CN114112959A

CN114112959A - Rapid high-precision hyperspectral depth imaging system

Info

Publication number: CN114112959A
Application number: CN202111607566.7A
Authority: CN
Inventors: 何赛灵; 罗晶
Original assignee: Taizhou Anqiling Intelligent Technology Co ltd
Current assignee: Taizhou Anqiling Intelligent Technology Co ltd
Priority date: 2021-12-27
Filing date: 2021-12-27
Publication date: 2022-03-01

Abstract

The invention discloses a rapid high-precision hyperspectral depth imaging system which comprises a projection structure light shape detection module, a pre-imaging unit, a tunable filtering spectrum scanning module and an information processing unit. The three-dimensional morphology data and the spectrum images of different wave bands can be fused into a four-dimensional hyperspectral morphology model. The invention combines the spectral detection of the tunable filter and the spectral morphology detection of the structured light into a whole, and overcomes the defects of low acquisition speed of other equipment, single and separate equipment, large data fusion error and the like. The invention has the advantages of both speed (acquisition speed is better than 1 second) and precision (three-dimensional reconstruction precision is better than 30 microns and spectral resolution is several nm), reduces the size of the tunable filter by design, and greatly reduces the cost and the occupied space. The method has important application value in the field of real-time spectrum morphology detection.

Description

Rapid high-precision hyperspectral depth imaging system

Technical Field

The invention belongs to the technical field of optics, and particularly relates to a rapid high-precision hyperspectral depth imaging system.

Background

With the continuous improvement of the development levels of material science, manufacturing process and industry, people have higher and higher requirements on industrial parts, and the requirements on use feeling and aesthetic degree are also higher. Industrial parts are used as components of various industrial products, and the appearance quality of the industrial parts directly influences the grade and precision level of the final products.

Due to the production environment and the manufacturing process of the industrial parts, some surface defects are inevitably generated, the overall quality of the industrial parts is influenced, and as the requirements for industrial production are gradually increased, the defect detection and screening of the industrial parts are increasingly indelible. The defects of the surface quality of the industrial parts mainly comprise dimensional precision difference and surface quality defects.

In the past decades, domestic machine vision detection of industrial part defects is mainly based on two-dimensional image feature extraction and identification. With the development of laser scanning technology, three-dimensional morphology recognition is gradually applied to three-dimensional detection of defects, richer space scale information is provided, and the detection accuracy is further improved. The traditional contact type three-dimensional topography measurement is gradually replaced by an optical non-contact type method because a probe is easy to scratch a sample and has an inefficient point scanning mode. The optical three-dimensional vision measuring method mainly comprises the following steps: photogrammetry, time-of-flight methods, triangulation, fringe projection methods, imaging surface positioning methods, interferometry, etc., confocal three-dimensional scanning. The confocal three-dimensional measurement has higher measurement accuracy compared with other methods, however, the conventional confocal three-dimensional scanning has low scanning efficiency and can only scan three-dimensional space point information. The information of the optical characteristics is lacking to judge the kind of defect.

Not only the difference of the geometric shapes of the samples to be detected is analyzed, but also the difference of materials is required to be detected, and the hyperspectral imaging technology is a suitable means. The hyperspectral imaging refers to an imaging mode capable of collecting continuous hundreds of wave bands, and each pixel point on an acquired image corresponds to a spectral curve. The hyperspectral imaging technology integrates the advantages of the imaging technology and the spectrum technology, and the information richness is greatly improved. Therefore, the detection of the three-dimensional shape of the object to be detected and the detection of the spectrum are of great help for analyzing the geometrical structure and the optics of the object. However, most of the existing devices combining two types of detection are low in acquisition speed, single and separated, large in data fusion error and difficult to have practical application value in the field of real-time high-precision spectrum morphology detection.

Disclosure of Invention

In order to solve the problems, the invention discloses a rapid high-precision hyperspectral depth imaging system which comprises a projection structure light shape detection module, a pre-imaging unit, a tunable filtering spectrum scanning module and an information processing unit. The information processing unit controls the tunable filter to keep the fixed wavelength transmission. And each time the projector projects a pair of stripe patterns to the surface of the sample to be measured, a trigger signal is sent out to control the camera to collect the stripe patterns for three-dimensional shape reconstruction. And when the information control unit receives the signals obtained after the projection of all the stripes is finished, the tunable filter is controlled to carry out wave band scanning through the illumination of the broadband light source, and the camera acquires all the scanning wave bands to obtain the spectral image. The three-dimensional morphology data and the spectrum images of different wave bands can be fused into a four-dimensional hyperspectral morphology model. Because the large-area tunable filter is difficult to manufacture and high in cost, the tunable filter is placed near a certain focus position of a light path, the area of the tunable filter can be greatly reduced, and the cost is reduced.

The invention has the beneficial effects that: the invention combines the spectral detection of the tunable filter and the spectral morphology detection of the structured light into a whole, and overcomes the defects of low acquisition speed of other equipment, single and separate equipment, large data fusion error and the like. The invention has the advantages of speed (the acquisition speed is better than 1 s) and precision (the three-dimensional reconstruction precision is better than 30 microns and the spectral resolution is several nm), reduces the size of the tunable filter by design, and reduces the cost and the occupied space. The method has important application value in the field of real-time spectrum morphology detection.

Drawings

FIG. 1 is a schematic diagram of a fast high-precision hyperspectral depth imaging system

As shown in fig. 1, the optical machine projection assembly 1, the lens 2, the imaging lens 3, the tunable filter 4, the collimating lens 5, the focusing lens 6, the camera 7, the information processing unit 8, and the broadband light source 9 are shown.

FIG. 2 is a schematic diagram of a high-dimensional imaging result of a rapid high-precision hyperspectral depth imaging system on broccoli.

Detailed Description

The invention is described below with reference to the drawings and specific examples.

A quick high accuracy hyperspectral depth imaging system which characterized in that: the device comprises a projection structure light shape detection module, a front-end imaging unit, a tunable filtering spectrum scanning module and an information processing unit. The information processing unit controls the projection structure light shape detection module to detect the three-dimensional shape data of the object to be detected, and the tunable filtering spectrum scanning module detects the spectrum data. The three-dimensional morphology data and the spectral data can be fused into a four-dimensional spectral morphology model.

A quick high accuracy hyperspectral depth imaging system, its characterized in that: the tunable filter of the tunable filter is designed to be placed near a certain focus (such as the focus of the collimating lens 5 in the front-end optical module) in the optical path, so that the area of the tunable filter can be greatly reduced, the cost is reduced (the cost is high due to the fact that the large-area tunable filter is difficult to manufacture), and the occupied space is also reduced.

A quick high accuracy hyperspectral depth imaging system, its characterized in that: the camera for collecting the spectral image modulated by the tunable filter and the structural light stripe image projected by the projector is the same camera, and the spectral data and the three-dimensional shape data have no fusion error.

The front-mounted optical module is characterized in that: comprises an imaging lens 3, a collimating lens 5 and a focusing lens 6. An image of an object to be measured is imaged at the focal position of the collimating lens 5 through the imaging lens 3, and then the image is collimated into parallel light of different angles by the collimating lens 5. All parallel light is focused on a camera 7 through a focusing lens 6 to form an image of an object to be measured. The image relay system consisting of the collimating lens 5 and the focusing lens 6 relays the image from the position of the tunable filter to the camera.

The projection structure light form detection module is characterized in that: the optical machine projection assembly projects structural light stripes onto the surface of an object to be measured, and the camera collects stripe image sequences. The information processing unit calculates the three-dimensional point cloud data of the object to be detected according to the position relation between the projection image pixel points of the optical machine projection assembly and the image pixel points collected by the camera.

The spectrum detection is characterized in that: the tunable filter realizes spectrum scanning by continuously changing the transmitted central wavelength. The tunable filter has short modulation time and can realize rapid spectrum morphology detection by matching with high-speed projection.

The projection structure light shape detection module can be replaced by other three-dimensional imaging detection modules, and the tunable filtering spectrum scanning module can be arranged on the same light path or arranged in front of a common camera on a conjugate light path.

To further illustrate the present example, four-dimensional topographic spectral data of broccoli was collected using the present system. The system writes a control program by C + +. A fast high-precision hyperspectral depth imaging system scans broccoli data as shown in figure II. The real size information of the broccoli can be obtained from the three-dimensional point cloud of the broccoli, and the point cloud detail distribution of the broccoli protruding buds is observed in the amplified three-dimensional point cloud appearance. The point cloud data can retrieve the corresponding spectral curve, so as to analyze the chlorophyll absorption of the surface of the broccoli.

Claims

1. A quick high accuracy hyperspectral depth imaging system which characterized in that: the system comprises a projection structure light shape detection module, a front-end imaging unit, a tunable filtering spectrum scanning module and an information processing unit; the information processing unit controls the projection structure light shape detection module to detect the three-dimensional shape data of the object to be detected, and the tunable filtering spectrum scanning module detects the spectrum data; the three-dimensional morphology data and the spectral data can be fused into a four-dimensional spectral morphology model.

2. The fast high-precision hyperspectral depth imaging system according to claim 1, wherein: the tunable filter of the tunable filter is designed to be placed near a certain focus (such as the focus of the collimating lens 5 in the front-end optical module) in the optical path, so that the area of the tunable filter can be greatly reduced, the cost is reduced (the cost is high due to the fact that the large-area tunable filter is difficult to manufacture), and the occupied space is also reduced.

3. The fast high-precision hyperspectral depth imaging system according to claim 1, wherein: the camera for collecting the spectral image modulated by the tunable filter and the structural light stripe image projected by the projector is the same camera, and the spectral data and the three-dimensional shape data have no fusion error.

4. The front optical module of claim 1, wherein: comprises an imaging lens 3, a collimating lens 5 and a focusing lens 6.

5. An image of an object to be measured is imaged at the focal position of the collimating lens 5 through the imaging lens 3, and then the image is collimated into parallel light with different angles by the collimating lens 5; all parallel light is focused on a camera 7 through a focusing lens 6 to form an image of an object to be measured; the image relay system consisting of the collimating lens 5 and the focusing lens 6 relays the image from the position of the tunable filter to the camera.

6. The projection structured light shape detection module of claim 1, wherein: the optical machine projection assembly is used for projecting structural light stripes onto the surface of an object to be measured, and a camera is used for collecting a stripe image sequence; the information processing unit calculates the three-dimensional point cloud data of the object to be detected according to the position relation between the projection image pixel points of the optical machine projection assembly and the image pixel points collected by the camera.

7. The spectroscopic probe of claim 1 wherein: the tunable filter realizes spectrum scanning by continuously changing the transmitted central wavelength; the tunable filter has short modulation time and can realize rapid spectrum morphology detection by matching with high-speed projection.

8. The projection structured light topography detection module according to claim 1, which is replaceable with other three-dimensional imaging detection modules, the tunable filter spectral scanning module being placed in the same optical path or being placed in front of a common camera in a conjugate optical path.