System and method for classification of objects by a body taper detection
A method is provided for classification of objects by body taper detection. The method includes receiving point cloud data of a plurality of objects from one or more 3D sensors disposed in a space to be monitored, pre-processing the fused point cloud data to search, detect and segment the plurality of objects, each segmented object comprising a set of horizontal layers of interception from the 3D sensor representing the structure of the segmented object, calculating an inference ratio of each segmented object using lengths of respective horizontal layers, distances between set of horizontal layers and a distance of each segmented object from the one or more 3D sensors, checking whether the inference ratio of the segmented object lies on best fit curves pre-generated using a plurality of golden ratios associated with a particular class of objects, and accordingly determining the class of each of the segmented object.
1. A method ( 200 ) for classification of objects by a body taper detection, the method ( 200 ) comprising:
receiving ( 210 ) point cloud data of a plurality of objects from one or more 3D sensors ( 102 ) disposed in a space to be monitored;
pre-processing ( 220 ) the point cloud data to search, detect and segment the plurality of objects, each segmented object comprising a set of horizontal layers of interception from the one or more 3D sensors ( 102 ) representing the structure of the segmented object;
calculating ( 230 ) an inference ratio of each segmented object using lengths of respective horizontal layers, distances between set of horizontal layers and a distance of each segmented object from the one or more 3D sensors ( 102 );
checking ( 240 ) whether the inference ratio of the searched, detected & segmented object lies on best fit curves pre-generated using a plurality of golden ratios associated with a particular class of objects; and
determining ( 250 ) the class of each of the searched, detected and segmented object when the inference ratio successfully lies on a best fit curve associated with the particular class of object, thereby classifying the plurality of objects in real-time.
2. The method ( 200 ) as claimed in claim 1 , further comprising a step of calibrating the point cloud data to a single origin, in case there are more than one 3D sensors disposed in the space to be monitored.
3. The method ( 200 ) as claimed in claim 1 , wherein the plurality of objects are selected from one or more of living and non-living objects including humans of multiple age groups, animals, plants, furniture, vehicles, natural resources, eatables, crops, infrastructure, stationery, sign boards, wearables, musical instruments, sports equipment, mechanical tools, electrical equipment and electronic equipment.
4. The method ( 200 ) as claimed in claim 1 , wherein the step of calculating the inference ratio of each segmented object includes determining a length of each of the set of horizontal layers of interception from the 3D sensor representing the segmented object.
5. The method ( 200 ) as claimed in claim 1 , wherein the class of objects are selected from humans, animals, furniture, automobiles, tools and electronics.
6. The method ( 200 ) as claimed in claim 5 , wherein the plurality of best fit curves associated with the plurality of classes are generated using machine learning by:
calculating a plurality of ratios of a set of objects belonging to a particular class at a plurality of distances from the one or more 3D sensors and obtaining a respective golden ratio for the particular class at the plurality of distances;
obtaining the golden ratios of the plurality of classes of the plurality of objects;
generating the best fit curve for each of the plurality of classes using the respective plurality of golden ratios and the plurality of distances;
wherein the golden ratio for each class of objects is continuously improved by increasing a number/size of the ‘set’ of objects belonging to the same class used for calculation of the golden ratio, during modelling and training.
7. The method ( 200 ) as claimed in claim 1 , wherein the one or more 3D sensors ( 102 ) are selected from a Light Detection and Ranging (LIDAR) sensor, Laser Detection and Ranging (LaDAR), Light Emitting Diode Detection and Ranging (LeDDAR) mmWave Radar, C or K Band Radar, Time of Flight (ToF) sensors, laser scanners or a combination thereof.
8. The method ( 200 ) as claimed in claim 1 , further comprising a step of visual identification of the plurality of objects after classification, using one or more visual cameras.
9. A system ( 100 ) for classification of objects by a body taper detection, the system ( 100 ) comprising:
one or more 3D sensors ( 102 ) disposed in a space to be monitored; and
a processing module ( 104 ) connected with each of the one or more 3D sensors ( 102 ), the processing module ( 104 ) comprising:
a memory unit ( 1042 ) configured to store machine-readable instructions; and
a processor ( 1044 ) operably connected with the memory unit ( 1042 ), the processor ( 1044 ) obtaining the machine-readable instructions from the memory unit ( 1042 ), and being configured by the machine-readable instructions to:
receive point cloud data of a plurality of objects from the one or more 3D sensors ( 102 );
pre-process the point cloud data to search, detect and segment the plurality of objects, each segmented object comprising a set of horizontal layers of interception from the 3D sensor representing the structure of the segmented object;
calculate an inference ratio of each segmented object using lengths of respective horizontal layers, distances between set of horizontal layers and a distance of each segmented object from the one or more 3D sensors ( 102 );
check whether the inference ratio of the searched, detected & segmented object lies on best fit curves pre-generated using a plurality of golden ratios associated with a particular class of objects; and
determine the class of each of the searched, detected and segmented object when the inference ratio successfully lies on a best fit curve associated with the particular class of object, thereby classifying the plurality of objects in real-time.
10. The system ( 100 ) as claimed in claim 9 , wherein the processor ( 1044 ) is further configured to calibrate the point cloud data to a single origin, in case there are more than one 3D sensors disposed in the space to be monitored.
11. The system ( 100 ) as claimed in claim 9 , wherein the plurality of objects are selected from one or more of living and non-living objects including humans of multiple age groups, animals, plants, furniture, vehicles, natural resources, eatables, crops, infrastructure, stationery, sign boards, wearables, musical instruments, sports equipment, mechanical tools, electrical equipment and electronic equipment.
12. The system ( 100 ) as claimed in claim 9 , wherein for calculating the inference ratio of each segmented object, the processor ( 1044 ) is configured to determine a length of each of the set of horizontal layers of interception from the 3D sensor representing the segmented object.
13. The system ( 100 ) as claimed in claim 9 , wherein the class of objects are selected from humans, animals, furniture, automobiles, tools and electronics.
14. The system ( 100 ) as claimed in claim 13 , wherein the plurality of best fit curves associated with the plurality of classes are generated using machine learning by:
calculating a plurality of ratios of a set of objects belonging to a particular class at a plurality of distances from the one or more 3D sensors and obtaining a respective golden ratio for the particular class at the plurality of distances;
obtaining the golden ratios of the plurality of classes of the plurality of objects;
generating the best fit curve for each of the plurality of classes using the respective plurality of golden ratios and the plurality of distances;
wherein the golden ratio for each class of objects is continuously improved by increasing a number/size of the ‘set’ of objects belonging to the same class used for calculation of the golden ratio, during modelling and training.
15. The system ( 100 ) as claimed in claim 9 , wherein the one or more 3D sensors ( 102 ) are selected from a Light Detection and Ranging (LIDAR) sensor, Laser Detection and Ranging (LaDAR), Light Emitting Diode Detection and Ranging (LeDDAR) mmWave Radar, C or K Band Radar, Time of Flight (ToF) sensors, laser scanners or a combination thereof.
16. The system ( 100 ) as claimed in claim 9 , wherein the processor ( 1044 ) is further configured to facilitate visual identification of the plurality of objects after classification, using one or more visual cameras.