How to select the best industrial camera
1 like212 views
The document provides guidance on selecting the best industrial camera for a system. It recommends a 3-step process: 1) Identify key camera parameters and prioritize requirements, 2) Compare specifications of existing cameras and develop a shortlist, 3) Evaluate top cameras by measuring image quality and system performance. The process aims to find the camera best matching needs within budget and supplier considerations.
How to select the best industrial camera
- 1. How to Select the Best Industrial Camera
- 2. introduction Choosing the correct industrial camera for your system from the beginning can eliminate costly redesigns or upgrades in the future. The question is, how do you determine the best camera? Best is always subject to your needs. There are hundreds of options for just high-performance industrial camera, so where do you begin? We suggest a step-by-step approach that will hopefully help in the camera selection process to find the best candidate. Here is an outline of the approach we will expand upon further. 1. Identify your key camera parameters a. b. 2. Define the ideal camera Prioritize your requirements c. Compare specifications of existing products Preliminary camera selection a. b. 3. Create list of possible products More thorough comparison of specifications c. Look for references to get to develop short list Camera Evaluation a. Look at an image b. Measure the Photon Transfer Curve c. Test within your system set-up At the end of these steps, you should know what camera is the best for you. Of course then you still have to determine if the company offers the right support and manufacturing controls as well as consider pricing versus value. Or, if at the end of this, you determine that none of the existing products will meet your needs, perhaps some minor modifications will make for a perfect fit.
- 3. Step 1 Identify your Key Camera Parameters So to begin, here are some tips on how to identify the key camera parameters that are most important to you. The best and most obvious place to start is at the end or in this case with the properties of the required resulting image. You can describe your image-needs in terms of: Contrast Contrast is the difference in brightness between the light and dark areas. It may be negatively influenced due to reflections in the optics and or between reflections between sensor and optics. For precise measurements these reflections can be detrimental to accuracy. This relates to the dynamic range (ratio of the maximum output signal to the dark level) and linearity (the percentage deviation from the ideal response of an image sensor) of the camera. In some applications it might lead to the requirement to have the coverglass removed from the image sensor in the camera. Resolution Resolution is the sensor resolution in combination with the field of view (FOV) that is being detected. It primarily defines the maximum spatial details that can be captured. The equipment you are building in combination with the measurement algorithm should give you numbers like pixels/object and objects per image. Sharpness Sharpness is affected by the combination of image sensor and optics. If both are not well matched, spatial details will be blurred. Distortion Distortion is an ‘error’ in the spatial representation of the 2D scene within the image. The combination of image sensor and optics give rise to this artifact.
- 4. Noise Noise can be shot noise at the sensor pixel level. At the image level view, noise is related to capability of the camera to represent a 2D scene with a certain uniformity. Non uniformity due to the camera is caused mainly by offset & gain differences (e.g. noise) between pixels, and by pixel-, row- and/or column defects. Extensive non-uniformity will limit the resolving capabilities of the camera, especially when image sensor resolutions increase (e.g. multi mega pixel field of views). When looking from image to image, it is the stability of the camera that is seen as noise. An unstable behavior will limit the ability to accurately capture subtle intensity variations between images. In other words, a badly designed camera can ruin the images from good image sensors. Define the ideal camera After considering this how do you then determine what camera you need? You want all of these to be the best possible right? This would mean a camera with the highest values in many parameters such as those listed are listed in the following list. Important camera parameters » » » » » » » » » » » » » Intra scene Dynamic Range Detection Linearity Pixel (shot) noise Image Uniformity Optical Reflections Modulation Transfer Function (MTF) Resolution within the Field Of View Geometrical Distortion Camera Stability Sensitivity Frame Speed Smear (Sensor Capturing Principle) Blooming Prioritize your requirements You can identify the ideal values for all of the above, and then accept that a camera meeting all of those is likely not cost effective. It is therefore a good idea, to prioritize this list for the 5 most important parameters for you to achieve the image details you require. First you can likely narrow down the basics of resolution, and frame speed necessary. Then it can get a little harder. For example, determine whether your measurement is read noise or shot noise limited. You should also consider the system requirements, such as what kind of illumination you have to determine the actual sensitivity required or what optics will be implemented. If you buy a high-performance camera and then an insufficient lens or vice versus, you will not get the desired result. There is also which interface is preferred, especially if the project is an upgrade of an existing system.
- 5. Compare specifications of existing products Now you are ready to search for products that match your key specifications. All of the usual ways of internet research, trade show attendance, or phone calls to suppliers should give you a good indication of what your options are.
- 6. Step 2 Preliminary Camera Selection Create a list of possible cameras with the desired functionality Suppose you have determined you need a lot of contrast in your image which means you need a high dynamic range, and your other system conditions have led you to conclude you need 4 Megapixels with a frame rate of at least 150 frames per second. The system is already based on the Camera Link interface and because several images are used for one measurement, you want global shutter technology to minimize the image-to-image effects. This can help you narrow down your search to cameras based on a specific sensor, in this case say the CMOSIS CMV 4000. As you begin your search, you may find it difficult to compare camera specifications as each supplier may have determine the values in a different way or under different conditions even though they all use the same image sensor. Another thing that is confusing is that cameras utilizing the same sensor can show different values for several parameters, such as dynamic range. Detailed comparison of cameras For example, the Point Grey Gazelle GZL-CL-41C6, the Adimec QUARTZ Q-4A180, and the Vieworks VC-4MC-M160 all utilize the CMOSIS CMV4000 image sensor. From readily available and downloadable documentation from the manufacturers, you can find the following: Point Grey Gazelle GZL-CL-41C6 SNR 58 dB Adimec QUARTZ Q-4A180m/CL Dynamic Range 60 dB Vieworks VC-4MC-M160 Dynamic Range 56 dB
- 7. These values can depend on the way in which the measurements were made, also Point Grey presents the information as signal to noise ratio (SNR) rather than dynamic range. The EMVA 1288 was developed to allow a consistent way of quantifying and comparing camera parameters. Some camera manufacturers have had this analysis done by a third party or have done it themselves and share the data. The cost for the analysis or equipment analysis is somewhat high so some have developed their own methods. Even with comparable data, you still need to know which exact value is required by your system. Consult experts and references to narrow down to 3 to 5 suppliers Also, you may still have a relatively long list of potential options at this point. Perhaps this list can be further narrowed down by asking either colleagues or industry professionals to share their experiences (LinkedIn groups are a great resource). As the cameras are going to be used in your systems, you may seek references and do some research on the stability and consistency of the manufacturer in addition to individual camera performance. Before camera evaluation starts . . . Before you start ordering evaluation cameras and talking to camera vendors try creating a checklist of items that are not camera properties but will affect the long-term buyer-supplier relationship you will have. These include items such as support level, quality of service, longterm supplier, markets served, etc. Once you have done this, you will likely have a shorter list of 3 -5 possibilities and now you are ready to start evaluations.
- 8. Step 3 Camera Evaluation and Final Selection Even if you have consistent data from the camera manufacturers and strong recommendations, you need to determine the necessary level of certain performance parameters since there are sometimes price consequences for more stringent requirements. Most industrial camera manufacturers have an evaluation/demo/loaner program to allow you to borrow a camera and see for yourself. This process can require a lot of time which is why it is so important to have previously minimized the number of possibilities. Though the evaluation process is lengthy, it allows you to: » » » » See the actual hardware, the software, and the integration for yourself Verify your assumptions Verify the specifications provided Check that it will enable the measurements you want to do Look at the image A good starting point is to simply get the camera up and running and look at the resulting image. You can play around with the gain settings to get a feel for defect pixels or other artifacts. You may want to quantify performance differences for yourself. The Photon Transfer Curve (PTC) characterization method can be used as a standardized test procedure to verify performance such as read noise, dark current, full well capacity, sensitivity, dynamic range, LOG (Noise) Measure the photon transfer Curve SLOPE=1 SLOPE=1/2 SLOPE=0 gain, and linearity. A curve rather than one data point can often give more insight if you know how to interpret it. There are several Full Well Read Noise Shot Noise Fixed Pattern Noise LOG (Signal) great sources on how complete a PTC calibration and calculate the camera parameters.
- 9. The links below are some other examples: http://www.couriertronics.com/docs/notes/cameras_application_notes/Photon_Transfer_Curve_ Charactrization_Method.pdf http://harvestimaging.com/blog/?p=784 Test within your system set-up To truly understand which camera will be best, you may need to try them directly in your system or at least in a test set-up, if possible. At the end of these steps, you should know what camera is the best for you. Of course then you still have to determine if the company offers the right support and manufacturing controls as well as consider pricing versus value. Or, if at the end of this, you determine that none of the existing products will meet your needs, perhaps some minor modifications will make for a perfect fit. Then you need to find a manufacturer with a well-matched starting product and the willingness and experience to tailor the camera.
- 10. adimec Adimec specializes in the development and manufacturing of high-performance cameras that meet the application-specific requirements of key market segments, including machine vision, medical imaging, and outdoor imaging. Founded in 1992, the company partners with major OEMs around the world to facilitate the creation of industry-leading cameras. The unique Adimec True Accurate Imaging® technology provides new levels of precision and accuracy to vision systems. Its diverse line of camera products meet a wide range of performance, size, cost, interface and application requirements. Adimec has offices around the world focused on creating customer value and satisfaction through local, personalized support. Need more inspiration? Contact us www.adimec.com