TWI518642B - Combination of infrared and image recognition device - Google Patents

Description

Device combining infrared and image recognition

The invention provides a technical field of motion detection and tracking of an object, in particular to a device for combining infrared and image recognition, wherein the embedded system module integrates whether the embedded image module moves and the image determines whether there is anybody. And the infrared sensing module senses whether there is any person, and performs comprehensive judgment to have an unmanned result in the environment, thereby improving the accuracy of no one to judge.

According to the idea of promoting smart families in recent years, we will implement home-based applications and extend to the environment, medical care, security, care, energy conservation and other energy-saving applications. In order to realize a smart home, the home automation system must be one of the core. The home automation system uses the network monitoring system combined with micro-processing electronic technology and embedded products to centrally manage and automatically control the electrical products and systems in the home, among which mobile Object detection and tracking technology accounts for the majority. On the other hand, the embedded product market has been increasingly valued by the market in recent years due to the continuous growth of the application industry and the field. As the application surface increases, it also penetrates the daily life of users and matures in network technology. The network also allows the user to achieve the convenience of remote operation control, and the person interacts with the machine. The opportunities are also more and more frequent. The more mature wireless transmission technologies on the market include wireless area network (WiFi), infrared, Bluetooth, ZigBee, etc. ZigBee is most suitable for indoor environment because of its regional networking interaction ability. Such as: family, office.

It can be seen from the above that to achieve smart homes need to have effective object motion detection and tracking technology, with the development of technology, the way of motion detection is more and more diverse, object positioning and tracking has always been a hot topic. The general motion detection method can be divided into Background Subtraction, Temporal Differencing, Optical Flow, and Pyroelectric Infrared Sensor using infrared module. , PIR). Background subtraction is one of the most commonly used methods in motion detection. It is a simple and effective method to detect moving regions by subtracting the current image from the background image, but for dynamic scene changes such as lighting and aliens. Interference from unrelated events is particularly sensitive, so background images must be updated as appropriate. The time series subtraction method uses continuous images to do one-to-one pixel subtraction. If the difference between the two is low, indicating that there is no movement, the image content repeatability is quite high, and only the moving part has a large change, but the disadvantage is that The measured moving image is incomplete and often breaks, which is quite unfavorable for detection and subsequent noise processing. The optical flow method refers to the moving speed of the object pixel in time. The method is to perform mathematical analysis on each pixel of the entire image, and calculate the moving speed of each pixel. The calculation amount is very large, and the hardware requirement is high. Miscellaneous and time consuming, not suitable for use on real-time systems. The pyroelectric type infrared sensor is the most commonly used type of motion detection module. The sensor uses the temperature change to generate a charge phenomenon, and receives infrared rays radiated from all the hot bodies to sense the presence or absence of the human body. The infrared radiation emitted by the radiation or the temperature change is used for switching control. The disadvantage is that it is susceptible to temperature change and misjudgment, and the infrared sensor is a kind of detector that has thermal movement to sense if the person is within the detection range. Still, there is no one in the room. In fact, there are still people. As a result, the lights are turned off, but more problems will arise.

Therefore, in view of the problems existing in the above-mentioned conventional devices, how to develop an innovative device that is more ideal and practical, the consumers are eagerly awaiting, and the relevant industry must strive to develop the goal and direction of breakthrough.

In view of this, the inventor has been engaged in the manufacturing development and design experience of related products for many years. After detailed design and careful evaluation, the inventor has finally obtained the practical invention.

The main object of the present invention is to provide an apparatus for combining infrared light and image recognition, which integrates an embedded image module with an embedded system module to determine whether a moving image module is moved and whether an image is detected by a person, and an infrared sensing module senses whether a person is present. Comprehensive judgment of the results of the unmanned environment, thereby improving the accuracy of no one judged.

In order to achieve the above object, the present invention provides an apparatus for combining infrared light and image recognition, which comprises: an embedded image module, a set of no The line transmission module, an embedded system module and an infrared sensing module, wherein the embedded system module integrates whether the embedded image module moves and the image determines whether there is a person, and the infrared sensing module senses whether someone is present. Conducting a comprehensive judgment environment has unmanned results, thereby improving the accuracy of unclaimed judgments, and having background learning and updating can reduce the variation of eigenvalues caused by external factors or light in the background, resulting in systematic comparison errors. It can also improve the possibility of a person misjudged in the detection environment for a period of time, and can use a plurality of wireless transmission modules of multiple devices to correspond to one receiving end, and use multiple wireless transmission modules to transmit the judgment result. The receiving end of the station is used to detect the environment without dead angle.

The above-mentioned objects, structures and features of the present invention will be described in detail with reference to the preferred embodiments of the present invention. .

1‧‧‧Infrared and image recognition device

10‧‧‧Embedded image module

11‧‧‧ CMOS camera module

20‧‧‧Infrared sensing module

30‧‧‧Embedded system modules

31‧‧‧Background learning algorithm

32‧‧‧Background update algorithm

40‧‧‧Wireless Transmission Module

50‧‧‧ Receiver

S_PIR ‧‧‧Infrared sensing module to determine if someone is

S_move ‧‧‧Embedded image module image to determine whether to move

S_object ‧‧‧ image embedded image module determines whether a person

The first figure is a schematic view of a device member of an embodiment of the present invention.

The second figure is a top plan view of an embodiment of the present invention taken at an angle of 30° to the corner of the wall.

The third figure is a top plan view of an embodiment of the present invention and 45 degrees of corner angle shooting.

The fourth figure is a plan view of the present invention and an angle of 60° taken as an example. Figure.

The fifth figure is a system judgment flow chart of the present invention.

The sixth figure is a schematic diagram of multiple sets of the present invention corresponding to one receiving end.

The present invention provides a designer of an apparatus for combining infrared and image recognition.

In order to enable the reviewing committee to have a better understanding and understanding of the purpose, features and functions of the present invention, the embodiments and the drawings are described in detail as follows: Referring to the first to sixth figures, the present invention provides a combination of infrared rays. And the image recognition device includes: an embedded image module (CMUcam) 10, the embedded image module 10 is provided with a CMOS camera module 11, the embedded image module 10 is capable of transmitting images, and is also built in The programmable image processing software can conveniently perform some arithmetic processing on the image, and the embedded image module 10 can perform moving or still image judgment or image judgment on someone or no one; the embedded image module 10 (CMUcam) judgment method Whether the image has a movement judgment: the embedded image module 10 reads the feature value of a real-time image every 0.5 seconds, and compares the feature values of the continuous image before and after, if the difference after the comparison is greater than the set valve The value is three times in succession, that is, it is judged that there is movement in the environment, and if it is less than the threshold value and three times in succession, it is judged that the environment is stationary. Whether the image is judged by someone: after the embedded image module 10 is turned on, the first image that will be captured is The eigenvalue is used as the background image feature value, and the feature value of the real-time image is compared with the feature value of the background image. If the difference is greater than the threshold set by the user and three times in succession, the environment is determined to be a person, and vice versa. If the threshold is three consecutive times, it is judged that there is no one in the environment. The sampling time, the number of judgments, and the threshold in the judgment method are all adjustable parameters.

An infrared sensing module (PIR Sensor) 20, the infrared sensing module 20 senses a person or a person; the infrared sensing module 20 determines the method, and the infrared sensing module 20 senses the body's radiation The infrared or temperature change determines whether there is anyone in the environment. If a change is sensed, it means that there is someone in the detection environment. If the change is not sensed, it means that the detection environment is unmanned. The infrared sensing module 20 senses whether or not a person has judged: the infrared sensing module 20 reads the infrared information in the environment every 0.5 seconds, and when three consecutive infrared change information is generated, it is determined to be in a human state. On the other hand, when there is no change in infrared information for three consecutive times, it is judged to be an unmanned state. The sampling time and the number of judgments in the judgment method are all adjustable parameters.

An embedded system module (Arduino Mega ADK) 30, the embedded system module 30 electrically connects the embedded image module 10 and the infrared sensing module 20, respectively, the embedded system module 30 integrates the embedded system Whether the image module 10 moves and the image determines whether there is a person, and the infrared sensing module 20 senses whether there is a person, and performs comprehensive judgment to have an unmanned result in the environment, thereby improving the accuracy of unclaimed judgment; a group of wireless transmission modules 40, the group of wireless transmission modules 40 The embedded system module 30 is electrically connected, and the wireless transmission module 40 wirelessly transmits the module data of the embedded system 30; wherein, the wireless transmission module 40, one embedded system module 30, and one The infrared sensing module 20 and the embedded image module 10 form a set of devices 1 for combining infrared and image recognition; and a receiving end 50, a plurality of sets of devices 1 for combining infrared and image recognition The wireless transmission module 40 has to transmit a determination result corresponding to a receiving end 50 to detect the environment without dead angle.

The embedded system module 30 integrates the determination method, and first defines three states S_PIR , S_move , and S_object , respectively, for the infrared sensing module 20 to determine whether someone ( S_PIR ), the embedded image module 10 image determines whether to move ( S_move ), and embed The image module 10 determines whether there is a person ( S_object ), and updates the state when the embedded system module 30 accumulates the same judgment result three times. The system sampling frequency is sampled once in 0.5 second. The image is judged whether the movement is the subtraction of the image features before and after, and the image is judged whether the image feature is reduced by the image feature. The judgment method is shown in the table below. All parameters and thresholds can be adjusted.

The device for combining infrared and image recognition, wherein the embedded system module 30 further includes a background learning algorithm 31. Since the environment has slight changes with time, in order to improve the problem and adapt the system to new In the environment, a background learning algorithm 31 is added to the embedded system module 30. The background learning algorithm 31 can reduce the variation of the eigenvalues caused by external factors or light in the background, and cause errors in the system comparison, and can also improve the possibility of misjudgment caused by the person not being in motion for a certain period of time in the detection environment. Sex. The background learning algorithm 31 is that when the S_PIR , S_move , and S_object states are all OFF, the embedded system module 30 learns a new background, and its formula is F_background = η. F_image , where F_background is the feature value of the background image, F_image is the feature value of the live image, and η is the learning rate. The learning rate η in the background learning algorithm 31 is an adjustable parameter.

The device is combined with the infrared and image recognition device, wherein the embedded system module 30 further includes a background update algorithm 32. Since the environment has slight changes with time, in order to improve the problem and adapt the system to new In the environment, a background update algorithm 32 is added to the embedded system module 30. By using the background update algorithm 32, the background value is changed by the external factors or the light, and the eigenvalue changes, which causes the system to compare errors, and can also improve the possibility that the person is completely inactive in the detection environment for a period of time. Sex. The background update algorithm 32 is when the S_PIR , S_move state is OFF and the S_object state is ON, and after accumulating for a period of time, the system updates the background, and the formula is F_background = F_image , where F_background is the feature value of the background image, and F_image is the instant. The eigenvalue of the image. The accumulated time in the background update algorithm 32 is an adjustable parameter.

The device combines infrared and image recognition, wherein the embedded image module 10 can be a CMUcam4 programmable embedded image sensor.

The device for combining infrared and image recognition, wherein the set of wireless transmission modules 40 can be a wireless local area network (WiFi), infrared, Bluetooth or ZigBee.

The infrared sensing module 20 combines the wavelength sensitivity of the detecting component to be almost constant within a range of 0.2 to 20 μm, and the infrared sensing module 20 is mounted with a wavelength. A filter having a range of 7 to 10 μm can be used for detecting infrared radiation of the human body, which is specifically used for detecting the radiation of the human body.

The receiving end 50 refers to any device that can read information and perform subsequent processing and calculation of information, such as an embedded system, a tablet computer, a mobile phone, and the like.

The present invention has been described with reference to the preferred embodiments of the present invention, and those skilled in the art can change and modify the present invention without departing from the spirit and scope of the invention. Such changes and modifications shall be covered by the scope of the following patent application.

In summary, the present invention provides a device for combining infrared light and image recognition, which has indeed achieved all the objects of the present invention, and the spatial pattern of the combined structure is not found in the same product, nor has it been disclosed before the application. In accordance with the provisions of the Patent Law, 提出 apply in accordance with the law.

1‧‧‧Infrared and image recognition device

10‧‧‧Embedded image module

11‧‧‧ CMOS camera module

20‧‧‧Infrared sensing module

30‧‧‧Embedded system modules

31‧‧‧Background learning algorithm

32‧‧‧Background update algorithm

40‧‧‧Wireless Transmission Module

Claims (7)

A device for combining infrared and image recognition includes: an embedded image module, the embedded image module is provided with a CMOS camera module, the embedded image module is capable of transmitting images, and a programmable image processing software is built in It is convenient to perform some arithmetic processing on the image. The embedded image module has to perform moving or still image judgment or image judgment on someone or no one; an infrared sensing module, the infrared sensing module has to sense someone or No. An embedded system module, the embedded system module electrically connects the embedded image module and the infrared sensing module respectively, and the embedded system module integrates whether the embedded image module moves and whether the image is judged whether Some people, and the infrared sensing module senses whether there is anyone, and the result of the unambiguous environment in the comprehensive judgment environment, thereby improving the accuracy of unclaimed judgment; and a set of wireless transmission modules, the wireless transmission module electrically connecting the embedded System module, the wireless transmission module can wirelessly transmit the embedded system module data; and form a combination of infrared and image processing The device is further configured to include a background learning algorithm, wherein the background learning algorithm determines whether the person ( S_PIR ) and the image of the embedded image module determine whether the image is moved by the infrared sensing module. ( S_move ) and the image of the embedded image module determine whether the state of any ( S_object ) is OFF, the embedded system module learns a new background, and its formula is F_background = η. F_image , where F_background is the feature value of the background image, F_image is the feature value of the instant image, η is the learning rate, and the learning rate η in the background learning algorithm is an adjustable parameter. The device of claim 1, wherein the embedded system module further includes a background update algorithm, wherein the background update algorithm is when the S_PIR and S_move states are OFF and S_object state is ON, and the accumulated period of time, the system update the background, the formula is F_background = F_image, which is the feature value F_background background image, F_image real-time image as the feature value, the cumulative time background update algorithm is to be Adjust the parameters. The device for combining infrared and image recognition according to claim 2, wherein a plurality of sets of the wireless transmission module combined with the infrared and image recognition device are corresponding to a receiving end to transmit a judgment result, Reach the environment without detection of dead ends. The device for combining infrared and image recognition according to claim 3, wherein the receiving end can be an embedded system, a tablet computer, a mobile phone or other information that can be read and the information is processed subsequently. And computing device. The device for combining infrared and image recognition according to claim 2, wherein the embedded image module is a CMUcam4 programmable embedded image sensor. The device for combining infrared and image recognition according to claim 2, wherein the wireless transmission module can be a wireless area network (WiFi), infrared, Bluetooth or ZigBee. The device for combining infrared and image recognition according to claim 2, wherein the infrared sensing module has a wavelength sensitivity of the detecting element which is almost constant within a range of 0.2 to 20 μm, and the infrared sensing is performed. The filter is assembled with a filter having a wavelength range of 7 to 10 μm, and the infrared sensing module specially used for detecting human body radiation can be obtained.