Sunday, June 30, 2019

HDR Pixel with Charge Splitting

MDPI publishes a paper "A Wide Dynamic Range CMOS Image Sensor with a Charge Splitting Gate and Two Storage Diodes" by Minho Lee, Min-Woong Seo, Juyeong Kim, Keita Yasutomi, Keiichiro Kagawa, Jang-kyoo Shin, and Shoji Kawahito from Shizuoka University, Japan, and Kyungpook National University, Korea.

"In this paper, a wide dynamic range (WDR) CMOS image sensor (CIS) with a charge splitting gate (SG) and two storage diodes (SDs) is presented. By using single-gate on/off control with the SG, photocurrent path to the first (SD1) or second storage diodes (SD2) is switched alternatively and periodically during exposure and signal electrons generated in a photodiode (PD) are transferred to and accumulated in the SD1 or SD2. By setting a large ratio of the off-time to on-time of the SG, two different sensitivity signals, which are originated by the same photodiode, are generated and a WDR image signal is obtained. This technique has a distinct advantage on mitigating the problem of motion artifact in WDR imaging with high and low sensitivity signals and flexible dynamic control of the dynamic range. An experimental WDR CMOS image sensor with 280 (H) × 406 (V)-pixel array consisting of 14 sub-arrays, each of which have 20 (H) × 406 (V) pixels, was implemented and tested. For the SG on/off-time ratio of 30 and 279, the DR of 93 dB and 104 dB, respectively, was demonstrated. The effect of the proposed WDR imaging operation on the reduced motion artifact was experimentally confirmed."

Friday, June 28, 2019

Albert Theuwissen Reviews Day 4 of IISW

Albert Theuwissen publishes his report about the last day of IISW 2019. This part talks about rad-hard image sensors and global shutter session. The report wraps up:

"The IISW2019 was again a big success, many excellent presentations, all high quality papers, good atmosphere, perfect organization. The interest of the participants can be best judged by the amount of people that are willing to stay till the very last paper is presented."

Thursday, June 27, 2019

Peter Centen Awarded by Order of the Netherlands Lion

Grass Valley congratulates Peter Centen on being awarded by the Order of the Netherlands Lion for his significant contributions to Dutch society and the broadcast industry. The Order is given on behalf of King Willem-Alexander of the Netherlands to only a limited amount of people each year who have distinguished themselves by exceptional artistic achievement, scientific performance, technical innovation or inventive entrepreneurship.

Centen played a major role in transitioning camera technology from CCD imagers to CMOS sensors — which enabled present-day HD, 4K UHD and HDR imaging. Centen invented HD-DPM technology and has been recognized worldwide for his work in imaging.

"This recognition is incredibly well deserved. Peter is a visionary and was a major asset to us at Grass Valley and the industry as a whole," said Marcel Koutstaal, GM and VP cameras, Grass Valley. "It is incredible to see him being honored for his impactful work, his passion for people and significant contributions to the industry over the last 34 years."

Centen has published over 65 international journal and conference articles on image sensor technologies and cameras.

Update: One more pictures from

SPAD Thesis

Delft University publishes a PhD thesis "CMOS SPAD Sensors for 3D Time-of-Flight Imaging, LiDAR and Ultra-High Speed Cameras" by Chao Zhang. This quite an extensive work over viewing different SPAD design techniques, ways to improve PDP and reduce DCR, photon counting electronics, etc. There is also a nice comparison between different ToF approaches:

Albert Theuwissen on Third Day of IISW

Albert Theuwissen posts his impressions from the Day 3 of IISW 2019 in Snowbird:

"Although the amount of paper submissions for this 2019 workshop was lower than in 2017, the quality of the presentations is most probably the best ever."

Oppo Demos its Hidden Under-Display Selfie Camera

Engadget quotes its colleagues from China visiting Oppo under-display camera demo at MWC at Shanghai:

"When the camera is idle, the screen works just as normal. However, when you look up close, the area above the camera appears to be more pixelated. According to Oppo, this zoned-out area features a highly-transparent material plus a redesigned pixel structure for improved light transmittance. In other words, this camera tech requires a customized display panel, because existing ones won't do the job -- their transparency properties are only good enough for in-display fingerprint readers, but not conventional cameras.

The under-screen camera itself also packs a larger sensor with bigger pixels, along with a larger aperture to get as much light as possible.

...based on our quick comparison, there's certainly room for improvement in terms of clarity and color accuracy.

Here is an example of selfie image quality of the Oppo camera (right) in comparison with a regular selfie camera (left):

Wednesday, June 26, 2019

AutoSens Brussels Agenda

Autosens Brussels scheduled for September 17-19, 2019 publishes its agenda with quite a lot of image sensor-related presentations:

  • Recent Developments in the CIS Auto World, 3 hour 15 min lecture by Albert Theuwissen, IISS
  • Overview of the challenges and requirements for state-of-the-art sensors across the major sensor types, Geoff Ballew, ON Semi
  • Panel Discussion: Who will be the winners and losers in AV and automotive sensors?
  • Snow-aware ADAS with Active Hyperspectral Sensing, Raul Bravo,
  • The race to a low-cost LIDAR system, Dexin Chen, IHS Markit
  • ShortWave Infrared Breaking the Status Quo - Identifying Hazards on the Road and Solving the Low Visibility Challenge, Uriel Levy, Trieye
  • The New Generation of MEMS LiDAR for Automotive Applications, Florian Petit, Blickfeld
  • A novel CCD LiDAR imager, Beat de Coi, Espros
  • Invisible integration of solid-state LIDAR to make beautiful self-driving cars, Filip Geuens, Xenomatix
  • Hybridization - To do, or not to do? An Update of use of Hybrid Lenses for Automotive ADAS Applications, Ingo Flodvari, Sunex
  • The evolution of automotive lighting, and considerations for the integration of illuminators and sensors into headlamps, Paul Lyon, Varroc Lighting Systems
  • Detectivity – Ranking Cameras for Machine Vision Tasks, Robin Jenkin, Nvidia
  • Testing automotive camera modules - the difference between theory and reality, Uwe Artmann, Image Engineering

Sony to Release Two 4K Sensors for Security Cameras

Sony announces the upcoming release of two new CMOS sensors: the IMX415 type 1/2.8 4K-resolution stacked one, the smallest 4K security sensor in the world (1.45um pixels), and the IMX485 type 1/1.2 4K-resolution BSI sensor with 3.3x low-light performance of previous generation IMX334 type 1/1.8 sensor. Sony has designed these two new sensors for security applications.

Postcard from IISW at Snowbird

Here is a group photo of IISW 2019 attendees (the full resolution picture can be downloaded here):

Thanks to EF for the picture!

Albert Theuwissen Reviews IISW - Day 2

Albert Theuwissen posted his review of the second day of the presentations. The big story of the day is ST pixel design with no photodiode implants.

Espros Unveils TOF>cam 635

ESPROS rolled out the first series production modules of the new 3D camera, TOF>cam 635. This module is based on ESPORS' ToF chip epc635 with 160 x 60 pixels and improved ambient light suppression up to 100Klux. The cameras firmware controls image acquisition and processing to provide a calibrated and filtered 3D point cloud. It also delivers a grayscale image. This combined with sophisticated algorithms allows interference and motion blur detection and suppression. The cost is said to be very low.

ESPROS also offers a full custom module design & manufacturing service.

Albert Theuwissen Reviews IISW - Day 1

Albert Theuwissen posts a short overview of Day 1 in International Image Sensor Workshop being held these days in Snowbird, UT. There were talks about small pixels, noise, and dark current.

Tuesday, June 25, 2019

Lucid Vision Labs Explains Sony ToF Sensor Operation

Lucid Vision Labs publishes a very nice animated explanation of how Sony IMX586 ToF sensor works inside Lucid's Helios camera:

"The IMX556 utilizes a current assisted photonic demodulator (CAPD) pixel structure that samples incoming light synchronously with emitted light modulation. CAPD allows for an alternating voltage inside each pixel’s photodiode, creating drift fields that divide and pull electrons to alternating detector junctions."

Interview with TowerJazz SVP of Image Sensor Business

Charbax publishes an interview with Avi Strum responsible for TowerJazz image sensor business:

Apple Stops Working on Quantum Dot Imager with Nanoco

iClarified quotes report that Apple stops working on Quantum Dot image sensor and does not extend a contract with Manchester, UK-based Nanoco supplying the QD materials. Apple "has now informed Nanoco that the project will not continue beyond the current contract, for reasons wholly unconnected to the performance of our materials and our service delivery."

Apple was expected to use the QD technology in a future iPhone image sensor that could have debuted as early as this year.

Thanks to JB for the pointer!

Monday, June 24, 2019

Corning Shipped 2 Million Varioptic Liquid Lenses

Corning announces that it has delivered its 2 millionth Corning Varioptic liquid lens for industrial applications.

Varioptic became a part of Corning in 2017 through an acquisition that included Varioptic and Invenios technologies,” said John Duke, VP and GM, Corning Glass Microsystems. “We believe the synergies from this acquisition will lead to exciting new liquid lens application opportunities and the advancement of our Microsystems business.

This milestone was driven by an increased demand for liquid lenses, particularly as a variety of industries integrate them into manufacturing and warehouse operations. This achievement strengthens Varioptic’s position as a leader in liquid lens solutions for machine vision, barcode readers, medical imaging, and other industrial applications that require robust, accurate, fast-focusing, and long-lasting lenses.

Our manufacturing capacity nearly doubled in 2018, enabling us to produce liquid lenses at a higher volume than ever before,” said Frédéric Laune, business manager, Corning Varioptic Lenses. “Our ability to innovate paid off with sales doubling in the last two years.

Varioptic lenses have no moving parts and feature electrowetting technology that uses electrostatic forces to change the radius of the spherical interface between two immiscible liquids. They also enable exceptional shock and vibration resistance, which is essential for current and future applications.

Omnivision Expands its NIR-Enhanced Sensor Portfolio

PRNewswire: OmniVision announces the OS04A10, a 2.9um pixel sensor with 4MP resolution, the latest addition to its growing family of Nyxel NIR and ultra-low light (ULL) sensors. The OS04A10 with QE is 60% at 850nm and 40% at 940nm, said to be 3x to 5x better than sensors without this technology.

"In addition to requiring top performance in all lighting conditions, security and surveillance camera designers are demanding higher resolution for greater facial recognition accuracy with longer zoom range," said Brian Fang, business development director at OmniVision. "By combining our industry-leading NIR and ULL technologies with 4 MP resolution, the OS04A10 is designed to meet the evolving needs of security and surveillance systems."

The amount of NIR light that a sensor requires to capture high quality images can be quantified with a new metric called NIR SNR1, which takes into account the QE, pixel size and read noise. The OS04A10 is said to achieve industry leading SNR1850nm and SNR1940nm performance, which is 2x to 3x smaller when compared with the leading known available competitor sensors.

The OS04A10 is said to also achieve the industry's lowest SNR1 while producing 4 MP and 1080p images for video recording. Additionally, OmniVision's dual-conversion-gain pixel provides good ULL and HDR performance. The 3-frame staggered shutter also minimizes motion artifacts and enables an HDR of 120 dB.

This image sensor is available in an extremely small package enabled by PureCel-S die stacking technology. The PureCel-S technology also reduces power consumption.

Canon Develops HDR Sensor and Improves its Older Sensors

PRNewswire: Canon Inc is pursuing development of the 3U3MRXSAAC, a 2.8MP, 1/2.32-inch CMOS sensor for HDR imaging in -40C to 105C temperature range.

"As Canon evolves in the sensor market, we are dedicated to utilizing our expertise to develop products built to meet current growing market trends," said Kazuto Ogawa, president and COO, Canon U.S.A., Inc. "The 3U3MRXSAAC CMOS sensor under development reflects the continuation of our business strategy that leverages the high-quality imaging technology Canon is able to achieve."

The 3U3MRXSAAC CMOS sensor is expected to achieve a DR of 120 dB at low noise levels. Even during normal drive operation, the sensor can achieve a dynamic range of 75 dB—greater than the sensors on many conventional digital cameras.

Additionally, the 3U3MRXSAAC CMOS sensor will also be capable of operating in environments with extreme temperatures ranging from -40C to 105C. Typically, in high-temperature conditions, the increase in dark current noise can affect the quality of the image. However, the 3U3MRXSAAC CMOS sensor will be equipped with functionality that helps correct black levels in real-time to help maintain high image quality.

The 3U3MRXSAAC CMOS sensor will feature a pixel size of 3.2 μm x 3.2 μm and a frame rate of up to 60 fps (30 fps during HDR drive operation). The sensor will be compatible with the MIPI CSI-2 interface.

PRNewswire: Canon re-announces its 35MMFHDXSMA, a 19um-large pixel 2.7MP sensor, and 120MXSI, a 2.2um 120MP APS-H sized sensor. Both sensors have been already presented earlier, albeit with slightly different part codes. It's not immediately clear what has changed in these new sensors.

Stacked SPAD for LiDAR Thesis

Delft University publishes a PhD thesis "Modular time-of-flight image sensor for light detection and ranging: A digital approach to LIDAR" by Augusto Ronchini Ximenes. Other than overview of the stacked SPAD chips and their components, the thesis has a nice system level model of a LiDAR, probably showing the problem that Fastree3D LiDAR startup dealing with.

The defense of the thesis is scheduled on July 12, so that it leaves enough time for critics.

"Direct time-of-flight (dTOF), specifically, requires more elaborate detectors and data processing, but it has the potential of reaching much longer distances at higher speed and accuracy, with the advantage of being robust to high background noise, making it suitable for space, automotive and consumer applications. One known drawback of dTOF, however, is data volume. For instance, automotive applications require over 100m range, only few centimeters accuracy, and multiple measurements for a reasonable precision, which produce data rates that can reach tens or even hundreds of Gbps, in large sensors, thus setting processing constraints to even very efficient GPUs, as well as chip readout capability. It is essential to provide as much on-chip processing as possible, in order to reduce data throughput, thus reducing power consumption and speeding up processing time. Some architectures have been proposed attempting to solve this problem, but the required memory renders them only feasible for an SiPM, single-pixel approach. Another known issue with light detection and ranging (LiDAR) is regarding the interference of multiple systems on each other. A software-based approach has been implemented, but requiring intensive post-processing resources. In this thesis, a novel approach for on-chip processing is proposed. With the use of cutting-edge 3D-stacking technologies, more flexibility and computational power can be spent on the chip, while not compromising fill factor. A novel proposal for dealing with external interferes is introduced, as well as novel phase/frequency locking solution at the sensor level, as a reference for timing measurements."

The system model of LiDAR analyses coincidence detection approach to remove sunlight for both scanning LiDAR (Condition 1) and flash one (Condition 2). The results show how difficult it is to achieve a long detection range on sunny days:

Sunday, June 23, 2019

XFAB Releases SPAD and APD Module in its 180nm Process

X-FAB Silicon Foundries announces the availability of APD and SPAD devices for implementation in the company’s 180nm high-voltage XH018 process. The APD has a high linear gain, and is fully scalable - going from just ten to several hundred micrometer dimensions. The proprietary X-FAB quenching circuit used in the SPAD results in a dead time of less than 15ns. In addition, its low dark count rate (less than 100 counts/s/µm²) means that it is less susceptible to thermal noise. The PDP of the SPAD is maintained across a wide range of wavelengths (e.g. 40% at 400nm). Furthermore, the low breakdown voltage (less than 20V) that has been achieved.

The X-FAB APD and SPAD can be utilized in a broad spectrum of different applications - including proximity sensing, LiDAR, time of flight (ToF), medical imaging (CT and PET) and scientific research. Being AEC-Q100 compliant, they are suitable for deployment within automotive systems.

Saturday, June 22, 2019

Startup Presentations at Embedded Vision Alliance

Entropix CEO Nathan Wheeler presents his resolution enhancement approach for security cameras:

Bo Zhu, CTO of BlinkAI, presents its AI algorithms for low-light and HDR imaging:

Friday, June 21, 2019

Smartphone Camera Resolution Grows

Counterpoint Research article talks about smartphone camera resolution:

"The trend of smartphone OEMs increasing the camera megapixels (MP) was clear in Q1 2019. According to Counterpoint Research’s data, smartphones with cameras of more than 12MP accounted for over 82% of the total shipments in Q1 2019.

During the quarter, the ratio of smartphones equipped with higher than 40MP rear camera reached nearly 3.6%, due to the increasing adoption of 48MP (0.8µm), which was applied in HONOR V20 and then widely used as the main camera within the triple camera setup.

In spite of a slight decrease in market share, the ratio of the 12~13MP camera remained over 57%. The massive adoption of 12MP is attributable to the trend of pursuing high-quality image, leveraging the relatively larger pixel size (1.22/1.4/1.55µm).

However, the share of 16~18MP was flat as compared to the previous quarter, while the ratio of smartphones equipped with an 8MP(or less) camera has been decreasing.

Samsung LSI just launched a new image sensor with resolution up to 64MP, aiming at the fall release cycle of smartphone flagships. This will further raise the level and quality of smartphone cameras, also resulting in an increased Bill of Materials (BoM) share.

Thursday, June 20, 2019

Guy Meynants Joins Photolitics

Bulgaria-based image sensor startup Photolitics is proud to announce that Guy Meynants has recently joined the management team of the young company focusing on speciality image sensor designs such as high speed linescan sensors, medical image sensors and highspeed imaging. Guy Meynants has been one of the first CMOS image sensor designers in Europe, starting with this “art” in 1994 at IMEC where he contributed amongst other to the first Web-cam flying to Mars, which is still sending images from its orbit around the red planet today (see

Guy was co-founder of Fillfactory which now is part of ON Semiconductor. In 2007, Guy co-founded CMOSIS and served as its CEO until 2009 when he became CTO and after AMS acquired CMOSIS he served as Engineering Fellow. Guy has always been working on image sensor design and technology development. He developed image sensors for SLR, space and industrial machine vision applications. He authored 60 publications and 30+ patent applications in the field. He is also board member and secretary of the International Image Sensor Society, an organisation to serve the CIS R&D community.

Yole and Xilinx Webcast on AI in Automotive Applications

Yole publishes its webcast on AI in automotive applications. Xilinx has joned the second part of the webcast:

Wednesday, June 19, 2019

Consolidation in LiDARs?

Crunchbase publishes an article "Consolidation In The Lidar Market May Be Coming" quoting Scott Burroughs, the co-founder and CEO of Sense Photonics:

"...there are about 70 to 80 companies that have attempted to enter the lidar market. Some have succeeded, some have died because they can’t get venture capital, and some are in stealth, he said.

Citing companies like Innoviz Technologies, VeloDyne, and Ouster, he added that he thinks “only a handful of lidar companies will be around in a couple years.”

When he started the company two years ago, he said the focus was on long-range lidar. Now, he said, there are different segments within the automation market, as different types of sensors can be applied to varied use-cases, whether it’s night sight or cutting through the fog.

“It used to be mostly for highway driving,” now there are lots of other segments and different cost points, he said. “Some companies will succeed going after the easy ones” like facial recognition, for example.

Tuesday, June 18, 2019

Market Data from TSR

Sony image sensor spin-off presentation quotes some interesting market data from Japan-based Techno Systems Research:

Monday, June 17, 2019

In Defense of Sony Semiconductor Spin-off

BusinessWire: Activist investor Third Point Highlights publishes explanations of its proposal to spin-off Sony image sensor business, opens A Stronger Sony dedicated web site promoting that, and publishes a 102-page long presentation with a lot of interesting market data, primarily from TSR.

The presentation gives a nice definition of image sensor on slide 27:

"Image sensor is an analog semiconductor with average price of $2 that processes light from the outside world and transforms those wavelengths into 1s and 0s."