Outdoor air quality estimation and early warning system based on double cameras and environment sensor data

[0017] The preferred embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0018] Such as figure 1 As shown, an outdoor air quality estimation and early warning system based on dual cameras and environmental sensor data is functionally composed of cameras, environmental sensors, image acquisition and preprocessing modules, environmental data acquisition modules, main control modules, and air It consists of a quality estimation module and an air quality warning module. It includes two identical cameras, environmental sensors, an environmental sensor control board with wireless transmission function, and a main control board. Preferably, the camera chooses two identical USB interface cameras. In order to save the Pin port of the control board, the environment sensor chooses the Bosch 4-in-1 environment sensor BME680, and the environment sensor control board with wireless transmission function chooses portable and low Power Bluetooth Adafruit Feather MO Bluefruit LE development board, the main control board selects Raspberry Pi 3 (Bluetooth function).

[0019] The air quality assessment and early warning system of the embodiment of the present invention includes the following steps:

[0020] Step 1. Turn on the two cameras;

[0021] Step 2. Run the image acquisition and preprocessing module, including the following sub-steps:

[0022] Step 2.1, according to the dual-camera ranging model, use the Raspberry Pi3 to control the two cameras to shoot outdoor scene targets. It is recommended to have at least 2 target areas with different distances in the scene;

[0023] Step 2.2: Obtain the corresponding parameters such as focal length f, parallax d, and camera center distance Tx through binocular calibration, calibration and matching;

[0024] Step 2.3, and then calculate the distance L1, L2 of the two target points respectively according to the parameters.

[0025] Step 2.4: Crop the target object images with distances of L1 and L2 respectively to obtain two sampled images of the same size.

[0026] Step 3. Run the environmental sensor data acquisition module, through the Adafruit Feather MO Bluefruit LE development board, according to the data transmission protocol, extract the environmental sensor data, including temperature, humidity, air pressure, VOC and other data, and then transmit to the Raspberry Pi via Bluetooth ;

[0027] Step 4. Run the ring air quality estimation module. The implementation process includes the following sub-steps:

[0028] Step 4.1, sample image input, input the cropped sample image obtained before into the system.

[0029] Step 4.2: Calculate the dark channel brightness value of each sampled image through the image defogging algorithm with dark channel priority:

[0030]

[0031] Among them, I (m, n, c) is the brightness value observed at the midpoint (m, n) of the color channel c, I aark (x, y) is the brightness value of the point (x, y) in the dark channel.

[0032] Step 4.3, establish the relationship model of extinction coefficient, target point distance and PM2.5 concentration according to the brightness value of the dark channel;

[0033] Step 4.4: Input the sampled image data into the model. According to the different extinction coefficients of the images at different distances, the estimated value of PM2.5 concentration can be judged by the relative extinction coefficient of the unit distance.

[0034] Step 5. Run the ambient air quality early warning module. The implementation process includes the following sub-steps:

[0035] Step 5.1, establish the relationship model between temperature, humidity, air pressure and PM2.5 concentration and the change trend of polluted gas concentration;

[0036] Step 5.2, input the change trend of the temperature, humidity, and air pressure in the past hour into the model to obtain the change trend of PM2.5 concentration and polluted gas concentration.

[0037] The specific embodiments described above are only examples of the present invention. Those skilled in the art to which the present invention belongs can modify or supplement the described specific implementations or use similar alternatives, but the changes in the design concept according to the present invention will not exceed the protection scope defined by the claims.