Method and system for processing spatial data

1. A method for classifying a spatial data carried out by a data computing environment, comprising:

receiving a spatial data from a data source;

generating a first feature from the spatial data;

dividing the first feature into a first sub-feature and a second sub-feature;

analysing the first sub-feature to derive a first sub-feature data;

analysing the second sub-feature to derive a second sub-feature data;

using the first sub-feature data and the second sub-feature data as a first input data for analysing the first feature;

analysing at least one of the first sub-feature data, the second sub-feature data, and the first feature to classify the spatial data into a plurality of object classes;

computing a benefit criteria from a division of each sub-feature; and

recursively dividing each sub-feature into a pair of further sub-features when the benefit criteria from the division is greater than a specified threshold and until a termination criteria is reached.

2. The method according to claim 1 , wherein the analysing of the first sub-feature comprises:

calculating a first representing value for the first sub-feature; and

dividing the first sub-feature to a third sub-feature and a fourth sub-feature if the first representing value is greater than a first threshold value.

3. The method according to claim 2 , wherein the analysing of the first sub-feature further comprises:

analysing the third sub-feature and the fourth sub-feature to derive a third sub-feature data and a fourth sub-feature data respectively; and

utilising the third sub-feature data and the fourth sub-feature data as a second input data to analyse the first sub-feature.

4. The method according to claim 1 , wherein the analysing of the second sub-feature comprises:

calculating a second representing value for the second sub-feature; and

dividing the second sub-feature to a fifth sub-feature and a sixth sub-feature if the second representing value is greater than a second threshold value.

5. The method according to claim 4 , wherein the analysing of the second sub-feature further comprises:

analysing the fifth sub-feature and the sixth sub-feature to derive a fifth sub-feature data and a sixth sub-feature data respectively; and

utilising the fifth sub-feature data and the sixth sub-feature data as a third input data to analyse the second sub-feature.

6. The method according to claim 1 , wherein the dividing of the first feature comprises:

creating a characteristic vector for the first feature using points associated with the first feature; and

using a trained deep neural network to determine whether to divide the first feature and a model pattern to divide the first feature into the first sub-feature and the second sub-feature based on the created characteristic vector.

7. The method according to claim 1 , wherein the benefit criteria comprises at least one of: a decrease in a computational cost at the data computing environment, an increase in performance defined by an accuracy of object classification, or an improvement in a computational cost to performance ratio, and wherein the termination criteria is reached when a number of points in a given sub-feature is less than the specified threshold number of points, or a physical dimension associated with the given sub-feature is less than a specified size threshold.

8. The method according to claim 1 , wherein the method further comprises generating a hierarchy of a plurality of features to determine a deep neural network structure, wherein the plurality of features comprises the first feature and at least one sub-feature.

9. The method according to claim 8 , wherein the generating the hierarchy of the plurality of features further comprises projecting each point associated with each feature of the plurality of features into a local coordinate system within a feature.

10. The method according to claim 1 , wherein the method further comprises applying the trained deep neural network to classify the spatial data into the plurality of object classes.

11. The method according to claim 1 , wherein the method further comprising iteratively reclassifying one or more object classes in the plurality of object classes to one or more new object classes based on an induced classification information associated with each previously classified object.

12. The method according to claim 1 , wherein the spatial data is a spatial point cloud data or a spatial two-dimensional data.

13. A data computing apparatus for classifying a spatial data the data computing apparatus comprising a non-transitory computer readable medium including machine readable instructions which when executed by a processor are configured to cause the data computing apparatus to:

receive a spatial data from a data source;

generate a first feature from the spatial data;

divide the first feature into a first sub-feature and a second sub-feature;

analyse the first sub-feature to derive a first sub-feature data;

analyse the second sub-feature to derive a second sub-feature data;

utilise the first and the second sub-feature data as a first input data for analysing the first feature;

analysing at least one of the first sub-feature data, the second sub-feature data, and the first feature to classify the spatial data into a plurality of object classes;

computing a benefit criteria from a division of each sub-feature; and

recursively dividing each sub-feature into a pair of further sub-features when the benefit criteria from the division is greater than a specified threshold and until a termination criteria is reached.

14. The data computing apparatus according to claim 13 , wherein the spatial data is a spatial point cloud data or a spatial two-dimensional data.

15. A computer program product for classifying a spatial data residing on a non-transitory computer readable medium comprising machine readable instructions which, when executed by a a processor, cause the processor to carry out steps of:

receiving, by the data computing environment, a spatial data from a data source;

forming a first feature from the spatial data;

dividing the first feature into a first sub-feature and a second sub-feature;

analysing the first sub-feature to derive a first sub-feature data;

analysing the second sub-feature to derive a second sub-feature data;

using the first and the second sub-feature data as a first input data for analysing the first feature;

analysing at least one of the first sub-feature, the second sub-feature, and the first feature to classify the spatial data into a plurality of object classes;

computing a benefit criteria from a division of each sub-feature; and

recursively dividing each sub-feature into a pair of further sub-features when the benefit criteria from the division is greater than a specified threshold and until a termination criteria is reached.

16. The computer program product according to claim 15 , wherein the spatial data is a spatial point cloud data or a spatial two-dimensional data.