CN109446950A - A kind of static gesture identification method based on thermal imaging system camera - Google Patents
A kind of static gesture identification method based on thermal imaging system camera Download PDFInfo
- Publication number
- CN109446950A CN109446950A CN201811200740.4A CN201811200740A CN109446950A CN 109446950 A CN109446950 A CN 109446950A CN 201811200740 A CN201811200740 A CN 201811200740A CN 109446950 A CN109446950 A CN 109446950A
- Authority
- CN
- China
- Prior art keywords
- gesture
- imaging system
- thermal imaging
- image
- system camera
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
- G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of static gesture identification methods based on thermal imaging system camera, comprising the following steps: the thermal imagery temperature data for obtaining thermal imaging system camera becomes image pixel value by Logarithm conversion；Image preprocessing；Feature extraction；Gesture identification.Due to converting after image pixel value row processing and identification again for thermal imagery temperature data, static gesture identification method provided by the invention based on thermal imaging system camera can effectively and accurately be partitioned into manpower region under the various environment of unglazed, dim light and complex colors background etc., and carry out accurate gesture classification identification, correct expected result is exported, and algorithm robustness is good.
Description
Technical field
The invention belongs to image procossing technical field of hand gesture recognition, are related to a kind of static gesture based on thermal imaging system camera
Recognition methods.
Background technique
In current social, the man-machine interaction mode based on keyboard and mouse still accounts for leading position in equipment, these
Mode requires to provide additional equipment for interactive device, is not suitable for the robot applied with locomotivity or portable
On smart machine.And based on the human-computer interaction of gesture identification based on directly operating, make human-computer interaction technology from machine
The heart be gradually transferred to it is human-centred, more meet people exchange habit.Therefore, gesture identification is just gradually by development and application in work
Cheng Zhong.
But traditional common camera can only carry out gesture identification under bright and clear environment, once insufficient light
Or no light conditions, then gesture identification resultant error is larger or even can not identify, this results in the application environment of gesture identification to have
There is limitation.Meanwhile common RGB camera is in complex environment, and such as: in the case of various colors or similar color, Hand Gesture Segmentation
Effect is very poor, this will affect final gesture identification result and inaccurate, and error is larger.
Summary of the invention
To solve the above problems, the invention discloses a kind of static gesture identification method based on thermal imaging system camera, energy
It is enough to realize convenient, accurate and efficient gesture identification under circumstances.Based on the method for the present invention, without external when gesture identification
Boundary's environment has an extra limitation, the power of ambient light will not influence gesture identification as a result, other reflection lights or phase
Nearly color equally will not influence the segmentation result of gesture, and solving common RGB camera can not be in the general scene in life
Under carry out using the problem of.
In order to achieve the above object, the invention provides the following technical scheme:
A kind of static gesture identification method based on thermal imaging system camera, includes the following steps:
Step 1, data are obtained
The thermal imagery temperature data that thermal imaging system camera is obtained becomes image pixel value by Logarithm conversion, i.e., by temperature
Data are non-linear to correspond to each image pixel value, so that data is converted to temperature pattern and shows, and has widened palm temperature simultaneously
Neighbouring corresponding image pixel value；
Step 2, image preprocessing
Edge is detected using canny edge detection algorithm, the edge of acquisition is filled using unrestrained water completion method, finally
Hand Gesture Segmentation, which is carried out, using image binaryzation obtains hand shape；
Step 3, feature extraction
Extract by the following feature in step 2 treated image: center of gravity square, convex closure convex defect, refer to position finger tip number
It sets；
Step 4, gesture identification
Using center of gravity square, finger tip number and refer to that being combined matching with the big feature in position three screens gesture, carries out gesture
Classification and Identification.
Further, during the Logarithm conversion, logarithm chooses the inverse function of sigmoid function
Its switch process is as follows:
Initial setting image displays temperature section, the gray value higher than 37 DEG C is 255, and the gray value lower than 13 DEG C is then
It is set as 0；
The highest temperature maxtemp of temperature range is enabled to correspond to sigmoid function abscissa x=0.7, mintemp pairs of lowest temperature
X=0.9 is answered, calculating acquires linear formula are as follows:
Sigmoid_x=0.0004*tempreture+1.58 (2)
It is one section neighbouring up and down to be rounded an image thermal imagery statistical average, calculates corresponding sigmoid_x's according to formula (2)
Value, then by formula (1), calculate the maximum value sigmoid_y_max and minimum value sigmoid_y_min of this section；
Whole image thermal-image data first acquires sigmoid_x according to formula (2), substitutes into formula (1), acquires sigmoid_
Y, finally according to formula (3)
It calculates gray value and shows image.
Further, center of gravity moment characteristics pass through formula (4) in the step 3
It acquires.
Further, step 3 middle finger is obtained with position by following procedure: being taken to each finger of gesture maximum external
Quadrangle obtains four vertex position informations of quadrangle, chooses the midpoint that two o'clock line is acquired on underlying two vertex, i.e.,
Refer to for correspondence with position.
Further, profile first searched to image binary map before the convex defects detection of convex closure in the step 3, then to obtaining
Profile carry out polygonal segments.
Further, the step 4 includes the following steps:
Step 4-1 just classifies gesture according to finger tip number；
Step 4-2 forms corner dimension and refers to and classified again with range information according to finger.
Further, classification includes: finger tip number for 0 gesture classification in the step 4-1, the hand that finger tip number is 1
Gesture classification, the gesture classification that finger tip number is 2, the gesture classification that finger tip number is 3, the gesture classification that finger tip number is 4.
Further, subseries, the classification of differentiation include: the step 4-2 again under the classification that step 4-1 is obtained
When finger tip number is 1, gesture of the angle less than 80 degree and angle are greater than 80 degree of gestures；
When finger tip number is 2, two Fingers are with the gesture respectively with angle formed by center of gravity line at 0 °~50 °, two hands
Refer to angle 50 °~80 ° and two refer to distance be less than thumb length 1.5 times of gesture, two finger angles 80 °~
180 ° and two refer to distance be greater than thumb length 1.5 times of gesture；
When finger tip number is 3, middle finger the gesture of the small Mr. Yu's distance of range difference and is greater than at a distance from other two fingers
The gesture of equal Mr. Yus' distance.
Compared with prior art, the invention has the advantages that and the utility model has the advantages that
It is provided by the invention based on heat due to converting after image pixel value row processing and identification again for thermal imagery temperature data
As the static gesture identification method of instrument camera can have under the various environment of unglazed, dim light and complex colors background etc.
Effect is accurately partitioned into manpower region, and carries out accurate gesture classification identification, exports correct expected result, and algorithm robust
Property is good.
Detailed description of the invention
Fig. 1 is 10 kinds of gesture classification figures that the embodiment of the present invention proposes.
Fig. 2 is the static gesture identification method flow chart provided by the invention based on thermal imaging system camera.
Fig. 3 is that thermal-image data Logarithm conversion is at temperature pattern display renderings in step 1 of the present invention, wherein left side is conversion
Before, right side is after converting.
Fig. 4 is canny edge detection effect picture in step 2 of the present invention.
Fig. 5 is that water completion method filling effect figure is overflow in step 2 of the present invention.
Fig. 6 is the location drawing of the finger tip that step 3 of the present invention is extracted and center of gravity square.
Fig. 7 is the result effect picture for using the method for the present invention to identify by taking gesture 6 as an example.
Specific embodiment
Technical solution provided by the invention is described in detail below with reference to specific embodiment, it should be understood that following specific
Embodiment is only illustrative of the invention and is not intended to limit the scope of the invention.
The present embodiment realizes the Classification and Identification of 10 kinds of gestures, and gesture classification is as shown in Figure 1.It is said below based on the present embodiment
The bright static gesture identification method provided by the invention based on thermal imaging system camera, specific steps are as shown in Fig. 2, include as follows
Step:
Step 1, data are obtained
The thermal imagery temperature data that thermal imaging system camera is obtained becomes image pixel value by Logarithm conversion, i.e., by temperature
Data are non-linear to correspond to each image pixel value, so that data is converted to temperature pattern and shows, and has widened palm temperature simultaneously
Neighbouring corresponding image pixel value, in order to subsequent image procossing.
Wherein Logarithm conversion specifically: the inverse function of logarithm selection sigmoid function
Steps are as follows:
Initial setting image displays temperature section (13 DEG C~37 DEG C), the gray value higher than 37 DEG C is 255, and is lower than 13
DEG C gray value be then set as 0；
The highest temperature maxtemp of temperature range is enabled to correspond to sigmoid function abscissa x=0.7, mintemp pairs of lowest temperature
X=0.9 (linear corresponding) is answered, calculating acquires linear formula are as follows:
Sigmoid_x=0.0004*tempreture+1.58 (2)
It is rounded an image thermal imagery statistical average (close to palm temperature, i.e. 29 degrees centigrades) (50) nearby one up and down
Section calculates the value of corresponding sigmoid_x according to formula (2), then by formula (1), calculates the maximum value of this section
Sigmoid_y_max and minimum value sigmoid_y_min；
Whole image thermal-image data first acquires sigmoid_x according to formula (2), substitutes into formula (1), acquires sigmoid_
Y, finally according to formula (3)
It calculates gray value and shows image.Before conversion and transformation result figure is as shown in Figure 3.
Step 2, image preprocessing
Edge is detected using canny edge detection algorithm, effect picture is as shown in figure 4, using unrestrained water completion method to acquisition
Edge is filled, and filling result is as shown in figure 5, last carry out Hand Gesture Segmentation acquisition hand shape using image binaryzation.
Step 3, feature extraction
Extract by the following feature in step 2 treated image: center of gravity square, convex closure convex defect, refer to position finger tip number
It sets.Center of gravity square passes through formula (4)
It acquires.Profile is searched to image binary map, polygonal segments are carried out to obtained profile, are utilized on this basis
Opencv function convexHull and convexityDefects carry out the convex defects detection of convex closure.It is obtained according to the convex defect analysis of convex closure
The position of finger tip and quantity information out.Finger tip and center of gravity square are drawn in Fig. 6.Refer to position acquisition process: each to gesture
Finger takes maximum external quadrangle, obtains four vertex position informations of quadrangle, chooses underlying two vertex and acquires two
The midpoint of point line, as correspondence refer to position.
Step 4, gesture identification
Using center of gravity square, finger tip number and refer to that being combined matching with the big feature in position three screens gesture, to realize to 10
Kind gesture carries out Classification and Identification.
The Classification and Identification of gesture is broadly divided into two steps: first is that just being classified according to finger tip number, pressing from both sides second is that being formed according to finger
Angle size and refer to classifying again with range information.Detailed process are as follows: ten kinds of gestures can be divided by six classes according to finger tip number,
Be divided into: finger tip number be 0 gesture 7, finger tip number be 1 gesture 1 and gesture 10, finger tip number be 2 gesture 2,6 and of gesture
Gesture 8, the gesture 3 and gesture 9 that finger tip number is 3, the gesture 5 that the gesture 4 and finger tip number that finger tip number is 4 are 5.Gesture
1 and 10 differentiation foundation refers to that angle is gesture 10 less than 80 degree, and angle is greater than with the line and horizontal angle with center of gravity
80 degree are gesture 1.Finger tip number be 2 when, two Fingers with respectively with angle formed by center of gravity line 0 °~50 ° be gesture
2, two finger angles 50 °~80 ° and two refer to distance to be less than 1.5 times of thumb length be gesture 8, two finger angles
80 °~180 ° and two refer to distance to be greater than 1.5 times of thumb length be gesture 6.When finger tip number is 3, in calculating
Refer at a distance from other two fingers, if two range differences are less than lesser distance (distance value should provide in advance), for hand
Otherwise gesture 3 is gesture 9.For example the recognition result of gesture 6 is as shown in Figure 7.
The technical means disclosed in the embodiments of the present invention is not limited only to technological means disclosed in above embodiment, further includes
Technical solution consisting of any combination of the above technical features.It should be pointed out that for those skilled in the art
For, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also considered as
Protection scope of the present invention.
Claims (8)
1. a kind of static gesture identification method based on thermal imaging system camera, which comprises the steps of:
Step 1, data are obtained
The thermal imagery temperature data that thermal imaging system camera is obtained becomes image pixel value by Logarithm conversion, i.e., by temperature data
It is non-linear to correspond to each image pixel value, so that data is converted to temperature pattern and show, and has been widened near palm temperature simultaneously
Corresponding image pixel value；
Step 2, image preprocessing
Edge is detected using canny edge detection algorithm, the edge of acquisition is filled using unrestrained water completion method, is finally utilized
Image binaryzation carries out Hand Gesture Segmentation and obtains hand shape；
Step 3, feature extraction
Extract by the following feature in step 2 treated image: center of gravity square, convex closure convex defect, refer to position finger tip number；
Step 4, gesture identification
Using center of gravity square, finger tip number and refer to that being combined matching with the big feature in position three screens gesture, classifies to gesture
Identification.
2. the static gesture identification method according to claim 1 based on thermal imaging system camera, which is characterized in that described right
In number conversion process, logarithm chooses the inverse function of sigmoid function
Its switch process is as follows:
Initial setting image displays temperature section, the gray value higher than 37 DEG C is 255, and the gray value lower than 13 DEG C is then set as
0；
It enables the highest temperature maxtemp of temperature range correspond to sigmoid function abscissa x=0.7, lowest temperature mintemp and corresponds to x=
0.9, calculating acquires linear formula are as follows:
Sigmoid_x=0.0004*tempreture+1.58 (2)
It is one section neighbouring up and down to be rounded an image thermal imagery statistical average, the value of corresponding sigmoid_x is calculated according to formula (2),
Again by formula (1), the maximum value sigmoid_y_max and minimum value sigmoid_y_min of this section are calculated；
Whole image thermal-image data first acquires sigmoid_x according to formula (2), substitutes into formula (1), acquires sigmoid_y, most
Afterwards according to formula (3)
It calculates gray value and shows image.
3. the static gesture identification method according to claim 1 based on thermal imaging system camera, which is characterized in that the step
Center of gravity moment characteristics pass through formula (4) in rapid 3
It acquires.
4. the static gesture identification method according to claim 1 based on thermal imaging system camera, which is characterized in that the step
Rapid 3 middle finger is obtained with position by following procedure: being taken maximum external quadrangle to each finger of gesture, is obtained four, quadrangle tops
Dot position information chooses the midpoint that two o'clock line is acquired on underlying two vertex, and as correspondence refers to position.
5. the static gesture identification method according to claim 1 based on thermal imaging system camera, which is characterized in that the step
Profile first is searched to image binary map before the convex defects detection of convex closure in rapid 3, then polygonal segments are carried out to obtained profile.
6. the static gesture identification method according to claim 1 based on thermal imaging system camera, which is characterized in that the step
Rapid 4 include the following steps:
Step 4-1 just classifies gesture according to finger tip number；
Step 4-2 forms corner dimension and refers to and classified again with range information according to finger.
7. the static gesture identification method according to claim 6 based on thermal imaging system camera, which is characterized in that the step
Classification includes: the gesture classification that finger tip number is 0, the gesture classification that finger tip number is 1, the gesture that finger tip number is 2 in rapid 4-1
Classification, the gesture classification that finger tip number is 3, the gesture classification that finger tip number is 4.
8. the static gesture identification method according to claim 6 based on thermal imaging system camera, which is characterized in that the step
Subseries, the classification of differentiation include: rapid 4-2 again under the classification that step 4-1 is obtained
When finger tip number is 1, gesture of the angle less than 80 degree and angle are greater than 80 degree of gestures；
When finger tip number is 2, two Fingers are with the gesture respectively with angle formed by center of gravity line at 0 °~50 °, two finger clamps
Angle 50 °~80 ° and two refer to distance be less than thumb length 1.5 times of gesture, two finger angles at 80 °~180 ° and
Two refer to distance be greater than thumb length 1.5 times of gesture；
When finger tip number is 3, middle finger the gesture of the small Mr. Yu's distance of range difference and is more than or equal at a distance from other two fingers
The gesture of certain distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811200740.4A CN109446950B (en) | 2018-10-16 | 2018-10-16 | Static gesture recognition method based on thermal imager camera |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811200740.4A CN109446950B (en) | 2018-10-16 | 2018-10-16 | Static gesture recognition method based on thermal imager camera |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109446950A true CN109446950A (en) | 2019-03-08 |
| CN109446950B CN109446950B (en) | 2022-02-15 |
Family
ID=65544948
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811200740.4A Active CN109446950B (en) | 2018-10-16 | 2018-10-16 | Static gesture recognition method based on thermal imager camera |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109446950B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111626364A (en) * | 2020-05-28 | 2020-09-04 | 中国联合网络通信集团有限公司 | Gesture image classification method and device, computer equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751141A (en) * | 2015-03-30 | 2015-07-01 | 东南大学 | ELM gesture recognition algorithm based on feature image full pixel gray values |
| CN106814859A (en) * | 2017-03-20 | 2017-06-09 | 肖赫 | A kind of man-machine interaction method of infrared gesture identification |
| CN107220584A (en) * | 2017-03-25 | 2017-09-29 | 南宁市广千信息技术有限公司 | Infrared gesture recognition system |
-
2018
- 2018-10-16 CN CN201811200740.4A patent/CN109446950B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104751141A (en) * | 2015-03-30 | 2015-07-01 | 东南大学 | ELM gesture recognition algorithm based on feature image full pixel gray values |
| CN106814859A (en) * | 2017-03-20 | 2017-06-09 | 肖赫 | A kind of man-machine interaction method of infrared gesture identification |
| CN107220584A (en) * | 2017-03-25 | 2017-09-29 | 南宁市广千信息技术有限公司 | Infrared gesture recognition system |
Non-Patent Citations (1)
| Title |
|---|
| KIM S等: "Tracking and classification of in-air hand gesture based on thermal", 《SENSORS》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111626364A (en) * | 2020-05-28 | 2020-09-04 | 中国联合网络通信集团有限公司 | Gesture image classification method and device, computer equipment and storage medium |
| CN111626364B (en) * | 2020-05-28 | 2023-09-01 | 中国联合网络通信集团有限公司 | Gesture image classification method, gesture image classification device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109446950B (en) | 2022-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107038424B (en) | Gesture recognition method | |
| CN106096610B (en) | A kind of file and picture binary coding method based on support vector machines | |
| CN112949564A (en) | Pointer type instrument automatic reading method based on deep learning | |
| CN110232311A (en) | Dividing method, device and the computer equipment of hand images | |
| CN103679145A (en) | Automatic gesture recognition method | |
| CN112634125B (en) | Automatic face replacement method based on off-line face database | |
| CN109426814B (en) | Method, system and equipment for positioning and identifying specific plate of invoice picture | |
| CN103971102A (en) | Static gesture recognition method based on finger contour and decision-making trees | |
| CN109637664A (en) | A kind of BMI evaluating method, device and computer readable storage medium | |
| CN101807257A (en) | Method for identifying information of image tag | |
| US20190236345A1 (en) | Hand detection method and system, image detection method and system, hand segmentation method, storage medium, and device | |
| CN109687382B (en) | Relay protection pressing plate switching state identification method based on color template matching | |
| CN109961016B (en) | Multi-gesture accurate segmentation method for smart home scene | |
| CN112232332A (en) | Non-contact palm detection method based on video sequence | |
| Vishwakarma et al. | Simple and intelligent system to recognize the expression of speech-disabled person | |
| Anis et al. | Digital electric meter reading recognition based on horizontal and vertical binary pattern | |
| CN110659637A (en) | Electric energy meter number and label automatic identification method combining deep neural network and SIFT features | |
| CN109724988A (en) | A kind of pcb board defect positioning method based on multi-template matching | |
| CN111539330A (en) | Transformer substation digital display instrument identification method based on double-SVM multi-classifier | |
| CN115082776A (en) | Electric energy meter automatic detection system and method based on image recognition | |
| CN109271882B (en) | Method for extracting color-distinguished handwritten Chinese characters | |
| CN113705564B (en) | Pointer type instrument identification reading method | |
| CN110263784A (en) | The English paper achievement of intelligence identifies input method | |
| CN109446950A (en) | A kind of static gesture identification method based on thermal imaging system camera | |
| CN109934152B (en) | Improved small-bent-arm image segmentation method for sign language image |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |