# US 12,118,796 B2 · Systems and methods for object detection in an environment > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,118,796 B2 - Document ID: US12118796B2 - Status: Granted - Kind: B2 - Application number: 17/580,023 - Filed: 2022-01-20 - Published: 2023-07-20 - Granted: 2024-10-15 - Earliest claimed: 2022-01-20 - Assignees: Sensormatic Electronics Corporation - Applicants: Sensormatic Electronics, LLC - Inventors: Michael C. Stewart; Natasha Mishell Astudillo; Harsh Gaurangbhai Motka; Joseph Celi - Web page: https://patentgrep.com/patent/US12118796B2 ## CPC - G06V 20/52: Surveillance or monitoring of activities, e.g. for recognising suspicious objects (recognising microscopic objects G06V20/69) - G06V 20/53: {Recognition of crowd images, e.g. recognition of crowd congestion} - G06T 7/97: {Determining parameters from multiple pictures (depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85)} - G06V 10/98: Detection or correction of errors, e.g. by rescanning the pattern or by human intervention; Evaluation of the quality of the acquired patterns - G06V 40/10: Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands - G06T 2207/30242: Counting objects in image - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL ## Abstract Aspects of the present disclosure relate generally to a vision system that detects persons exiting an environment, comprising detecting, by a processor using a plurality of image frames from at least one sensor, persons that exited the environment during a first period of time. The vision system further determines, by the processor, an exit count for the first period of time and retrieves, from a database, historical egress data comprising a detected historic exit count and a corrected historic exit count of the environment for a second period of time corresponding to the first period of time. The vision system calculates an error rate for the second period of time based on a ratio of the detected historic exit count and the corrected historic exit count, determines and stores a corrected exit count for the first period of time by adjusting the exit count using the error rate. ## Claims ### Claim 1 A vision system for detecting persons exiting an environment, comprising: at least one sensor; a memory; and a processor communicatively coupled with the memory and the at least one sensor and configured to: detect, using the at least one sensor, persons that exited the environment during a first period of time; determine an exit count for the first period of time; retrieve, from a database in the memory, historical egress data for a second period of time corresponding to the first period of time, wherein the historical egress data comprises a detected historic exit count and a corrected historic exit count of the environment; calculate an error rate for the second period of time based on a ratio of the detected historic exit count and the corrected historic exit count; determine a corrected exit count for the first period of time by adjusting the exit count using the error rate; and store the corrected exit count in the database. ### Claim 2 (depends on 1) The vision system of claim 1, wherein the processor is further configured to determine the corrected exit count by: determining an occupancy count of the environment during the first period of time, wherein the occupancy count is a difference between a total entry count and the exit count; determining whether the adjusting the exit count using the error rate will cause the occupancy count to become a negative value; and in response to determining that the occupancy count will become a negative value, adjusting the exit count such that the occupancy count becomes zero. ### Claim 3 (depends on 1) The vision system of claim 1, wherein the first period of time and the second period of time are at a same time of day across different days. ### Claim 4 (depends on 3) The vision system of claim 3, wherein the different days are a same day across different weeks. ### Claim 5 (depends on 1) The vision system of claim 1, wherein the processor is further configured to: transmit an alert to a user indicating the corrected exit count. ### Claim 6 A vision system for detecting persons entering an environment, comprising: at least one sensor; a memory; and a processor communicatively coupled with the memory and the at least one sensor and configured to: detect, using a first sensor of the at least one sensor, persons that enter the environment during a period of time by: identifying moving objects in a plurality of image frames captured by the first sensor during the period of time; and determining, for each respective object, a likelihood that the respective object is a person, wherein a likelihood greater than a threshold likelihood is indicative of a person detection; increment an entry count of the period of time for each distinct person detected; determine a corrected entry count for the period of time by, without manual input from a user, removing ghost detection counts and adding missed detection counts; and store the corrected entry count in a database. ### Claim 7 (depends on 6) The vision system of claim 6, wherein the processor is configured to increment the entry count in response to determining that a threshold number of consecutive image frames are associated with likelihoods indicative of at least one person being detected. ### Claim 8 (depends on 7) The vision system of claim 7, wherein the processor is configured to identify a ghost detection by: identifying a set of consecutive image frames that have likelihoods indicative of at least one person and have a frame count equal to or greater than the threshold number of consecutive image frames; determining that an object classified as a person in at least one of the set of consecutive image frames is not a person; updating the frame count of the set of consecutive image frames by removing counts of objects that are not persons; and identifying the ghost detection in response to determining that the updated frame count is less than the threshold number of consecutive image frames. ### Claim 9 (depends on 7) The vision system of claim 7, wherein the processor is configured to identify a missed detection by: identifying a set of consecutive image frames that have likelihoods indicative of at least one person and have a frame count less than the threshold number of consecutive image frames; determining that a person identified in the set of consecutive image frames appears in non-consecutive image frames captured by the first sensor during the period of time; calculating a sum of frame counts between the set of consecutive image frames and the non-consecutive image frames; and identifying the missed detection in response to determining that the sum is not less than the threshold number of consecutive image frames. ### Claim 10 (depends on 6) The vision system of claim 6, wherein the processor is configured to identifying a missed detection by: detecting, by the first sensor, a first amount of persons entering the environment during the period of time; detecting, by a second sensor of the at least one sensor, a second amount of persons in the environment during the period of time; identifying the missed detection in response to determining that the second amount of persons is greater than the first amount of persons. ### Claim 11 (depends on 6) The vision system of claim 6, wherein the processor is further configured to: transmit an alert to the user indicating the corrected entry count. ### Claim 12 (depends on 6) The vision system of claim 6, wherein the processor is further configured to: re-train an object detection model using identified ghost detections and missed detections, wherein the object detection model detects the persons entering the environment. ### Claim 13 A vision system for detecting persons exiting an environment, comprising: at least one sensor; a memory; and a processor communicatively coupled with the memory and the at least one sensor and configured to: detect, using the at least one sensor, persons that enter the environment during a first period of time by: identifying moving objects in a plurality of image frames captured by the at least one sensor during the first period of time; and determining, for each respective object, a likelihood that the respective object is a person, wherein a likelihood greater than a threshold likelihood is indicative of a person detection; increment an entry count of the first period of time for each distinct person detected; determine a corrected entry count for the first period of time by, without manual input from a user, removing ghost detection counts and adding missed detection counts; detect, using the at least one sensor, persons that exited the environment during the first period of time; determine an exit count for the first period of time; retrieve, from a database in the memory, historical egress data for a second period of time corresponding to the first period of time, wherein the historical egress data comprises a detected historic exit count and a corrected historic exit count of the environment; calculate an error rate for the second period of time based on a ratio of the detected historic exit count and the corrected historic exit count; determine a corrected exit count for the first period of time by adjusting the exit count using the error rate; and store the corrected exit count and the corrected entry count in the database. ### Claim 14 A method for detecting persons exiting an environment, comprising: detecting, using at least one sensor, persons that exited the environment during a first period of time; determining an exit count for the first period of time; retrieving, from a database, historical egress data for a second period of time corresponding to the first period of time, wherein the historical egress data comprises a detected historic exit count and a corrected historic exit count of the environment; calculating an error rate for the second period of time based on a ratio of the detected historic exit count and the corrected historic exit count; determining a corrected exit count for the first period of time by adjusting the exit count using the error rate; and storing the corrected exit count in the database. ### Claim 15 (depends on 14) The method of claim 14, wherein determining the corrected exit count further comprises: determining an occupancy count of the environment during the first period of time, wherein the occupancy count is a difference between a total entry count and the exit count; determining whether the adjusting the exit count using the error rate will cause the occupancy count to become a negative value; and in response to determining that the occupancy count will become a negative value, adjusting the exit count such that the occupancy count becomes zero. ### Claim 16 (depends on 14) The method of claim 14, wherein the first period of time and the second period of time are at a same time of day across different days. ### Claim 17 (depends on 16) The method of claim 16, wherein the different days are a same day across different weeks. ### Claim 18 (depends on 14) The method of claim 14, further comprising: transmitting an alert to a user indicating the corrected exit count. ### Claim 19 A method for detecting persons entering an environment, comprising: detecting, using a first sensor of at least one sensor, persons that enter the environment during a period of time by: identifying moving objects in a plurality of image frames captured by the first sensor during the period of time; and determining, for each respective object, a likelihood that the respective object is a person, wherein a likelihood greater than a threshold likelihood is indicative of a person detection; incrementing an entry count of the period of time for each distinct person detected; determining a corrected entry count for the period of time by, without manual input from a user, removing ghost detection counts and adding missed detection counts; and storing the corrected entry count in a database. ### Claim 20 (depends on 19) The method of claim 19, wherein incrementing the entry count is in response to determining that a threshold number of consecutive image frames are associated with likelihoods indicative of at least one person being detected. ### Claim 21 (depends on 20) The method of claim 20, wherein identifying a ghost detection further comprises: identifying a set of consecutive image frames that have likelihoods indicative of at least one person and have a frame count equal to or greater than the threshold number of consecutive image frames; determining that an object classified as a person in at least one of the set of consecutive image frames is not a person; updating the frame count of the set of consecutive image frames by removing counts of objects that are not persons; and identifying the ghost detection in response to determining that the updated frame count is less than the threshold number of consecutive image frames. ### Claim 22 (depends on 20) The method of claim 20, wherein identifying a missed detection further comprises: identifying a set of consecutive image frames that have likelihoods indicative of at least one person and have a frame count less than the threshold number of consecutive image frames; determining that a person identified in the set of consecutive image frames appears in non-consecutive image frames captured by the first sensor during the period of time; calculating a sum of frame counts between the set of consecutive image frames and the non-consecutive image frames; and identifying the missed detection in response to determining that the sum is not less than the threshold number of consecutive image frames. ### Claim 23 (depends on 19) The method of claim 19, wherein identifying a missed detection further comprises: detecting, by the first sensor, a first amount of persons entering the environment during the period of time; detecting, by a second sensor of the at least one sensor, a second amount of persons in the environment during the period of time; identifying the missed detection in response to determining that the second amount of persons is greater than the first amount of persons. ### Claim 24 (depends on 19) The method of claim 19, further comprising: transmitting an alert to the user indicating the corrected entry count. ### Claim 25 (depends on 19) The method of claim 19, further comprising: re-training an object detection model using identified ghost detections and missed detections, wherein the object detection model detects the persons entering the environment.