Method and system for AR-based graded management education of laboratory safety risk
A method and system for AR-based graded management education of laboratory safety risk are provided, which includes following steps: collecting laboratory image data and sensor data and obtaining a risk grade matching rule of a laboratory; identifying image feature data of the image data and sensor feature data of the sensor data, converting matched image feature data into a laboratory equipment name and converting matched sensor feature data into an equipment model; inputting the equipment name and a corresponding equipment model and quantity into the risk grade matching rule of the laboratory, and outputting a risk grade of a current laboratory according to the risk grade matching rule of the laboratory; and fusing and matching a AR template image corresponding to the identified and output risk grade and real image information according to preset AR template images for various risk grades, so as to output alarm information.
1. A method for AR-based graded management education of laboratory safety risk, comprising:
collecting a laboratory image data and a sensor data and obtaining a risk grade matching rule of a laboratory;
identifying an image feature data of the laboratory image data and a sensor feature data of the sensor data, converting the image feature data into a laboratory equipment name after matching and converting the sensor feature data into an equipment model after matching;
inputting the equipment name and a corresponding equipment model and quantity into the risk grade matching rule of the laboratory, and outputting a risk grade of a current laboratory according to the risk grade matching rule of the laboratory; and
according to a present plurality of AR template images for various risk grades, fusing and matching the AR template image corresponding to the identified image feature data and output risk grade to output an alarm information;
wherein identifying the image feature data comprises:
applying a gray threshold T to the laboratory image data using a gray threshold segmentation method, wherein f(x,y) represents a pixel value of a point (x,y) and pixel values of f(x,y) greater than the threshold T are segmented to obtain the segmented image g(x,y); and
by using a Hough transform, selecting any point (a,b) within a shape of the segmented image g(x,y) as a reference point, and
calculating a tangent direction ϕ and an offset vector r to a position of the reference point (a,b) and an included angle α between r and an x axis, from each point on an edge of a graph of an arbitrary shape,
wherein calculating the reference point (a,b) is as follows:
a=x+r (ϕ)cos(α(ϕ))
b=x+r (ϕ)sin(α(ϕ)).
2. The method for AR-based graded management education of laboratory safety risk according to claim 1 , wherein after the feature extraction is completed, a classifier is further configured to classify extracted features, a classification method comprises:
calculating a total number M of classes of current equipment samples, a current sample class extracted from the sample library being wi={1, 2, . . . , M}, a prior probability of the current sample being expressed as P(wi) and a class conditional probability density of the current sample being expressed as P(x|wi); calculating a posterior probability P(wi|x) of the current sample by using a Bayesian formula, a class corresponding to a maximum value of the posterior probability being a final classified class, j representing a subscript of a summation term, n being the total number of classes of the equipment, where the posterior probability P(wi|x) is calculated as follows:
P
(
w
i
❘
x
)
=
P
(
x
❘
w
i
)
P
(
w
i
)
∑
j
=
0
n
P
(
x
❘
w
j
)
P
(
w
j
)
.
3. The method for AR-based graded management education of laboratory safety risk according to claim 1 , wherein the risk grade matching rule of the laboratory comprises: establishing equipment classes and quantities for various risk grades, calculating equipment classes corresponding to all of the image feature data in a laboratory image identified by a classifier; and if an equipment class identified by the classifier completely cover an equipment class and quantity for a corresponding risk grade, determining that a current laboratory scene is a laboratory matched with covered risk grade rules, and outputting the AR template image corresponding to the risk grade of the laboratory so as to be fused with a real laboratory image.
4. The method for AR-based graded management education of laboratory safety risk according to claim 1 , wherein the risk grade matching rule of the laboratory comprises a decision rule for various priorities, and the decision rule for various priorities comprises: establishing a first priority risk grade rule and a second priority risk grade rule; and if an equipment class and quantity identified in the collected laboratory image data not only cover an equipment class and quantity corresponding to the first priority risk grade rule, but also cover an equipment class and quantity corresponding to the second priority risk grade rule, the first priority risk grade being higher than the second priority risk grade, directly calling an AR template image corresponding to the first priority risk grade to be fused with a real laboratory image.
5. The method for AR-based graded management education of laboratory safety risk according to claim 1 , wherein the risk grade matching rule of the laboratory comprises: collecting the sensor data, setting a risk grade range of sensor values, and if a collected current sensor value is at a corresponding risk grade, and an equipment class and quantity in the image obtained by a classifier match a same risk grade in the grade range where the sensor data is located, directly outputting a corresponding AR module image with a same risk grade to be fused with a real laboratory image.
6. The method for AR-based graded management education of laboratory safety risk according to claim 1 , wherein the risk grade matching rule of the laboratory comprises: collecting the sensor data, setting a risk grade range of sensor values, and if there is difference between a corresponding risk grade where the collected current sensor value is located and a risk grade determined by an equipment class and quantity in the image obtained by a classifier, selecting a risk grade with higher priority to output an AR module image corresponding to the risk grade to be fused with a real laboratory image.
7. A system for AR-based graded management education of laboratory safety risk, wherein the system implements the method for AR-based graded management education of laboratory safety risk according to claim 1 .
8. A non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores a computer program, and the computer program can be executed by a processor to implement the method for AR-based graded management education of laboratory safety risk according to claim 1 .