DESC:
F O R M 2

THE PATENTS ACT, 1970
(39 of 1970)

COMPLETE SPECIFICATION
(See section 10 and rule 13)

TITLE
METHOD FOR CAMERA COVERAGE OPTIMIZATION UNDER ENERGY CONSTRAINTS

APPLICANT
AMRITA VISHWA VIDYAPEETHAM
Amritapuri Campus
Amritapuri, Clappana P. O.
Kollam - 690525 Kerala, India

THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED:
METHOD FOR CAMERA COVERAGE OPTIMIZATION UNDER ENERGY CONSTRAINTS
CROSS-REFERENCES TO RELATED APPLICATION
This application take priority to Provisional Patent Application No. 202441068354 titled “METHOD FOR CAMERA COVERAGE OPTIMIZATION UNDER ENERGY CONSTRAINTS” filed on September 10, 2024.
FIELD OF INVENTION
The present disclosure relates to surveillance network, more particularly, it relates tooptimization of camera coverage in surveillance network.
DESCRIPTION OF THE RELATED ART
The advent of multi-camera surveillance networks has revolutionized the field of security, providing comprehensive spatial coverage and continuous monitoring capabilities. These systems are essential for enhancing safety and security in a wide range of scenarios, from urban surveillance and traffic monitoring to border security and wildlife conservation. However, the deployment and operation of such networks entail significant challenges, primarily related to coverage optimization and energy efficiency
An optimal coverage over a surveillance region is crucial for ensuring effective monitoring over every intended area, minimizing redundancy and gaps. Traditionally, coverage maximization using a minimum number of cameras is the cornerstone while designing a surveillance network. Such networks are hardly adaptable to face any dynamic changes in the environment arising from power fluctuations. Maintaining continuous surveillance in multi-camera networks is costly and environmentally unfriendly, especially when operating on backup power supply.
Efficient utilization of backup power is a key challenge when a surveillance network switches to an Uninterrupted Power Supply (UPS) system during power outages. Instead of operating all the cameras in such energy constrained scenarios, the surveillance system can exhibit graceful degradation by operating with a reduced number of cameras for a prolonged operation. Towards this end, randomly switching off some of the cameras does help to reduce the overall energy consumption in the network. However, randomly switching off some cameras for energy gains is an arbitrary measure and might lead to losing coverage over certain high priority regions instead of low priority regions. Different regions of a surveillance scenario might be associated with varying degrees of priority as specified by the end-user. For example, in a banking surveillance scenario, the regions where the lockers are located will have higher priority due to security reasons.
Various publications have tried to address the problems encountered when maintaining surveillance during power/energy fluctuation. Chinese publication 113890986Adiscloses method for starting of monitoring cameras based on state of power supply. high-reflectivity surface dynamic three-dimensional measurement method. mechanisms to manage permissions to access user data in a distributed ledger trust network. US publication2021195096A1discloses an assessment system for prioritization among cameras of a multi-camera arrangement. Liet. al in “Using Evolutionary Approaches to Manage Surveillance Cameras in Dynamic Environments” discusses identifying efficient surveillance configurations. In “Adaptive Monitoring Relevance in Camera Networks for Critical Surveillance Applications”, Costaet. al. mentions dynamically assigning relevancies to cameras that view the area of critical events employing scalar sensors and a decentralized decision mechanism.
Presently, there is a requirement of adaptively optimizing camera coverage while operating under the constraint of depleting or limited energy availability and considering the priority of region covered.
SUMMARY OF THE INVENTION
The present subject matter relates to optimizing critical surveillance coverage in energy constrained conditions.
In one embodiment of the present subject matter, the method comprises obtaining location and orientation of a plurality of cameras N placed in a surveillance area, segmenting the surveillance area into plurality of regions (Rh, Rm, Rn) with each region being assigned a priority of high Ph, medium Pm or normal Pn, determining a score sc for each of the cameras in the surveillance area and generating a first list SC with the plurality of cameras N and their corresponding score sc. The method further comprises sorting the first list SC to obtain the cameras ci with the highest scores sci, generating a second list of ranked cameras RC having the camera ci with the highest score sci, updating the priorities of the region covered by the camera ci with the highest score sci, evaluating percentage of high ACh and medium ACm priority areas covered by the camera with the highest score sci and resetting the priority of the high priority and medium areas to normal if the percentage covered for high ACh and medium ACm priority areas individually is above 50%. Further, the method comprises ranking the remaining cameras in the first list SC after re-calculating the score for the remaining cameras, updating the score scjof all non-ranked cameras cjbased on the extent of their coverage overlap with the coverage of ranked cameras ciand generating a rank for the plurality of cameras N present in the surveillance area.
In various embodiments, the sorting of the first list is performed in descending order.
In various embodiments, all the non-ranked cameras are subject to a penalty based on the extent of their coverage overlap with ranked cameras.
In various embodiments, the orientations of the plurality of cameras N is ?1, ?2…. ?N.
In various embodiments, score sc for each of the cameras is sum of priorities of the surveillance area covered by the camera.
This and other aspects are described herein.

BRIEF DESCRIPTION OF THE DRAWINGS
The invention has other advantages and features, which will be more readily apparent from the following detailed description of the invention and the appended claims, when taken in conjunction with the accompanying drawings, in which:
FIG. 1 illustrates method for optimizing critical surveillance coverage, according to an embodiment of the present subject matter.
FIG. 2A, 2B and 2C illustrate initial scoring of the cameras, rescoring after first camera ranking and second camera ranking respectively, according to an embodiment of the present subject matter.
FIG. 3 illustrates complexity analysis of the method for optimizing critical surveillance coverage, according to an embodiment of the present subject matter.
FIG. 4A, 4B and 4C illustrate plan of office space with camera locations, existing surveillance network and priority map respectively for implementation in small indoor surveillance regions, according to an embodiment of the present subject matter.
FIG. 5 illustrates ranking sequence of cameras in small indoor surveillance regions, according to an embodiment of the present subject matter.
FIG. 6Aand 6B illustrate region wise coverage results and coverage results after iterative removal of least ranked cameras respectively, according to an embodiment of the present subject matter.
FIG. 7A, 7B and 7C illustrate plan of office space with camera locations, existing surveillance network and priority map respectively for implementation in large indoor surveillance regions, according to an embodiment of the present subject matter.
FIG. 8A and 8B illustrate region wise coverage results and coverage results after iterative removal of least ranked camerasrespectively, according to an embodiment of the present subject matter.
FIG. 9A, 9B and 9C illustrate plan of crossroad junction with camera locations, existing surveillance network and priority map respectively for implementation in outdoor surveillance, according to an embodiment of the present subject matter
FIG. 10 illustrates ranking sequence of cameras in outdoor surveillance, according to an embodiment of the present subject matter.
FIG. 11A and 11B illustrate region wise coverage results and coverage results after iterative removal of least ranked camerasrespectively, according to an embodiment of the present subject matter.
Referring to the figures, like numbers indicate like parts throughout the various views.
DETAILED DESCRIPTION OF THE EMBODIMENTS
While the invention has been disclosed with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the invention. In addition, many modifications may be made to adapt to a particular situation or material to the teachings of the invention without departing from its scope.
Throughout the specification and claims, the following terms take the meanings explicitly associated herein unless the context clearly dictates otherwise. The meaning of “a”, “an”, and “the” include plural references. The meaning of “in” includes “in” and “on.” Referring to the drawings, like numbers indicate like parts throughout the views. Additionally, a reference to the singular includes a reference to the plural unless otherwise stated or inconsistent with the disclosure herein.
The present subject matter describesa method for adaptive coverage optimization of energy-efficient multi-camera surveillance networks during fluctuations in energy availability. It is crucial to ensure that security systems remain operational and effective even when regular power sources are unavailable, in such scenarios there may be a need to curtail the number of cameras in the surveillance network to save power. In such cases it is essential to rank the cameras over a surveillance network based on their area of coverage, priority of the covered region, and extent of overlap with neighbouring cameras. The method ranks the cameras over the surveillance region to address adaptive camera optimization in surveillance networks.
A method for optimizing critical surveillance coverage in energy constrained conditions is illustrated in FIG. 1. The method 100 is based on area of camera coverage, priority of the covered region, and extent of overlap with neighboring cameras. In step 102 the location and orientation of a plurality of cameras N placed in a surveillance area is obtained, wherein C = c1, c2, . . . ,cN configured at orientations {?1, ?2, . . . , ?N}. In step 104 the entire surveillance area is divided in plurality of regions (Rh, Rm, Rn) and each region is assigned a priority of high Ph, medium Pm or normal Pn, whereinPh>Pm>Pn and Rhindicates high priority region, Rm indicates medium priority region while Rn indicates normal priority region. A score sc for each of the cameras in the surveillance area is determined as provided in step 106, wherein each camera ci, is configured at an orientation ?j. The score sc is the sum of priorities of the entire region Ri with an area Ai covered by camera ci.

In step 108, a first list SC is generated with the cameras ciand their corresponding score sci.

Further, in step 110 the first list SC is sorted to obtain the cameras ci with the highest scores sci, wherein the first list SCmaybe sorted in descending order. A second list of ranked cameras RC is generated with camera cihaving the highest score sci in the first list SCas mentioned in step 112.

Updating the priorities of region Ri covered by camera ci to zero, and the score sciassociated with ci is also updated to zero as provided in step 114.

In step 116, the percentage of area covered in the high-priority region denoted as ACh, and the percentage of area covered in the medium-priority region denoted as ACm by the camera ci? RC is evaluated. If the percentage of area covered over both of these regions individually is above 50%, the priority of the respective region is updated to normal priority Pn as provided in step 118.

Wherein, Ah is the total area of high-priority region Rh and Am is the total area of medium-priority region Rm. For every iteration, a rank is assigned to one of the cameras ci with the highest score as mentioned in step 120. Further, in step 122 the score of all non-ranked cameras cj RC is updated based on the extent of its coverage overlap with the coverage of ranked cameras. All the non-ranked cameras are subject to a penalty based on their extent of coverage overlap. The updated score scof a non-ranked camera cj is calculated as:

Where, OAj is the intersection of the region covered by camera cj with the regions covered by all the ranked cameras. The score scj is recalculated for all non-ranked cameras, while keeping the score of all ranked cameras at zero. Thereafter, the first list SC is sorted to find the next camera with the highest score to append to the second list of ranked cameras RC until all the N cameras in the surveillance region are assigned with a rank as provided in step 124.
In various embodiments, FIG. 2A illustrates a surveillance region R with three cameras c1, c2, c3, and the boxes inside R indicate the priorities within the region, whereinwhere 3 signifies high priority, 2 represents medium priority, and 1 indicates normal priority. The score for all three cameras is determined and camera c1 gets the highest score as it covers most of the high priority regions.After calculating the scores for all the cameras, the order of the cameras in the list SC, arranged according to their scoreswill be {c1, c2, c3}. The highest score, camera c1 is initially inserted to second list of ranked cameras RC,resulting in RC = {c1}. Subsequently, the priorities of the region covered by camera c1 is updated to zero as shown in FIG. 2B with the score associated with camera c1 being reassigned to zero. It is determined whether the ranked camerac1 covers more than 50% of priority regions Rh or Rm. Asthe ranked camera c1 covers more than 50% of Rh, hence the priority for the remaining Rhis assigned with normal priority Pn= 1, as illustrated in FIG. 2B. Further, the score for the remaining non-ranked cameras c2 andc3is recalculated considering the priority of the region they cover and the extent of overlap with the ranked camera c1. FIG. 2B shows the camera c3covers a larger area of the medium priority region Rm and hasless overlap with the previously ranked camera c1, compared toc2.Based on the initial scoring the camera c2 obtained a higher score than c3, whereas after updating the scores, the camera c3obtained a higher score than c2 and hence secured the rank two as illustrated in FIG. 2C. Therefore, the final ranked list generated is RC = {c1, c2, c3}.
In various embodiments, FIG. 3 illustrates the complexity analysis of the method 100 based on the time required for ranking cameras with the number of cameras ranging from 1 to 16. The complexity analysis of the method 100 indicate its computational demand as thenumber of cameras increases. The methodincludes two internal loops as part of execution to iterate over variouscamera orientations and locations. Eachiteration involves sorting the list of cameras, which has acomplexity of O(NlogN). Since cameras are ranked oneat a time and their scores updated, this action repeats Ntimes, leading to a complexity of O(N2logN) for the entireranking process. The increasing trend ofthe elapsed time indicates that the method’s computational cost grows with the number of cameras. This growth appearsto be more than linear, indicating a possibility of polynomialcomplexity.
A method for optimizing critical surveillance coverage in energy constrained conditions has several advantages over prior art particularly for maintaining coverage in critical and attain optimum usage of the power available. Energy efficiency in surveillance networks during power outages is driven by the critical need to prolong the monitoring in the absence of a reliable power supply. The method ranks all the cameras ina surveillance network in terms of their relative importance based on their area of coverage, priority of the covered region, and extent of overlap with neighboring cameras. Using this method various strategies can addressed that arise during power shortage such as reducing the power consumption to half, selecting the best k cameras in the case of critical power shortage, selecting a minimal number of cameras to attain a coverage threshold CT and performing sequential elimination of the least significant cameras. Additionally, the method is scalable and adaptable to various surveillance scenarios, making it a versatile solution for diverse settings. Further, the method is cost-effective and computationally efficient which includes extended operational lifespan of surveillance systems, reduced maintenance costs, and improved adaptability to varying surveillance requirements. This facilitates in balancing the need for vigilant monitoring with the imperative of energy conservation.
Although the detailed description contains many specifics, these should not be construed as limiting the scope of the invention but merely as illustrating different examples and aspects of the invention. It should be appreciated that the scope of the invention includes other embodiments not discussed herein. Various other modifications, changes and variations which will be apparent to those skilled in the art may be made in the arrangement, operation and details of the system and method of the present invention disclosed herein without departing from the spirit and scope of the invention as described here, and as delineated in the claims appended hereto.
EXAMPLES
All indoor experiments are based on the specifications of the ‘AXIS P1357’ static network cameras which include a 1/3.2” RGB CMOS image sensor, varifocal lens (2.8 – 8 mm), and a horizontal field of view (92° to 32°). The working range of this camera is fixed to 630 cm and the horizontal field of view is set at 54°. For outdoors experiments, the specifications of the DS-2CE16C2T-IR camera for the crossroad junction scenario was used.
EXAMPLE 1: Implementation in Indoor Scenarios: Small Surveillance Regions
Floor plan of a 3800 square feet office building with an existing surveillance network with 27 AXIS P1357 static cameras which provide coverage across 94% of the total area was considered. FIG. 4A shows the plan image of the office building with installed cameras, FIG. 4B shows the existing surveillance over the building, and FIG. 4C represents the priority map of the office building. The red regions in the priority map indicate the high-priority regions, here in this case one document storage room and the areas near the main entrance door are considered as high-priority regions (Rh). The blue regions in the priority map such as the corridors indicate the medium-priority regions (Rm). The green regions in the priority map individual staff spaces are considered normal-priority regions (Rn).
The ranking sequence of the method over the floor plan of the office building is illustrated in FIG. 5, wherein the cameras are turned on based on their rank. The method initially ranks the cameras that cover the high priority region Rh and avoids prioritizing cameras that overlap with the ranked cameras even though it covers the high priority or medium priority regions. Thus, the first camera to be turned on covers a majority of the high-priority region, followed by second-ranked camera which covers the next available high priority region (near the main entrance door), which does not overlap with the coverage of the first selected camera. The method avoided cameras that covers the high priority region which has an overlap with the first ranked camera. Specifically in this example, the camera which is covering the high priority region and has an overlap with the first-ranked camera secured a rank of 20. The method assigns a higher rank to the cameras that not only cover high-priority regions but also minimize overlaps. The first 13 ranked cameras cover unique areas in the surveillance region without any coverage overlap. If there arises a need for energy saving and reducing the number of cameras to half, the system can still attain a coverage of 65% with the first 13 ranked cameras.
FIG. 6A shows the coverage attained over various priority regions while turning on high-ranked cameras one at a time. It can be observed that the top-ranked cameras maximize coverage over high and medium-priority region. The method focuses on normal-priority regions after attaining sufficient coverage over the high and medium-priority regions. TABLE 1 shows the coverage over regions Rh, Rm, and Rnwhile placing the first ranked camera, top-ranked ?N/(4 )? cameras, top-ranked ?N/(2 )?, top-ranked ?3N/(4 )?, and all N cameras. Here N is the total number of cameras over the surveillance region, which in this case is 27.
TABLE 1: Coverage results over various priority regions of the office building while the top N ranked cameras are turned on

FIG. 6B shows the coverage results over the surveillance region after removing the least priority cameras from the surveillance network iteratively one at a time. It can be visualized that initially switching off these least priority cameras does not reduce the total coverage drastically, and instead, reduces the coverage overlap over the surveillance region. This indicates that the method correctly identifies the least significant cameras in the surveillance region and assigns them with the lowest-rank.
EXAMPLE 2: Implementation in Indoor Scenarios: Large Surveillance Regions
Floor plan of a 7800 square feet university building with an existing surveillance network containing 46 ‘AXIS P1357’ static cameras, which covers 84% of the total area is considered. FIG. 7A shows the plan image of the university building with installed cameras, where the cameras are highlighted in red. FIG. 7B shows the existing surveillance over the building, where blue triangles represent the coverage of cameras and FIG. 7C represents the priority map of the office building. Here, two rooms in the university building are highlighted as high-priority regions in the priority map, corridors are considered as medium-priority regions, and all other areas are considered as normal-priority regions.
FIG. 8A shows the coverage attained over various priority regions while turning on high-ranked cameras one at a time. It can be observed that the 8 top-ranked cameras focus on maximizing coverage over the high-priority region and the next 8 ranked cameras focus on maximizing coverage over the medium-priority region. The method focuses on normal priority regions after attaining sufficient coverage over the high and medium-priority regions. TABLE 2 shows the coverage over regions Rh, Rm, and Rn while placing the first ranked camera, top-ranked ?N/(4 )? cameras, top-ranked ?N/(2 )?, top-ranked ?3N/(4 )?, and all N cameras. Here N refers to the total number of cameras over the surveillance region, which in this case is 46.
TABLE 2: Coverage results over various priority regions of university building while the top N ranked cameras are turned on

FIG. 8B shows the coverage results over the surveillance region after removing the least priority cameras from the surveillance network iteratively one at a time. Here, it can be visualized that initially switching off these least priority cameras does not reduce the total coverage drastically, and instead, reduces the coverage overlap over the surveillance region. There is no coverage overlap between the 8 top-ranked cameras and only 1% coverage overlap between the next 8 ranked cameras.

EXAMPLE 3: Implementation in Outdoor Scenarios: Crossroad Junction
A 500 square meter crossroad junction with an existing surveillance network containing 8 ‘DS-2CE16C2TIR’ static cameras, which covers 93% of the total area. FIG. 9A shows the map image of the crossroad junction with installed cameras, FIG.9B shows the existing surveillance over the building and FIG.9C represents the priority map of the office building. Here, the center portion of the crossroad is considered ashigh-priority regions as these regions are accident-prone areaand need to be monitored for traffic rule violations. The areas adjacent to center portion of the crossroad are consideredmedium-priority regions, and all other areas are consideredas normal-priority regions.
The ranking sequence of the cameras using the method over the cross-road junction is illustrated in FIG. 10, where the cameras are turned on based on their rank. Initially, the method ranks the camera covering the largest portion of the high priority region Rh. The second-ranked camera spans a smaller segment of Rhand a substantial part of the medium priority region Rm. The cameras ranked three and four focus primarily on the medium priority regions compared to other non-ranked cameras. At this stage, both high priority and medium priority regions achieve coverage exceeding 50%, leading to an update in their associated priorities to normal priority. Consequently, the fifth to eighth camera focuses on the normal priority regions. The first four cameras collectively achieve 68% coverage, indicating that even if the system is scaled down to half its capacity for energy savings, it can still maintain coverage of more than half of the area. The method prioritizes cameras that cover high priority regions effectively while also minimizing overlaps.
FIG. 11A represents the coverage achieved across various priority regions by sequentially activating high ranked cameras. Initially, the first camera maximizes coverage over the high priority area, covering 73% of the critical region. By the time the third camera is activated, it manages to cover 56.15% of the medium priority regions. Early selections of cameras predominantly focus on these high and medium priority re- gions. Subsequently, the method shifts its emphasis towards cameras that enhance coverage over areas of normal priority.TABLE 3 shows the coverage over regions Rh, Rm, and Rn, while placing the first ranked camera, top-ranked ?N/(4 )? cameras, top-ranked ?N/(2 )?, top-ranked ?3N/(4 )?, and all N cameras top-ranked. Here N is the total number of cameras over the surveillance region, which in this case is 8.
TABLE 3: Coverage results over various priority regions ofthe crossroad junction while the top N ranked cameras areturned on

FIG. 11B shows the coverage results over the surveillance region after removing the least priority cameras from thesurveillance network iteratively one at a time. Here, it can be visualized that initially switching off these least prioritycameras does not reduce the total coverage drastically, andinstead, reduces the coverage overlap over the surveillanceregion. Especially there is no coverage overlap between the 8 top-ranked cameras and only 1% coverage overlap betweenthe next 8 ranked cameras.
,CLAIMS:WE CLAIM:
1. A method (100) for optimizing critical surveillance coverage in energy constrained conditions, the method comprising:
obtaining (102) location and orientation of a plurality of cameras N placed in a surveillance area;
segmenting (104) the surveillance area into plurality of regions (Rh, Rm, Rn) with each region being assigned a priority of high Ph, medium Pm or normal Pn;
determining (106) a score sc for each of the cameras in the surveillance area;
generating (108) a first list SC with the plurality of cameras N and their corresponding score sc;
sorting (110) the first list SCto obtain the cameras ci with the highest scores sci;
generating (112) a second list of ranked cameras RC having the camera ci with the highest score sci;
updating(114) the priorities of the region covered by the camera ci with the highest scoresci;
evaluating(116) percentage of high ACh and medium ACm priority areas covered by the camera with the highest scoresci;
resetting(118) the priority of the high priority and medium areas to normal if the percentage covered for high ACh and medium ACm priority areasindividually is above 50%;
ranking (120) the remaining cameras in the first list SC after re-calculating the score for the remaining cameras;
updating (122) the score scj of all non-ranked cameras cj based on the extent of their coverage overlap with the coverage of rankedcamerasci; and
generating (124) a rank for the plurality of cameras N present in the surveillance area.

2. The method (100) as claimed in claim 1, wherein the sorting of the first list is performed in descending order.

3. The method (100) as claimed in claim 1, wherein all the non-ranked cameras are subject to a penaltybased on the extent of their coverage overlap with ranked cameras.

4. The method (100) as claimed in claim 1, wherein the orientations of the plurality of cameras N is ?1, ?2….?N.

5. The method (100) as claimed in claim 1, wherein score sc for each of the cameras is sum of priorities of the surveillance area covered by the camera.

Dr V. SHANKAR
IN/PA-1733
For and on behalf of the Applicants