KR101475849B1 - PC Automatic Energy Saving System by Using Thermal Image Sensor and Neural Network Model and Method thereof - Google Patents

Abstract

The present invention relates to a personal computer (PC) automatic energy saving system using a thermal image sensor and a neural network model and a method thereof. The PC automatic energy saving system includes: a thermal imaging module to sense temperature data of an object positioned at a predetermined distance from a front surface of the PC and to transmit the temperature data to the PC; the PC to extract a foreground based on thermal image information received from the thermal imaging module, and to analyze thermal image data using the neural network model to determine the existence of a person so that a body and a monitor are turned on or turned off; and a server to monitor on-off states of multiple PCs located at an office.

Description

Technical Field [0001] The present invention relates to a PC automatic energy saving system using a thermal image sensor and a neural network model, and an automatic power saving method using the thermal image sensor and a neural network model,

The present invention relates to saving energy consumption of a PC using a thermal image sensor and a neural network model. Usually, a home has a small number of PCs, but in the case of an office, it has about one PC, and the energy used by the PC is a major factor in office energy consumption. Therefore, the present invention can reduce energy consumption in offices using a large number of PCs, thereby reducing costs and carbon emissions, thereby having a good effect on environmental problems.

The prior art related to the present invention is disclosed in Korean Patent Laid-Open Publication No. 10-2011-0049043 (published on May 12, 2011). FIG. 1 is a block diagram of a conventional monitor power control system for detecting infrared rays of a human body. 1, a conventional monitor power control system for detecting infrared rays of a human body is constructed by attaching a super-infrared sensor 2 to the lower front of an LCD or LED monitor 1. When a person 3 is positioned in front of a monitor The super-infrared sensor 2 senses the infrared rays of the human body 3 and controls the monitor backlight inverter 5. [ In the circuit structure of the monitor, the super-infrared sensor 2 senses the human infrared light in front of the monitor 1 and transmits it to the control circuit 6. The control circuit receives the signal from the super-infrared sensor 2 The backlight inverter or the LED converter 5 is turned on and the backlight inverter or the LED converter 5 is turned off if the signal is a signal that the human body infrared ray is not detected. Depending on the ON or OFF state of the backlight inverter or the LED converter 5, the backlight of the monitor or the LED 4 is turned on or off. In this way, you can turn off the LCD or LED monitor automatically when it is turned on and off when there is a person, thereby saving power on the monitor and extending the life of the monitor.

The conventional monitor power control system using the infrared sensor has a problem that it is difficult to accurately determine whether the object to be detected by the infrared sensor is a human being. In addition, the conventional monitor power control system using the infrared sensor as described above can not reduce the power consumption of the PC main body by controlling the power of each monitor, and also judges whether a person is re- There is a problem that can not be controlled. Therefore, an object of the present invention is to accurately determine whether a person is re-installed through a program by using thermal image information received from an infrared sensor and to save energy consumption by turning off the PC if there is no person. In addition, since each PC is connected to the network in the office, it monitors whether or not the PC is turned off so that the entire office can be checked again.

The PC automatic power saving system and the automatic power saving method using the thermal image sensor and the neural network model having the above-mentioned object have a function of sensing the temperature data (thermal image information) An image module, a PC for extracting a foreground based on thermal image information received from the thermal image module, analyzing thermal image data using a neural network model to determine whether a person exists, and controlling the main body and the monitor to be turned on / off And a server for monitoring on-off status of a plurality of PCs located in an office.

The PC automatic power saving system and the automatic power saving method using the thermal image sensor and the neural network model constructed as described above have the effect of reducing the power consumption of the main body and the monitor by applying the power saving function to the PC using the thermal image sensor . Also, by monitoring the status of each PC, it is possible to judge whether or not a person is re-installed. In addition, the neural network model can be used to accurately determine the presence of a person using a PC.

FIG. 1 is a block diagram of a conventional monitor power control system for detecting infrared rays of a human body,
FIG. 2 is a general configuration diagram of a PC automatic power saving system using a thermal image sensor and a neural network model of the present invention,
3 is a control flow diagram of a method for automatic power saving of a PC using a thermal image sensor and a neural network model according to the present invention,
4A is a flowchart of a foreground extraction method applied to the present invention,
FIG. 4B is a control flowchart for a method of separating foreground and background applied to the present invention; FIG.
5 is a block diagram of a neural network model applied to the present invention,
6 is a control flowchart of a method for determining whether a person exists using a neural network model applied to the present invention.

The PC automatic power saving system and the automatic power saving method using the thermal image sensor and the neural network model having the above-described objects will be described with reference to FIGS. 2 to 6. FIG.

2 is a diagram showing the entire configuration of a PC automatic power saving system using a thermal image sensor and a neural network model according to the present invention. 2, the PC automatic power saving system using the thermal image sensor and the neural network model according to the present invention is for transmitting thermal image information of an object positioned on the front side of a PC, and includes a USB module and a thermal imaging module 100 A PC 200 for extracting the foreground based on the thermal image information received from the thermal image module, analyzing the thermal image data using the neural network model, determining whether a person is present and turning off the main body and the monitor, And a server 200 for monitoring on-off status of a plurality of PCs connected to the plurality of PCs located in the office through a network.

3 is a control flowchart of a method for automatically saving power of a PC using a thermal image sensor and a neural network model of the present invention. 3, a PC automatic power saving method using a thermal image sensor and a neural network model according to the present invention includes a step S11 of initializing a PC, a step S11 of transferring thermal image data of an object sensed by the sensor to a PC A step (S13) of PC extracting and labeling the foreground by a histogram method, a step (S14) of calculating the feature quantity of the label by the PC by the feature quantity normalization calculation formula, (S15) of predicting the presence or absence of a person using the neural network model based on the presence or absence of a person, a step (S16) of determining whether or not a person is present in the PC, and a step (S17) . The characteristic quantities of the label include the area of the label, the maximum temperature of the label, the movement distance of the label, the temperature standard deviation, the density of the label, the continuity of the label, It can be explained by the differential temperature of the label. In this case, the area of the label is the average value of the number of pixels of the tracking labels, the maximum temperature of the label is the maximum temperature average value of the tracking labels, and the movement distance of the label is the maximum distance (unit is pixels) And the temperature standard deviation represents the temperature standard deviation value of the current frame and the density of the label represents the number of pixels that are true in the AND operation of the 3x3 mask having the current label and all of true, The differential temperature is a value obtained by calculating the differential temperature value between the previous label and the current label. In addition, the normalization calculation formula for estimating each characteristic quantity of the label is calculated as (measurement value-minimum value) / (maximum value-minimum value), and the estimation formula of each characteristic quantity can be calculated by the following equation.

     - Down -

1. Area of label = (measuring area - 4) / (40 - 4),

2. Maximum temperature of label = (Measuring temperature - 28) / (38 - 28),

3. Label travel distance = (measurement travel distance - 0) / (10 - 0),

4. Temperature standard deviation = (standard deviation of measurement - 0) / (3 - 0),

5. Density of label = (measurement density - 0) / (9 - 0),

6. Label continuity = (measurement continuity - 16) / (16 - 0),

7. Differential temperature of label = (measured differential temperature - 0) / (50-0)

4A is a flowchart of a foreground extraction method applied to the present invention. 4A is a flow chart of a foreground extraction method applied to the present invention, wherein (a) is to receive temperature data from a thermal imaging module to form a thermal image, and (b) is based on the received temperature data (C) shows an image obtained by extracting a foreground using the histogram, and FIG. 5 (C) shows an image obtained by extracting a foreground by analyzing the histogram and calculating a threshold value.

4B is a control flowchart for a foreground extraction method applied to the present invention. 4B, the foreground extraction method according to the present invention includes a step S21 of determining whether a standard deviation of a single frame received from the thermographic module is greater than a predetermined condition 1 and a maximum temperature is greater than a predetermined condition 2, A step S22 of expressing the temperature data, a step S23 of setting a threshold value based on the histogram, and a step S24 of separating the foreground and the background based on the threshold value It is characterized by. In addition, in step S21, the maximum temperature may be extracted (S26) if the predetermined condition 1 is less than or equal to the predetermined condition 1 and the predetermined condition 2 is satisfied. In step S21, based on each cell temperature data of a single frame composed of 64 cells received from the thermo-image module, the standard deviation is greater than a constant condition 1 of 0.75 ° C to 0.85 ° C and a maximum temperature of 32 ° C to 34 ° C 2 < / RTI >

5 is a block diagram of a neural network model applied to the present invention. In FIG. 4, the neural network model to be applied to the present invention can be classified into a DB type used for learning, a case where a thermal image sensor faces the head, a thermal image sensor faces the face, In the case of drinking a beverage, drinking a cold drink, making a phone call, or hanging at a distance of 1.5 m from a thermal image sensor, the neural network model to be used has a two-level hidden radar, Area, label temperature, label travel distance, temperature standard deviation, label density, label continuity and label differential temperature, and the output information indicates the presence or absence of a person.

FIG. 6 is a control flowchart of a method for determining whether a person exists using a neural network model applied to the present invention. 5, the method for determining the presence of a person using the neural network model using the neural network model according to the present invention includes the steps of: determining whether the n-second average value of the predicted value, which is the output of the neural network model, If the value is larger than the certain condition 3 value, it is determined that the person exists and the process ends. If it is determined in step S31 that the value is less than the predetermined condition 3, it is determined that no person exists (step S33). In this case, n seconds may be a value ranging from 2 seconds to 4 seconds, and if the maximum probability is 1, it can be determined that a person exists if the predetermined condition 3 value is 0.8 or more.

1: LCD monitor, 2: Infrared sensor,
5: LED converter, 6: control circuit,
100: thermal imaging module, 200: PC,
300: server

Claims (22)

delete delete delete delete A PC automatic power saving method for performing automatic power saving of a PC by determining the presence or absence of a person,
The PC automatic power saving method includes:
A step (S11) of initializing the PC;
Transmitting thermal image data of the object sensed by the thermal imaging module to the PC (S12);
A step (S13) of the PC extracting and labeling the foreground by a histogram method;
Calculating a feature quantity of the label consisting of the area of the label, the maximum temperature of the label, the movement distance of the label, the temperature standard deviation, the density of the label, the continuity of the label and the differential temperature of the label by the feature quantity normalization calculation formula S14);
(S15) of predicting the presence or absence of a person using the neural network model based on the feature quantity;
A step (S16) of determining whether the PC is present or not;
And turning the PC into a power saving mode (S17).
6. The method of claim 5,
The neural network model includes:
And a hidden layer of two stages.
6. The method of claim 5,
The area calculating equation of the label is:
(Measured area - 4) / (40 - 4).
6. The method of claim 5,
The maximum temperature calculation equation of the label is:
(Measured temperature - 28) / (38 - 28).
6. The method of claim 5,
The movement distance calculation formula of the label is,
(Measured moving distance - 0) / (10 - 0).
6. The method of claim 5,
The temperature standard deviation calculation formula is:
(Standard deviation of measurement - 0) / (3 - 0).
6. The method of claim 5,
The density calculation formula of the label,
(Measured density - 0) / (9 - 0).
6. The method of claim 5,
The continuity calculation equation of the label is:
(Measurement continuity - 16) / (16 - 0).
6. The method of claim 5,
The differential temperature calculation equation of the label is:
(Measured differential temperature - 0) / (50 - 0).
A method for extracting a foreground in order to perform automatic power saving of a PC by determining the presence or absence of a person,
The method of extracting the foreground includes:
A step (S21) of judging whether a temperature standard deviation of a single frame received from the thermal imaging module is greater than a predetermined condition 1 and a maximum temperature is greater than a predetermined condition 2
A step (S22) of expressing temperature data by a histogram and
Setting a threshold based on the histogram (S23);
And separating the foreground and the background based on the threshold value (S24).
15. The method of claim 14,
In the constant condition 1,
0.75 캜 to 0.85 캜.
15. The method of claim 14,
In the constant condition 2,
Is within a range of 32 ° C to 34 ° C.
A method for determining the presence of a person by a neural network model to determine whether a person exists or not and to perform automatic power saving of a PC,
A method of determining whether a person exists by using the neural network model,
Operating a neural network model to determine whether an n-second average of the predicted value, which is an output of the neural network model, is greater than a predetermined condition 3;
And determining that a person is present if the predicted value is greater than a predetermined condition 3 value.
18. The method of claim 17,
Wherein n is a number
Wherein the human being is within a range of 2 seconds to 4 seconds.
18. The method of claim 17,
The constant condition 3 value is a value
Wherein the probability is 0.8 or more.
18. The method of claim 17,
The neural network model includes:
And determining the presence or absence of a person based on the neural network model.
18. The method of claim 17,
Wherein the input data of the neural network model comprises:
Wherein the label is composed of the area of the label, the maximum temperature of the label, the moving distance of the label, the temperature standard deviation, the density of the label, the continuity of the label, and the differential temperature of the label.
18. The method of claim 17,
The DB type used for the learning of the neural network model is,
If the thermal image sensor is head-oriented or if the thermal image sensor is facing the face, or if the thermal image sensor is aimed at the body, or if you are drinking hot beverages, drinking cold drinks, The method according to any one of claims 1 to 3,

Images