TWI750673B - System of estimating ambient light - Google Patents

Abstract

A system of estimating ambient light includes an image sensor that captures an image; a region of interest (ROI) selector that determines at least one ROI on the image; an occupancy detector that determines existence status of an object disposed on the at least one ROI; and an ambient light estimator that estimates illumination of ambient light of the at least one ROI according to luminance of the at least one ROI on the image.

Description

Ambient Light Estimation System

The present invention relates to lighting, and more particularly to an ambient light estimation system.

The smart parking lot uses an occupancy sensor to sense the occupancy status of the parking space, which provides great convenience for the driver. However, the artificial lighting in conventional parking lots is adjusted manually or at automatic timing according to a prior plan. Therefore, energy consumption cannot be greatly reduced, and artificial lighting cannot be dynamically adjusted according to the parking status of vehicles in the parking lot.

Although some parking lots use ambient light sensors, they actually sense the light reflected from the vehicle, not the real ambient light of the parking space. For example, even if the ambient lighting remains the same, the ambient light sensor may sense too much light reflected from a white vehicle and too little light from a black vehicle.

Therefore, it is urgent to propose a novel mechanism to overcome many shortcomings of traditional parking lots.

In view of the above, one of the objectives of the embodiments of the present invention is to provide an ambient light estimation system, which is suitable for parking lot lighting to save a lot of energy and effectively enhance driving and personal safety.

According to an embodiment of the present invention, an ambient light estimation system includes an image sensor, a region of interest selector, an occupancy detector, and an ambient light estimator. The image sensor captures an image. The region of interest selector determines at least one region of interest in the image. Occupancy detectors determine the presence of objects in the area of interest. The ambient light estimator estimates the illuminance of ambient light in the region of interest based on the brightness of the region of interest in the image.

Figure 1 A shows a block diagram of an ambient light estimation system 100 according to an embodiment of the present invention, which is suitable for parking lot lighting. The parking lot lighting is only an example, and the ambient light estimation system (hereinafter referred to as the system) 100 can also be applied to other applications.

In this embodiment, the system 100 may include an image sensor 11 whose field of view (FOV) can capture images of a plurality of (consecutive) parking spaces of a (outdoor or indoor) parking lot. In a preferred embodiment, the image sensor 11 is a visible light image sensor, such as a red, green and blue image sensor or a monochrome image sensor. The system 100 can include a region of interest (ROI) selector 12 for determining at least one region of interest of the image.

The second figure illustrates the image sensor 11 and the light source 21, which are disposed above the two parking spaces 22 of the parking lot. Incident light (or ambient light) 23 from the light source 21 illuminates the surface of the parking grid 22 . Part of the incident light 23 is reflected by the parking grid 22 (referred to as reflected light 24 ) and captured by the image sensor 11 , and other incident light 23 is absorbed by the parking grid 22 . The illumination of the reflected light 24, the relative luminance of the captured image, the exposure time and gain of the image sensor 11, and the reflectance of the parking grid 22 have the following relationships (1) to (2) : Reflected light∝(relative brightness)/(exposure time x gain) (1) Reflected light∝(incident light x reflectance) (2) where ∝ represents proportional to.

Therefore, the illuminance of the incident light 23 can be expressed by the following formula (3): Incident light = Kx relative brightness / (exposure time x gain x reflectivity) (3) where K represents the proportionality constant.

The system 100 of this embodiment may include a calibration device 10 , which receives the image of the idle parking bay 22 captured by the image sensor 11 , and determines the proportionality constant K accordingly. The third figure shows a schematic diagram of the configuration of determining the proportional constant K, which includes an image sensor 11 , a light source 21 and a light meter 31 , which are arranged above the empty parking space 22 . The photometer 31 can measure the illuminance of the incident light 23 . According to the relative brightness of the captured image, the exposure time and gain of the image sensor 11 and the reflectivity of the (empty) parking grid 22, the proportionality constant K can be obtained from equation (3). In another embodiment, if the reflectivity of the (empty) parking bay 22 is unknown, then (K/reflectance) is determined instead.

The system 100 of this embodiment may include an occupancy detector 13 for determining the presence status of objects in the region of interest. In this embodiment, as shown in the detailed block diagram of the occupancy detector 13 in FIG. 1 B, the occupancy detector 13 may include a motion detector 131 for detecting moving vehicles. In one embodiment, the motion detector 131 can detect the moving vehicle by comparing the current image (captured by the image sensor 11 ) with the previous image. When the difference between the current image and the previous image is greater than a preset threshold, a moving vehicle is detected. In this embodiment, the motion detector 131 can perform motion detection on the image captured by the image sensor 11 . In another alternative embodiment, passive infrared (PIR) sensors or ultrasonic sensors (not shown) may be used to detect moving vehicles.

The occupancy detector 13 of the present embodiment may include a state detector 132 that is triggered by the motion detector 131 to determine the presence state (ie, empty or occupied) of the parking bay 22 . The state detector 132 may perform image processing on the image captured by the image sensor 11 to perform detection. In one embodiment, the state detector 132 may use feature-based object detection, such as histogram of oriented gradient (HOG) or scale-invariant feature transformation (SIFT) ). In another embodiment, the state detector 132 may use a neural network, such as a convolutional neural network (CNN). In an alternative embodiment, occupancy detector 132 may use non-image-based detection, such as ultrasonic distance measurement or earth induction.

In this embodiment, the occupancy detector 13 may include a reflectivity estimator 133 , and when the state detector 132 detects the occupancy state, the reflectivity estimator 133 is used to estimate the reflectivity of the occupied parking space 22 . According to Equation (3), when the incident light of the light source 21 and the exposure time and gain of the image sensor 11 remain the same, the ratio of the relative brightness to the reflectivity of the empty parking space will be the same as the relative brightness and reflectance of the occupied parking space. The ratio of , as shown in formula (4): (relative brightness/reflectivity) _empty = (relative brightness/reflectivity) _occupied (4)

Therefore, the reflectivity of the occupied parking space can be estimated according to the relative brightness of the empty parking space, the relative luminance of the occupied parking space, and the reflectivity of the empty parking space.

In this embodiment, the system 100 may include an ambient light estimator 14, which receives the image captured by the image sensor 11, and estimates the area of interest (ie, an empty or occupied parking space) according to the brightness of the area of interest in the image. The illuminance of the ambient light (or incident light 23). As shown in Equation (3), the illuminance of ambient light (or incident light) can be estimated according to the relative brightness of the captured image, the exposure time, gain of the image sensor 11 and the reflectivity of the (empty or occupied) parking space . The fourth illustration shows the estimation of ambient light within the estimation range 41 (shown in dashed lines) of each parking slot 22 . In one embodiment, when estimating the ambient light, the specular reflection area 42 whose brightness value is greater than a predetermined threshold value is excluded. Therefore, as illustrated in the fourth figure, a reduced estimation range 43 is used to estimate ambient light. In another embodiment, a median filter may be performed on the captured image to remove the specular region 42 .

According to the above embodiment, the artificial lighting provided in the parking lot can be adjusted according to the estimated ambient light. For example, if the estimated ambient light is high, reduce artificial lighting to save energy; if the estimated ambient light is low, increase artificial lighting to enhance driving and personal safety. Additionally, a small estimated ambient light may also indicate that the artificial lighting has aged or failed, requiring immediate repair or replacement. The fifth figure illustrates the illuminance timing diagram of the artificial lighting of the parking lot. Before the vehicle enters, the illuminance of artificial lighting is adjusted according to the estimated ambient light. When (the motion detector 131 and the state detector 132 ) detect the vehicle entering, the illumination of artificial lighting is increased to enhance driving safety. When the vehicle is parked, reduce the illuminance of artificial lighting to the original level to save energy.

The above descriptions are only preferred embodiments of the present invention and are not intended to limit the scope of the patent application of the present invention; all other equivalent changes or modifications made without departing from the spirit disclosed in the invention shall be included in the following within the scope of the patent application.

100: Ambient Light Estimation System 10: Correction device 11: Image sensor 12: Region of interest selector 13: Occupancy Detector 131: Motion Detector 132: Status Detector 133: Reflectivity Estimator 14: Ambient Light Estimator 21: Light source 22: Parking grid 23: Incident light 24: Reflected Light 31: Photometer 41: Estimated range 42: Specular reflection area 43: Narrow down your estimates

The first figure A shows a block diagram of an ambient light estimation system according to an embodiment of the present invention, which is suitable for parking lot lighting. Figure 1 B shows a detailed block diagram of the occupancy detector of Figure 1 A. The second figure illustrates the image sensor and the light source, which are arranged above the two parking spaces of the parking lot. The third figure shows a schematic diagram of the configuration for determining the proportionality constant, which includes an image sensor, a light source and a light meter, and is arranged above the empty parking space. The fourth legend shows the estimation of ambient light within the estimated range of each parking space. The fifth figure illustrates the illuminance timing diagram of the artificial lighting of the parking lot.

100: Ambient Light Estimation System 10: Correction device 11: Image sensor 12: Region of interest selector 13: Occupancy Detector 14: Ambient Light Estimator

Claims (9)

An ambient light estimation system, comprising: an image sensor for capturing an image; a region of interest selector for determining at least one region of interest in the image; an occupancy detector for determining the size of an object in the region of interest Existence state; an ambient light estimator for estimating the illuminance of ambient light in the region of interest according to the brightness of the region of interest in the image; and a light source for lighting above the region of interest, the region of interest The reflectivity of the light source changes according to the existence state of the object, thereby affecting the illuminance of the ambient light, and the light source adjusts the illumination of the light source according to the estimated illuminance of the ambient light. The ambient light estimation system of claim 1, wherein the illuminance of the ambient light illuminating the object is a function of the relative brightness of the image, the exposure time and gain of the image sensor, and the reflectivity of the region of interest, expressed as follows: Incident light = Kx relative brightness / (exposure time x gain x reflectance) where K represents the proportionality constant. The ambient light estimation system of claim 2, wherein the occupancy detector comprises: a reflectivity estimator for estimating the reflectivity of the region of interest in which the object is placed according to the relationship shown below, and in extracting the The incident light, the exposure time and gain of the image sensor are maintained during imaging: (relative brightness/reflectivity) _{without objects} = (relative brightness/reflectivity) _{with objects} . The ambient light estimation system of claim 2, further comprising: a light meter for measuring the illuminance of the incident light; and a calibration device for receiving the image without the object, and according to the relative brightness of the image, the image sensitivity The proportionality constant K is determined according to the exposure time and gain of the detector, and the reflectivity of the object, or the value of (K/reflectivity) is determined according to the relative brightness of the image, the exposure time and gain of the image sensor. The ambient light estimation system of claim 1, wherein the occupancy detector comprises: a motion detector for detecting movement of the object in the region of interest. The ambient light estimation system of claim 5, wherein the motion detector detects the movement of the object according to the difference between the current image and the previous image captured by the image sensor. The ambient light estimation system of claim 5, wherein the occupancy detector comprises: a state detector, when the motion detector detects the movement of the object and triggers the state detector to determine the interest The presence status of the area. The ambient light estimation system of claim 1, wherein the image sensor comprises a visible light image sensor. The ambient light estimation system of claim 1, wherein when estimating the ambient light, a specular reflection area with a brightness value greater than a predetermined threshold value is excluded.

Images