Detection of area of abnormal air quality within a geographical area

Patent metadata

Field	Value
Publication number	US-11307187-B2
Application number	US-201916589714-A
Country	US
Kind code	B2
Filing date	Oct 1, 2019
Priority date	Oct 1, 2019
Publication date	Apr 19, 2022
Grant date	Apr 19, 2022

Abstract

An abnormal area is detected using an initial spatial weights matrix between pairs of air quality sensors in a plurality of air quality sensors distributed across a geographical area and air quality data for each air quality sensor. The spatial weights matrix utilizes a distance between pairs of air quality sensors and wind direction through the geographical area. The initial spatial weights matrix and air quality data are used to calculate a plurality of local moran's indexes, one for each air quality sensor. The plurality of local moran's indexes are used to divide the plurality of air quality sensors into four groups. The groups are classified as proper or improper, and the proper groups are identified as abnormal areas.

First claim

What is claimed is:

1. A method for identifying an abnormal area, the method comprising:
identifying an initial spatial weights matrix between pairs of air quality sensors from a plurality of air quality sensors distributed across a geographical area, the spatial weights matrix utilizing a distance between each given pair of air quality sensors and wind direction through the geographical area;
obtaining air quality data from each air quality sensor;
using the initial spatial weights matrix and air quality data to calculate a plurality of local moran's indexes, one for each air quality sensor;
using the plurality of local moran's indexes to divide the plurality of air quality sensors into four groups;
classifying each one of the four groups as a proper group or an improper group; and
designating each group classified as a proper group for supervision as an abnormal area.

2. The method of claim 1 , wherein identifying the initial spatial weights matrix comprises:
calculating a distance between each given pair of air quality sensors using location coordinates associated with each air quality sensor;
determining a node vector between each given pair of air quality sensors;
determining the wind direction; and
setting a spatial weight between each given pair of air quality sensors based on a comparison between the wind direction and the node vector associated with the given pair.

3. The method of claim 2 , wherein setting the spatial weight between each given pair of air quality sensors based on a comparison between the wind direction and the node vector further comprises:
defining the spatial weight between the given pair of air quality sensors as a cosine of an angle between the wind direction and the node vector divided by the distance between the given pair of air quality sensors when the given pair of air quality sensors has at least a node vector component extending in the wind direction; and
defining the spatial weight between the given pair of air quality sensors as the negative of the cosine of the angle between the wind direction and the node vector divided by the distance between the given pair of air quality sensors when the given pair of air quality sensors lack at least a node vector component extending in the wind direction.

4. The method of claim 1 , wherein the four groups comprise a statistically significant cluster group of high air quality values, a statistically significant cluster group of low air quality values, a high value outlier group comprising a high air quality value surrounded by low air quality values, and a low value outlier group comprising a low air quality value surrounded by high air quality values.

5. The method of claim 1 , wherein classifying each one of the four groups comprises using a supervised classification model.

6. The method of claim 5 , wherein the method further comprises building the supervised classification model by:
obtaining a plurality previously defined groups of air quality sensors, each previously defined group comprising air quality sensor locations, air quality data and a classification label, the classification label comprising proper or improper; and
using each previously defined group to train the supervised classification model with features as model inputs and the classification label as a model output.

7. The method of claim 6 , wherein the features comprise a largest inner group air quality data difference, a median inner group air quality difference, a largest whole group air quality data difference, and a geographical size.

8. The method of claim 1 , wherein the method further comprises:
subdividing each group classified as an improper group into four sub-groups.

9. The method of claim 8 , wherein subdividing each improper group comprises:
obtaining air quality data from each air quality sensor in each improper group;
using the initial spatial weights matrix and air quality data to calculate a plurality of local moran's indexes, one for each air quality sensor in each improper group;
using the plurality of local moran's indexes to divide the plurality of air quality sensors in each improper group into four sub-groups; and
classifying each one of the four sub-groups from each improper group as either a proper sub-group or an improper sub-group using the supervised classification model.

10. The method of claim 9 , wherein the four sub-groups comprise a statistically significant cluster of high air quality values, a statistically significant cluster of low air quality values, a high value outlier comprising a high air quality value surrounded by low air quality values, and a low value outlier comprising a low air quality value surrounded by high air quality values.

11. A computer-readable medium containing a computer-readable code that when read by a computer causes the computer to perform a method for identifying an abnormal area, the method comprising:
identifying an initial spatial weights matrix between pairs of air quality sensors from a plurality of air quality sensors distributed across a geographical area, the spatial weights matrix utilizing a distance between each given pair of air quality sensors and wind direction through the geographical area;
obtaining air quality data from each air quality sensor;
using the initial spatial weights matrix and air quality data to calculate a plurality of local moran's indexes, one for each air quality sensor;
using the plurality of local moran's indexes to divide the plurality of air quality sensors into four groups;
classifying each one of the four groups as a proper group or an improper group; and
designating each group classified as a proper group for supervision as an abnormal area.

12. The computer-readable medium of claim 11 , wherein identifying the initial spatial weights matrix comprises:
calculating a distance between each given pair of air quality sensors using location coordinates associated with each air quality sensor;
determining a node vector between each given pair of air quality sensors;
determining the wind direction; and
setting a spatial weight between each given pair of air quality sensors based on a comparison between the wind direction and the node vector associated with the given pair.

13. The computer-readable medium of claim 11 , wherein setting the spatial weight between each given pair of air quality sensors based on a comparison between the wind direction and the node vector further comprises:
defining the spatial weight between the given pair of air quality sensors as a cosine of an angle between the wind direction and the node vector divided by the distance between the given pair of air quality sensors when the given pair of air quality sensors has at least a node vector component extending in the wind direction; and
defining the spatial weight between the given pair of air quality sensors as the negative of the cosine of the angle between the wind direction and the node vector divided by the distance between the given pair of air quality sensors when the given pair of air quality sensors lack at least a node vector component extending in the wind direction.

14. The computer-readable medium of claim 11 , wherein the four groups comprise a statistically significant cluster group of high air quality values, a statistically significant cluster group of low air quality values, a high value outlier group comprising a high air quality value surrounded by low air quality values, and a low value outlier group comprising a low air quality value surrounded by high air quality values.

15. The computer-readable medium of claim 11 , wherein the method

Assignees

• IBM

Inventors

• Qu Junmei
• Wang Lingyun
• Bai Xin Xin
• Xia Xi
• Shao Jin Yan

Classifications

• H04L67/52

  specially adapted for the location of the user terminal · CPC title

• G01N33/0075 Primary

  for multiple spatially distributed sensors, e.g. for environmental monitoring · CPC title

• G08B21/12

  responsive to undesired emission of substances, e.g. pollution alarms · CPC title

• G01W1/02

  Instruments for indicating weather conditions by measuring two or more variables, e.g. humidity, pressure, temperature, cloud cover or wind speed (G01W1/10 takes precedence) · CPC title

• G08C25/04

  by recording transmitted signals · CPC title