Patent · US11113833B2 · B2 · US
Object detection system
- (11) Publication number
- US11113833B2
- (21) Application number
- 16/493,121
- (22) Filing date
- 2018-03-05
- (30) Priority date
- 2017-03-14
- (43) Publication date
- 2021-09-07
- (45) Date of grant
- 2021-09-07
- (51) IPC
- G01S 17/04; G01S 17/89; G06T 7/246; G06T 7/521; G06T 7/579
- (52) CPC
- G06T Image data processing or generation, in general: 7/579, 2207/10028, 2207/30232, 7/194, 7/246, 7/254, 7/521
- G01S Radio direction-finding; radio navigation; determining distance or velocity by use of radio waves; locating or presence-detecting by use of the reflection or reradiation of radio waves; analogous arrangements using other waves: 17/04, 17/50, 17/89, 7/4802
- (73) Assignee
- Konica Minolta Inc
- (72) Inventors
- Shunsuke TAKAMURA
- (54) Title
- Object detection system
- (57) Abstract
An object detection system includes a depth image detector and a moving object extractor. The depth image detector detects a depth image from an external environment. The moving object extractor extracts a moving object desired to be extracted from the depth image. The moving object extractor registers in advance the depth image in a memory as a background while the moving object to be extracted does not exist, and extracts only a pixel whose current depth is present on a nearer side than a depth of the background as a candidate for a pixel corresponding to the moving object to be extracted.
- Full text
- View on Google Patents
Claims (2)
- An object detection system comprising: a depth image detector that detects a depth image from an external environment; and a moving object extractor that extracts a moving object desired to be extracted from the depth image, wherein the moving object extractor registers in advance the depth image in a memory as a background while the moving object to be extracted does not exist, and extracts only a pixel whose current depth is present on a nearer side than a depth of the background as a candidate for a pixel corresponding to the moving object to be extracted; wherein the moving object extractor further makes a determination to extract, as the moving object to be extracted, a pixel from a pixel group composed of pixels of the candidates having been extracted, and for a portion not extracted as the moving object to be extracted, updates and registers a depth of the pixel in the memory as the depth of the background; and wherein the moving object extractor determines a size of the pixel group, compares the size of the pixel group to a threshold value, and when the size of the pixel group exceeds the threshold value the pixel group is registered as background.
- The object detection system according to claim 1, wherein when determining whether a pixel is present on a nearer side than the background registered in the memory, the moving object extractor sets a minimum threshold value of a changing depth, and extracts only a pixel present on a nearer side than the depth of the background by more than or equal to the threshold value as a candidate for a pixel corresponding to the moving object to be extracted.
Description
The present invention relates to an object detection system.
Recently, an object detection system that, for the purpose of detecting an intruder or a vehicle in a monitoring space, detects an object in a depth image detected by a laser radar or the like is used. Here, the depth image is obtained by mapping the value of a depth to the object as two-dimensional coordinates.
As a method of extracting pixels including a moving object from a depth image, a method of extracting pixels different in depth from a previously acquired background (a depth image while the moving object does not exist) is known as described in Patent Document 1 as well. A background depth map generator in Patent Document 1 generates a background depth map indicating information about a depth from a laser radar in a detection region to an object present in the background on the basis of a measurement result of the laser radar. A moving object detector detects a moving object in the detection region from a difference between the depth information in the background depth map stored in advance in a background depth map memory and the measurement result of the laser radar.
Patent Document 1: JP 2005-300259 A
However, the above-described method has a problem in that even an essentially unnecessary pixel is extracted in cases as will be described below.
Citations (11)
- JP2004191095A
- JP2005300259A
- JP2006064695A
- EP1772752A1
- JP2008275442A
- JP2009085927A
- US20110164185A1
- JP2011185664A
- WO2016013719A1
- US20170206423A1
- US20160292522A1
Record as JSON
{
"publication_number": "US11113833B2",
"country": "US",
"kind": "B2",
"title": "Object detection system",
"abstract": "An object detection system includes a depth image detector and a moving object extractor. The depth image detector detects a depth image from an external environment. The moving object extractor extracts a moving object desired to be extracted from the depth image. The moving object extractor registers in advance the depth image in a memory as a background while the moving object to be extracted does not exist, and extracts only a pixel whose current depth is present on a nearer side than a depth of the background as a candidate for a pixel corresponding to the moving object to be extracted.",
"claims": [
"1. An object detection system comprising: a depth image detector that detects a depth image from an external environment; and a moving object extractor that extracts a moving object desired to be extracted from the depth image, wherein the moving object extractor registers in advance the depth image in a memory as a background while the moving object to be extracted does not exist, and extracts only a pixel whose current depth is present on a nearer side than a depth of the background as a candidate for a pixel corresponding to the moving object to be extracted; wherein the moving object extractor further makes a determination to extract, as the moving object to be extracted, a pixel from a pixel group composed of pixels of the candidates having been extracted, and for a portion not extracted as the moving object to be extracted, updates and registers a depth of the pixel in the memory as the depth of the background; and wherein the moving object extractor determines a size of the pixel group, compares the size of the pixel group to a threshold value, and when the size of the pixel group exceeds the threshold value the pixel group is registered as background.",
"2. The object detection system according to claim 1, wherein when determining whether a pixel is present on a nearer side than the background registered in the memory, the moving object extractor sets a minimum threshold value of a changing depth, and extracts only a pixel present on a nearer side than the depth of the background by more than or equal to the threshold value as a candidate for a pixel corresponding to the moving object to be extracted."
],
"description_excerpt": "The present invention relates to an object detection system.\n\nRecently, an object detection system that, for the purpose of detecting an intruder or a vehicle in a monitoring space, detects an object in a depth image detected by a laser radar or the like is used. Here, the depth image is obtained by mapping the value of a depth to the object as two-dimensional coordinates.\n\nAs a method of extracting pixels including a moving object from a depth image, a method of extracting pixels different in depth from a previously acquired background (a depth image while the moving object does not exist) is known as described in Patent Document 1 as well. A background depth map generator in Patent Document 1 generates a background depth map indicating information about a depth from a laser radar in a detection region to an object present in the background on the basis of a measurement result of the laser radar. A moving object detector detects a moving object in the detection region from a difference between the depth information in the background depth map stored in advance in a background depth map memory and the measurement result of the laser radar.\n\nPatent Document 1: JP 2005-300259 A\n\nHowever, the above-described method has a problem in that even an essentially unnecessary pixel is extracted in cases as will be described below.",
"cpc": [
"G06T 7/579",
"G01S 17/04",
"G01S 17/50",
"G01S 17/89",
"G01S 7/4802",
"G06T 2207/10028",
"G06T 2207/30232",
"G06T 7/194",
"G06T 7/246",
"G06T 7/254",
"G06T 7/521"
],
"ipc": [
"G01S 17/04",
"G01S 17/89",
"G06T 7/246",
"G06T 7/521",
"G06T 7/579"
],
"assignees": [
"Konica Minolta Inc"
],
"inventors": [
"Shunsuke TAKAMURA"
],
"filing_date": "2018-03-05",
"publication_date": "2021-09-07",
"grant_date": "2021-09-07",
"priority_date": "2017-03-14",
"application_number": "US-201816493121-A",
"family_id": "63523303",
"cited_by_count": 1,
"citations": [
"JP2004191095A",
"JP2005300259A",
"JP2006064695A",
"EP1772752A1",
"JP2008275442A",
"JP2009085927A",
"US20110164185A1",
"JP2011185664A",
"WO2016013719A1",
"US20170206423A1",
"US20160292522A1"
]
}
Record 1,494 of 8,000 in Patents full text (MLC-0201). Request the full dataset.