficial effects of the present invention are as follows: the target detection and recognition method based on the neural network of the present invention adopts target detection and recognition to perform face recognition, the operation speed is fast, the judgment result is accurate, and the detection accuracy is effectively improved. and recognition efficiency.

Detailed ways

The technical solutions in the embodiments of the present invention will be described clearly and completely below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Embodiment 1: a neural network-based target detection and recognition method, comprising the following steps:

A. Use the image acquisition device to collect the picture information in the detection area;

B. Perform identification processing on the collected images, and comprehensively compare the collected images within a certain period of time. If the comparison results are consistent, it is judged that the picture within this period of time is still and no dynamic objects appear, and the image is automatically deleted. The image information collected over a period of time is therefore determined to be an invalid image. When the comparison results are inconsistent, the image with the difference is extracted, and the image in this period of time is determined to be the image to be recognized.

C. Establish a neural network model through a computer;

D. Define the main modules of the convolutional neural network;

E. Use the neural network model calculation to detect the collected picture information, and complete the output of the numerical results; format and write the label information of each image into a txt file, and at the same time also change the image in the image different from the original picture. The image points are extracted as the difference image;

F. Mark the image according to the detection result, mark the position and category of the mark in each target with a rectangular frame; judge the results of detection and recognition;

G. Verify the results of identification and detection.

H. Carry out algorithm simulation, and build an experimental platform for the improved identification algorithm to verify, wherein the algorithm simulation is realized by MATLAB software.

Wherein, step A is implemented by a camera. You can also use mobile phones, digital cameras and other devices to achieve.

It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, but that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments are to be regarded in all respects as illustrative and not restrictive, and the scope of the invention is to be defined by the appended claims rather than the foregoing description, which are therefore intended to fall within the scope of the claims. All changes within the meaning and scope of the equivalents of , are included in the present invention.

In addition, it should be understood that although this specification is described in terms of embodiments, not each embodiment only includes an independent technical solution, and this description in the specification is only for the sake of clarity, and those skilled in the art should take the specification as a whole , the technical solutions in each embodiment can also be appropriately combined to form other implementations that can be understood by those skilled in the art.

Claims (5)

1. a kind of object detection and recognition method neural network based, which is characterized in that comprise the steps of:

Use the image information in image acquisition device detection zone；

Identifying processing is carried out to acquired image,

Neural network model is established by computer；

Define the main modular of convolutional neural networks；

Collected image information is detected using neural network model calculating, and completes numerical result output；

It makes mark on the image according to testing result, position and the classification of the mark in each target is marked with rectangle frame；Sentence The result of disconnected detection and identification；

The result of identification and detection is verified.

2. a kind of object detection and recognition method neural network based according to claim 1, which is characterized in that described Step A is realized by camera.

3. a kind of object detection and recognition method neural network based according to claim 2, which is characterized in that described Step B is specifically: the image in collected certain time being carried out global alignment, if the comparison results are consistent, then is judged as Picture still in this time occurs without dynamic object, is automatically deleted this section of time acquired image information, therefore sentence Fixed this section of temporal image is invalid image, when comparison result is inconsistent, then comes out the image zooming-out for difference occur, determines the section The image of time is images to be recognized.

4. a kind of object detection and recognition method neural network based according to claim 3, which is characterized in that described Step E be specifically the label information of every image is formatted and is written in a txt file, while also by the image with original The different picture point of beginning picture extracts, as difference image.

5. a kind of object detection and recognition method neural network based according to claim 4, which is characterized in that also wrap It includes step H: carrying out algorithm simulating, and build experiment porch for improved recognizer and verified, wherein algorithm simulating is logical Cross MATLAB software realization.