CN108038849A - A kind of excellent robotic vision system of recognition performance - Google Patents
A kind of excellent robotic vision system of recognition performance Download PDFInfo
- Publication number
- CN108038849A CN108038849A CN201711289370.1A CN201711289370A CN108038849A CN 108038849 A CN108038849 A CN 108038849A CN 201711289370 A CN201711289370 A CN 201711289370A CN 108038849 A CN108038849 A CN 108038849A
- Authority
- CN
- China
- Prior art keywords
- mrow
- target
- msub
- msubsup
- module
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Withdrawn
Links
- 238000011156 evaluation Methods 0.000 claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 11
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000005352 clarification Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims 1
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/12—Edge-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of excellent robotic vision system of recognition performance, including image capture module, image pre-processing module, characteristic extracting module, picture recognition module and identification and evaluation module, described image acquisition module is used to be acquired image by camera, described image pretreatment module is used to pre-process the image of collection, the characteristic extracting module is used to extract pretreated characteristics of image, obtain target feature vector, described image identification module is matched according to target feature vector and target template, target is identified, the identification and evaluation module is used to evaluate the recognition performance of picture recognition module.Beneficial effects of the present invention are：Realize the evaluation of identification and recognition performance of the robot to target.
Description
Technical field
The present invention relates to robotic technology field, and in particular to a kind of excellent robotic vision system of recognition performance.
Background technology
One of the main means of machine vision as acquisition environmental information, can increase the capacity of will of industrial robot,
Improve its flexibility.Industrial robot, how according to acquired image information, correctly, is extracted go to work in real time by vision
Part characteristic parameter simultaneously judges the location of workpiece, is machine vision applications in one of key technology of industrial circle.
The method of image recognition is very much, can generally be summarized as statistical picture identification, structural images identification, fuzzy set figure
As identification, images match identification.Images match is method that is wherein most representative, being most widely used, it is in moving target
The fields such as tracking, remote sensing images identification, robot vision have all been applied, but existing recognition methods recognition performance is poor,
Can not effectively it be identified.
The content of the invention
A kind of in view of the above-mentioned problems, the present invention is intended to provide excellent robotic vision system of recognition performance.
The purpose of the present invention is realized using following technical scheme：
Provide a kind of excellent robotic vision system of recognition performance, including image capture module, image preprocessing mould
Block, characteristic extracting module, picture recognition module and identification and evaluation module, described image acquisition module are used for by camera to figure
As being acquired, described image pretreatment module is used to pre-process the image of collection, and the characteristic extracting module is used for
Pretreated characteristics of image is extracted, obtains target feature vector, described image identification module according to target signature to
Amount and target template are matched, and target is identified, and the identification and evaluation module is used for the identification to picture recognition module
Performance is evaluated.
Beneficial effects of the present invention are：Realize the evaluation of identification and recognition performance of the robot to target.
Brief description of the drawings
Using attached drawing, the invention will be further described, but the embodiment in attached drawing does not form any limit to the present invention
System, for those of ordinary skill in the art, without creative efforts, can also obtain according to the following drawings
Other attached drawings.
Fig. 1 is the structure diagram of the present invention；
Reference numeral：
Image capture module 1, image pre-processing module 2, characteristic extracting module 3, picture recognition module 4, identification and evaluation mould
Block 5.
Embodiment
The invention will be further described with the following Examples.
Referring to Fig. 1, a kind of excellent robotic vision system of recognition performance of the present embodiment, including image capture module 1,
Image pre-processing module 2, characteristic extracting module 3, picture recognition module 4 and identification and evaluation module 5, described image acquisition module 1
For being acquired by camera to image, described image pretreatment module 2 is used to pre-process the image of collection, institute
State characteristic extracting module 3 to be used to extract pretreated characteristics of image, obtain target feature vector, described image identification
Module 4 is matched according to target feature vector and target template, and target is identified, and the identification and evaluation module 5 is used for
The recognition performance of picture recognition module 4 is evaluated.
The present embodiment realizes the evaluation of identification and recognition performance of the robot to target.
Preferably, described image pretreatment module 2 includes first object segmentation submodule and the second pretreatment submodule, institute
State first object segmentation submodule to extract the outer edge of target in image using Canny operators, by multiple mesh in image
Mark is divided into simple target image, and the second pretreatment submodule carries out at greyscale transformation and filtering simple target image
Reason.
Due to difference between different target, this preferred embodiment first object splits submodule by the multiple target in image point
Multiple simple targets are cut into, each target image is pre-processed, easy to subsequent extracted different target feature；Due in image
In gatherer process, inevitably it is subject to various interference and is mixed into random noise, this preferred embodiment second pre-processes submodule
Gray processing and filtering process are carried out to image, helps to reduce identification error, preferably keeps the marginal information of target.
Preferably, the characteristic extracting module 3 includes fisrt feature extracting sub-module, second feature extracting sub-module and spy
Sign vector generation submodule, the fisrt feature extracting sub-module are used for the fisrt feature for extracting target, and the second feature carries
Submodule is taken to be used for the second feature for extracting target, described eigenvector generates submodule and given birth to according to fisrt feature and second feature
Into clarification of objective vector.
The fisrt feature extracting sub-module is used for the fisrt feature for extracting target, is specially：Extract the interior outside of target
Edge, obtains the coordinate of target outer contour pixel and each Internal periphery pixel, and the fisrt feature of target is：
T1=[W, N1,…,NL]
In above-mentioned formula, T1Represent the fisrt feature of target, W represents the characteristic value of target outer contour, Nl(l=1,2 ...,
L the characteristic value of l-th of Internal periphery of target) is represented, L represents the number of the Internal periphery of target；
The characteristic value of the target outer contour obtains in the following manner：Binary conversion treatment is carried out to image, by outer contour
Grey scale pixel value is denoted as 1, remaining position grey scale pixel value is denoted as 0；Calculate the characteristic value W of target outer contour：
In above-mentioned formula, I (i, j) represents the gray value that pixel position is (i, j), and n and m represent target image respectively
It is wide and high, represent target outer contour number of pixels；
The characteristic value of the target outer contour obtains in the following manner：Binary conversion treatment is carried out to image, by target l
The grey scale pixel value of a Internal periphery is denoted as 1, remaining position grey scale pixel value is denoted as 0；Calculate the characteristic value of l-th of Internal periphery of target
Nl：
The second feature extracting sub-module is used for the second feature for extracting target, is specially：The outer edge of target is extracted,
The pixel of the outer contour inclusion region of target is obtained, binary conversion treatment is carried out to image, by the pixel ash of outer contour inclusion region
Angle value is denoted as 1, remaining position grey scale pixel value is denoted as 0, calculates the Second Eigenvalue of target：
In above-mentioned formula, T2Represent the second feature of target；
Described eigenvector generates submodule and generates clarification of objective vector according to fisrt feature and second feature, specifically
For：By the fisrt feature of target and second feature constitutive characteristic vector：T=[T1,T2], in above-mentioned formula, T represents the spy of target
Sign vector.
This preferred embodiment extracts target signature by characteristic extracting module, is established for succeeding target matching and identification
Basis is determined, by establishing clarification of objective vector, more accurate identification can have been carried out to target, specifically, fisrt feature
Taken into full account the outer contour and Internal periphery of target, second feature has taken into full account whole pixel in outer contour, target signature to
Amount combines fisrt feature and second feature, has obtained more complete target signature.
Preferably, the identification and evaluation module 5 includes the first evaluation submodule, the second evaluation submodule and overall merit
Module, the first evaluation submodule are used for the first evaluation of estimate for obtaining recognition performance, and the second evaluation submodule is used to obtain
The second evaluation of estimate of recognition performance is taken, the overall merit submodule is used for according to the first evaluation of estimate and the second evaluation of estimate to target
Recognition performance carries out overall merit；
The first evaluation submodule is used for the first evaluation of estimate for obtaining recognition performance, is specially：
In above-mentioned formula, S1Represent the first evaluation of estimate, A represents the target numbers included in image, A1Represent what can be identified
Target number, A2Represent the target number correctly identified；
The second evaluation submodule is used for the second evaluation of estimate for obtaining recognition performance, is specially：
In above-mentioned formula, S2Represent the second evaluation of estimate, B represents picture number to be identified, B1The first evaluation of estimate represented is big
In the picture number of given threshold；
The overall merit submodule is used to carry out target identification performance according to the first evaluation of estimate and the second evaluation of estimate comprehensive
Evaluation is closed, is specially：Calculate the overall merit factor：
In above-mentioned formula, S represents the overall merit factor；The overall merit factor is bigger, represents that target identification performance is better.
This preferred embodiment realizes the evaluation of picture recognition module recognition performance by identification and evaluation module, ensure that mesh
Mark is not horizontal, specifically, the first evaluation of estimate considers the identification accuracy of target, the second evaluation of estimate considers the identification of target
Stability, target identification calculate the overall merit factor by the first evaluation of estimate and the second evaluation of estimate, being capable of concentrated expression identification
Performance.
Target is identified using recognition performance of the present invention excellent robotic vision system, chooses 5 mesh to be identified
Mark is tested, and is respectively target 1 to be identified, target to be identified 2, target to be identified 3, target to be identified 4, target to be identified 5,
Recognition efficiency and identification accuracy are counted, compared compared with robotic vision system, generation has the beneficial effect that table
It is shown：
| Recognition efficiency improves | Identify that accuracy improves | |
| Target 1 to be identified | 29% | 27% |
| Target 2 to be identified | 27% | 26% |
| Target 3 to be identified | 26% | 26% |
| Target 4 to be identified | 25% | 24% |
| Target 5 to be identified | 24% | 22% |
Finally it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than the present invention is protected
The limitation of scope is protected, although being explained with reference to preferred embodiment to the present invention, those of ordinary skill in the art should
Work as understanding, can be to technical scheme technical scheme is modified or replaced equivalently, without departing from the reality of technical solution of the present invention
Matter and scope.
Claims (7)
1. the excellent robotic vision system of a kind of recognition performance, it is characterised in that including image capture module, image preprocessing
Module, characteristic extracting module, picture recognition module and identification and evaluation module, described image acquisition module are used to pass through camera pair
Image is acquired, and described image pretreatment module is used to pre-process the image of collection, and the characteristic extracting module is used
Extracted in pretreated characteristics of image, obtain target feature vector, described image identification module is according to target signature
Vector sum target template is matched, and target is identified, and the identification and evaluation module is used for the knowledge to picture recognition module
Other performance is evaluated.
2. the excellent robotic vision system of recognition performance according to claim 1, it is characterised in that described image is located in advance
Managing module includes first object segmentation submodule and the second pretreatment submodule, and the first object segmentation submodule uses
Canny operators extract the outer edge of target in image, and multiple Target Segmentations in image is described into simple target image
Second pretreatment submodule carries out greyscale transformation and filtering process to simple target image.
3. the excellent robotic vision system of recognition performance according to claim 2, it is characterised in that the feature extraction
Module includes fisrt feature extracting sub-module, second feature extracting sub-module and feature vector generation submodule, and described first is special
Sign extracting sub-module is used for the fisrt feature for extracting target, and the second feature extracting sub-module is used for the second spy for extracting target
Sign, described eigenvector generate submodule according to fisrt feature and second feature generation clarification of objective vector.
4. the excellent robotic vision system of recognition performance according to claim 3, it is characterised in that the fisrt feature
Extracting sub-module is used for the fisrt feature for extracting target, is specially：The outer edge of target is extracted, obtains target outer contour pixel
With the coordinate of each Internal periphery pixel, the fisrt feature of target is：
T1=[W, N1,…,NL]
In above-mentioned formula, T1Represent the fisrt feature of target, W represents the characteristic value of target outer contour, Nl(l=1,2 ..., L) table
Show the characteristic value of l-th of Internal periphery of target, L represents the number of the Internal periphery of target；
The characteristic value of the target outer contour obtains in the following manner：Binary conversion treatment is carried out to image, by outer contour pixel
Gray value is denoted as 1, remaining position grey scale pixel value is denoted as 0；Calculate the characteristic value W of target outer contour：
<mrow>
<mi>W</mi>
<mo>=</mo>
<msqrt>
<mrow>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mfrac>
<mrow>
<mi>i</mi>
<mo>&times;</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mfrac>
<mrow>
<mi>j</mi>
<mo>&times;</mo>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
</mrow>
In above-mentioned formula, I (i, j) represents the gray value that pixel position is (i, j), n and m represent respectively target image width and
Height, represents target outer contour number of pixels；
The characteristic value of the target outer contour obtains in the following manner：Binary conversion treatment is carried out to image, by l-th of target
The grey scale pixel value of profile is denoted as 1, remaining position grey scale pixel value is denoted as 0；Calculate the characteristic value N of l-th of Internal periphery of targetl：
<mrow>
<msub>
<mi>N</mi>
<mi>l</mi>
</msub>
<mo>=</mo>
<msqrt>
<mrow>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mfrac>
<mrow>
<mi>i</mi>
<mo>&times;</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mrow>
<mo>&lsqb;</mo>
<mfrac>
<mrow>
<mi>j</mi>
<mo>&times;</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<mo>-</mo>
<msubsup>
<mi>&Sigma;</mi>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&rsqb;</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
<mo>.</mo>
</mrow>
5. the excellent robotic vision system of recognition performance according to claim 4, it is characterised in that the second feature
Extracting sub-module is used for the second feature for extracting target, is specially：The outer edge of target is extracted, the outer contour for obtaining target includes
The pixel in region, carries out binary conversion treatment to image, the grey scale pixel value of outer contour inclusion region is denoted as 1, remaining position picture
Plain gray value is denoted as 0, calculates the Second Eigenvalue of target：
<mrow>
<msub>
<mi>T</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<msup>
<mi>e</mi>
<mrow>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</msubsup>
<msubsup>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</msubsup>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>+</mo>
<mn>2</mn>
<mi>&pi;</mi>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>m</mi>
</munderover>
<mi>I</mi>
<mrow>
<mo>(</mo>
<mi>i</mi>
<mo>,</mo>
<mi>j</mi>
<mo>)</mo>
</mrow>
</mrow>
In above-mentioned formula, T2Represent the second feature of target；
Described eigenvector generates submodule according to fisrt feature and second feature generation clarification of objective vector, is specially：Will
Fisrt feature and second feature the constitutive characteristic vector of target：T=[T1,T2], in above-mentioned formula, T represent clarification of objective to
Amount.
6. the excellent robotic vision system of recognition performance according to claim 5, it is characterised in that the identification and evaluation
Module includes the first evaluation submodule, the second evaluation submodule and overall merit submodule, and the first evaluation submodule is used for
The first evaluation of estimate of recognition performance is obtained, the second evaluation submodule is used for the second evaluation of estimate for obtaining recognition performance, described
Overall merit submodule is used to carry out overall merit to target identification performance according to the first evaluation of estimate and the second evaluation of estimate.
7. the excellent robotic vision system of recognition performance according to claim 6, it is characterised in that first evaluation
Submodule is used for the first evaluation of estimate for obtaining recognition performance, is specially：
<mrow>
<msub>
<mi>S</mi>
<mn>1</mn>
</msub>
<mo>=</mo>
<mfrac>
<mrow>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mo>+</mo>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
</mrow>
<mrow>
<mi>A</mi>
<mo>+</mo>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
</mrow>
</mfrac>
<mo>&times;</mo>
<mi>lg</mi>
<mrow>
<mo>(</mo>
<mfrac>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
<mi>A</mi>
</mfrac>
<mo>+</mo>
<mfrac>
<msub>
<mi>A</mi>
<mn>2</mn>
</msub>
<msub>
<mi>A</mi>
<mn>1</mn>
</msub>
</mfrac>
<mo>+</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
In above-mentioned formula, S1Represent the first evaluation of estimate, A represents the target numbers included in image, A1Represent the target that can be identified
Number, A2Represent the target number correctly identified；
The second evaluation submodule is used for the second evaluation of estimate for obtaining recognition performance, is specially：
<mrow>
<msub>
<mi>S</mi>
<mn>2</mn>
</msub>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mfrac>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
<mi>B</mi>
</mfrac>
<mo>+</mo>
<mfrac>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
<mrow>
<mi>B</mi>
<mo>+</mo>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
<mo>&times;</mo>
<msqrt>
<mrow>
<mfrac>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
<mi>B</mi>
</mfrac>
<mo>+</mo>
<mfrac>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
<mrow>
<mi>B</mi>
<mo>+</mo>
<msub>
<mi>B</mi>
<mn>1</mn>
</msub>
</mrow>
</mfrac>
</mrow>
</msqrt>
</mrow>
In above-mentioned formula, S2Represent the second evaluation of estimate, B represents picture number to be identified, B1The first evaluation of estimate represented, which is more than, to be set
Determine the picture number of threshold value；
The overall merit submodule is used to target identification performance integrate commenting according to the first evaluation of estimate and the second evaluation of estimate
Valency, is specially：Calculate the overall merit factor：
<mrow>
<mi>S</mi>
<mo>=</mo>
<mn>1</mn>
<mo>+</mo>
<msqrt>
<mrow>
<msup>
<mi>e</mi>
<mrow>
<msub>
<mi>S</mi>
<mn>1</mn>
</msub>
<mo>&times;</mo>
<msub>
<mi>S</mi>
<mn>2</mn>
</msub>
</mrow>
</msup>
<mo>+</mo>
<msub>
<mi>S</mi>
<mn>1</mn>
</msub>
<mo>&times;</mo>
<msub>
<mi>S</mi>
<mn>2</mn>
</msub>
</mrow>
</msqrt>
</mrow>
In above-mentioned formula, S represents the overall merit factor；The overall merit factor is bigger, represents that target identification performance is better.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711289370.1A CN108038849A (en) | 2017-12-07 | 2017-12-07 | A kind of excellent robotic vision system of recognition performance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711289370.1A CN108038849A (en) | 2017-12-07 | 2017-12-07 | A kind of excellent robotic vision system of recognition performance |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108038849A true CN108038849A (en) | 2018-05-15 |
Family
ID=62096025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711289370.1A Withdrawn CN108038849A (en) | 2017-12-07 | 2017-12-07 | A kind of excellent robotic vision system of recognition performance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108038849A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108839032A (en) * | 2018-06-27 | 2018-11-20 | 深圳大图科创技术开发有限公司 | A kind of intelligent robot |
| CN111609852A (en) * | 2019-02-25 | 2020-09-01 | 北京奇虎科技有限公司 | Semantic map construction method, sweeping robot and electronic equipment |
| CN111679661A (en) * | 2019-02-25 | 2020-09-18 | 北京奇虎科技有限公司 | Semantic map construction method based on depth camera and sweeping robot |
-
2017
- 2017-12-07 CN CN201711289370.1A patent/CN108038849A/en not_active Withdrawn
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108839032A (en) * | 2018-06-27 | 2018-11-20 | 深圳大图科创技术开发有限公司 | A kind of intelligent robot |
| CN111609852A (en) * | 2019-02-25 | 2020-09-01 | 北京奇虎科技有限公司 | Semantic map construction method, sweeping robot and electronic equipment |
| CN111679661A (en) * | 2019-02-25 | 2020-09-18 | 北京奇虎科技有限公司 | Semantic map construction method based on depth camera and sweeping robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110097093B (en) | Method for accurately matching heterogeneous images | |
| CN106709950B (en) | Binocular vision-based inspection robot obstacle crossing wire positioning method | |
| CN108335331B (en) | Binocular vision positioning method and equipment for steel coil | |
| CN104268519B (en) | Image recognition terminal and its recognition methods based on pattern match | |
| CN103198477B (en) | Apple fruitlet bagging robot visual positioning method | |
| CN102722731A (en) | Efficient image matching method based on improved scale invariant feature transform (SIFT) algorithm | |
| CN107464252A (en) | A kind of visible ray based on composite character and infrared heterologous image-recognizing method | |
| CN106408597A (en) | Neighborhood entropy and consistency detection-based SAR (synthetic aperture radar) image registration method | |
| CN110782477A (en) | Moving target rapid detection method based on sequence image and computer vision system | |
| CN103136525B (en) | High-precision positioning method for special-shaped extended target by utilizing generalized Hough transformation | |
| CN104268602A (en) | Shielded workpiece identifying method and device based on binary system feature matching | |
| CN104715251B (en) | A kind of well-marked target detection method based on histogram linear fit | |
| CN108257155B (en) | Extended target stable tracking point extraction method based on local and global coupling | |
| CN108447016B (en) | Optical image and SAR image matching method based on straight line intersection point | |
| CN112634335A (en) | Method for extracting characteristic point pairs of robust remote sensing image facing to nonlinear radiation distortion | |
| CN106023187A (en) | Image registration method based on SIFT feature and angle relative distance | |
| Han et al. | An improved corner detection algorithm based on harris | |
| CN104123554A (en) | SIFT image characteristic extraction method based on MMTD | |
| CN108038849A (en) | A kind of excellent robotic vision system of recognition performance | |
| CN105405138A (en) | Water surface target tracking method based on saliency detection | |
| CN110335280A (en) | A kind of financial documents image segmentation and antidote based on mobile terminal | |
| CN104077769A (en) | Error matching point pair removing algorithm in image registration | |
| CN115239882A (en) | Crop three-dimensional reconstruction method based on low-light image enhancement | |
| CN104715476A (en) | Salient object detection method based on histogram power function fitting | |
| CN106355576A (en) | SAR image registration method based on MRF image segmentation algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180515 |