Get the (x,y) position of a given node with id = "node_id" Arguments: node_id {[int]} -- The node id Returns:

Method to log all the clustered points. The points will be appended to the cluster log file as indicated by the global variable

Formats the search query

Returns the required attributes of the car object in a particular format.

Gets the resolved results and formats the car objects

A generic function which receives a queryObject and uses it to obtain results from cassandra. The queryObject consists of the query and the params.

This method computes the eucledian distance between the detections at timestep (t-1) (indicated by prev_json_list) and detetions at

Method to log all the matched points. The tracker matches clusters at tiemstamp (t-1) with those at timestamp t. The points will be a

The tracker adds additional fields to the detection dictionaries that are passed. This will violate the day2 schema. Use this method

Transfer vehicle attributes such as license plate, make, color from one representative record to all the records in the list

If the vehicle detected does not have any attributes (e.g., license plate, color), then add dummy attributes Arguments:

Returns if two detections are from conflicting cameras or not Arguments: ele1 {[dict]} -- first detection (in day2 schem

A generic function which receives a queryObject and uses it to obtain results from cassandra. The queryObject consists of the query and the params.

Get all rectangles for the point the_pt Arguments: rect {KDTree rect} -- Input rectangle kdtree the_pt {[type]} -- [descript

Get the list of ids associated with a given detection Arguments: json_ele {[dict]} -- Day2 schema dictionary Returns:

This method maps all points in the json_list to the nearest points on the map_info Arguments: json_list {[list]} -- The l

This method normalizes the distance between an array of (x,y) points based on the minimum x,y and range of x,y. Specifically,

This is the main method that will be called for multicam tracking. The detections are passed in day2 schema in a list (all_json_list)

This method selects one representative dict from the list of vehicle detections (in json_list) Returns: [dict] -

This method replaces a vehicle detections in a list by a representative point (x,y). Currently, the logic is: computed (x

Truncate the angle between [0, 360) Arguments: the_angle {float} -- angle Returns: float -- Angle between [0,360)]

Build the network and create an indexing structure (kdtree)

Build the road-network

Returns if two detections are from overlapping cameras or not Arguments: ele1 {[dict]} -- first detection (in day2 schem

Check if a point is present. A point will be added if there are no other points within the "min_dist" neighborhood Arguments:

This method closes all the open debug files. This function has to be explicity called after multi-cam tracking module is finished

Each camera may have multiple frames within a given resample period. In such cases, single object may be detected multiple times. If

This method collates objects of same tracker id and from the same sensor into a single object Arguments: json_li

Given all the objects in json_list, create a list of all ids for as a list and return the list of ids Arguments:

Create a kdtree of points

Densify a line ("line") Arguments: line {[type]} -- [description] hypotenuse {[type]} -- [description] Returns:

Write all the tracking timings into the file specified by DEFAULT_TRACKER_LOGS_FILE

Flush the log files, if logging is enabled

Generates the cassandra aisle query for live/playback system * and formats the results

The result is compressed so that a variety of events can be obtained for various detected objects. * Moving events for a particular license plate is

Queries and provides the result for entry and exit flowrate

Returns the formatted timestamp Range

Used to get the moving car dictionary

Calculates latitude of object

Calculates longitude of object

Resolves the queries generated for live and playback mode

Generates subsequent cassandra parking spot queries for live/playback system

Generates the first cassandra parking spot query for playback system

Generates cassandra query for playback system

Used to decide which kind of parking spot query needs to be made

Generates cassandra query for live system

Generates the first cassandra parking spot query for live system

Calculates the number of occupied and available spots based on the parked events

Aggregates the parking spot and aisle related results into a single list after post processing

Used to generate parkingspot and aisle results asynchronously

Used to obtain the start and end index of search tokens in a complex search query string

Get the angle between two points defined by (var_x1, var_y1) and (var_x2, var_y2) Arguments: var_x1 {float} -- X-value for the f

This method clusters all vehicle detections in the json_list. Currently it: 1. computes distance matrix between vehicle detec

Get the direction between detections ele1 and ele2 detections. This is usually used if the same object moves between two locations

Get the nearest point to "the_pt" using the the kd=tree index Arguments: kd {[KDTree]} -- The KD-tree of all the points the_

Get the object id value of a detected vehicle. The object id returned in this function is a string of the the object id Arguments:

This method computes the ID distance between the detections at timestep (t-1) (indicated by prev_json_list) and detetions at

Get the shortest path between given source and dest Arguments: start_point_id {[list]} -- [x,y] of the source point d

Given a point (argument "point") and road_network, return the nearest point to given "point" on road-network Arguments:

This method parses the json_list and categorizes them into four types of records based on the event observed. a. Parked recor

This method is called as an init method before passing the day2 schema. Some of the work done by this method is: 1. Change th

Checks if the day2 dict is within the given timerange [start_time, end_time) NOTE: start_time is inclusive, and end_time is exclusive If s

Method to log all the input points that the tracker is receiving. The points will be appended to the input log file as indicated by t

Main function. Starts multicam tracker and runs continiously until killed

Maintain the object ids of the clustered trajectories Arguments: clustered_json_list {list} -- list of detections in a single clu

This method matches the vehicle detection (argument "json_ele") to the road-network. The (x,y) will be snapped to the road-network ed

This is the main matching function that matches detections at timestep (t-1) (prev_json_list) with detections at timestep (t)

Method to merge all ids for a single cluster given that two or more ids (in merge_obj_id_list) Arguments: merge_

This method inputs two distance matrices: a. spatial distance matrix (dist_matrix): This is a matrix with distances betwee

Prune the carry over list. The cars that are not matched are carried over to the next timestep. However, if the car that was carried

Prune cluster object ids which were updated long back. The time thershold is given by self.config.cluster_obj_id_staytime_in_sec Argu

This method will cluster all nearby moving points. The following is done: 1. The clustering of moving points is done by calli

This method clusters all points in the list (json_list). The method returns a list of points, where nearby points (in one cluster) ha

Refine the graph by removing close points Arguments: lines_arr {[list]} -- List of lines max_point_dist {[fl

If an empty spot is declared, then the car that was pulled has to be removed from the parking-spot-state-store. Arguments:

Remove all unnecessary records Unnecessary records = ["reset] Arguments: gt_json_list {[list]} -- List of Day2 detection dictionaries

The cars in the aisle which haven't moved for certain amount of time are retired * using this function.

Used to send the processed message to web socket client at regular intervals of time

Car objects in the list are sorted on the basis of event timestamp

This method: 1. Continiously listens to an input kafka (given by in_kafkaservers and in_kafkatopics) 2. Performs mult

A sanity checker to validate the total occupied and available spots

Store the state of a parking spot Arguments: json_list {[list]} -- List of detections in day2 schema

This method performs the reid for a given set of detection records (all_json_list). It returns re-identified set of json objects.

Used to maintain state of moving cars so that it can be retired * after certain amount of time