PRNewswire: CEVA announces that Altek, a leading Taiwanese imaging systems provider, has licensed CEVA's imaging and vision DSP to add imaging and deep learning capabilities to its imaging solutions and dual camera technologies targeting smartphones, ADAS, AR/VR, drones and other smart camera devices. Altek and its customers will be able to utilize the open, programmable vision DSP to deploy convolutional neural networks (CNN) for deep learning tasks tailored to specific applications or use cases, enabling true product differentiation.
PRNewswire: CEVA announces that ON Semiconductor has licensed CEVA's imaging and vision platform for its ADAS product lines.
"The automotive industry requires cost- and power-efficient vision-based ADAS solutions to address the growing end-customer demand and safety regulations across all tiers of the automotive industry," said Ross Jatou, VP and GM of the Automotive Solutions Division of ON Semiconductor. "CEVA's industry-leading vision processing IP provides us with a comprehensive solution that enables us to integrate a host of innovative and intelligent system features into our ADAS product offerings."
PRNewswire: Rockchip presents three comparison videos based on RV1108 visual processing platform with 600MHz CEVA XM4 visual processor:
Thursday, January 05, 2017
Wednesday, January 04, 2017
Autonomous Driving News
Valeo publishes a nice video about its hands-off journey across the USA:
Engadget shares its experience with Delphi-Mobileye self-driving car with 9 cameras and 10 radars:
Wired posts a video about an affordable Huyndai autonomous car equipped with 4 cameras, 3 "cheaper" LiDARs and 3 radars:
Engadget shares its experience with Delphi-Mobileye self-driving car with 9 cameras and 10 radars:
Wired posts a video about an affordable Huyndai autonomous car equipped with 4 cameras, 3 "cheaper" LiDARs and 3 radars:
Changhong Smartphone Features Embedded Spectrum Analyzer
PRNewswire: Changhong H2 smartphone features a spectrum analyzer inside. Changhong has worked closely with Consumer Physics and ADI to develop and manufacture this first-of-its-kind smartphone. Since collaborating in early 2016, ADI and Consumer Physics have been committed to miniaturizing the SCiO sensor for integration into a wide variety of connected devices.
"Changhong H2, with the integrated, miniaturized near-infrared spectrometer, marks a huge step forward for connected devices and makes it possible to see the physical world around us like never before via a mobile phone, which is the most common smart device in the world," said Martin Cotter, SVP, Worldwide Sales and Digital Marketing, Analog Devices.
Update: EETimes article shows Analog Devices' 4x8x4 mm3 spectrometer module used in the phone:
"Changhong H2, with the integrated, miniaturized near-infrared spectrometer, marks a huge step forward for connected devices and makes it possible to see the physical world around us like never before via a mobile phone, which is the most common smart device in the world," said Martin Cotter, SVP, Worldwide Sales and Digital Marketing, Analog Devices.
Update: EETimes article shows Analog Devices' 4x8x4 mm3 spectrometer module used in the phone:
Socionext Announces 4K ISP & Vision Processor for Drones
PRNewswire: Socionext introduces the SC2000, a new image processing IC designed for drones and action cameras. The SC2000 is a new product from the "M10V" series, an eighth-generation version of Socionext's Milbeaut Image Processors. The new SC2000 has a digital signal processor optimized for computer vision. It also features advanced functionalities such as 360-degree, real-time panorama stitching with four cameras, image stabilization without mechanical gimbals, and rolling shutter correction. The image processing rate is 1.2 gigapixels per second, three times faster than the company's previous processor, while the power consumption is as low as 1.7 Watt in typical operating conditions.
Snapdragon 835 Features Improved ISP and Vision Processing Functions
The new Qualcomm Snapdragon 835 features Spectra 180 ISP and Hexagon 682 DSP with support for TensorFlow for machine learning and Halide for image processing.
The Snapdragon 835 features "smooth optical zoom capabilities, fast auto-focus technologies and HDR true-to-life colors with perceptual quantization video encoding. At the core of the capture experience is the Qualcomm Spectra 180 camera ISP, featuring dual 14-bit ISPs that will enable up to 32MP single or dual 16MP cameras."
Italian site HWUpgade published Qualcomm's slides explaining the new camera features:
HotHardware site adds few more slides and explanations:
"When it comes to image capture, the Spectra 180 features two separate ISPs, one which can interface with a wide-angle lens and one that can accommodate a telephoto lens. EIS3.0 brings improved video stabilization with gyro-based pitch, yaw and roll correction along with enhanced rolling shutter correction."
"On the security front, the Haven Security Platform offers eye- and face-based secure camera authentication:"
The Snapdragon 835 features "smooth optical zoom capabilities, fast auto-focus technologies and HDR true-to-life colors with perceptual quantization video encoding. At the core of the capture experience is the Qualcomm Spectra 180 camera ISP, featuring dual 14-bit ISPs that will enable up to 32MP single or dual 16MP cameras."
Italian site HWUpgade published Qualcomm's slides explaining the new camera features:
HotHardware site adds few more slides and explanations:
"When it comes to image capture, the Spectra 180 features two separate ISPs, one which can interface with a wide-angle lens and one that can accommodate a telephoto lens. EIS3.0 brings improved video stabilization with gyro-based pitch, yaw and roll correction along with enhanced rolling shutter correction."
"On the security front, the Haven Security Platform offers eye- and face-based secure camera authentication:"
Tuesday, January 03, 2017
Fraunhofer facetVISION Camera is 2mm Thin
Fraunhofer researchers have developed a process enabling a two millimeter flat 1MP camera. Similar to the eyes of insects, its lens is partitioned into 135 tiny facets. The researchers have named their mini-camera concept facetVISION. The cameras are said to be suitable for use in the automotive and printing industries and in medical engineering. Their basic principle is said to be able to change the design of future smartphones.
"With a camera thickness of only two millimeters, this technology, taken from nature’s model, will enable us to achieve a resolution of up to four megapixel", says Andreas Brückner, project manager at the Fraunhofer Institute for Applied Optics and Precision Engineering IOF in Jena. "This is clearly a higher resolution compared to cameras in industrial applications – for example in robot technology or automobile production."
"It will be possible to place several smaller lenses next to each other in the smartphone camera. The combination of facet effect and proven injection molded lenses will enable resolutions of more than 10 megapixels in a camera requiring just a thickness of around three and a half millimeters," adds Brückner.
Update: As mentioned in comments, Fraunhofer has opened www.facetvision.de web site devoted to the project and its applications. LaserFocusWorld published an expanded article on the new camera and its future developments.
"With a camera thickness of only two millimeters, this technology, taken from nature’s model, will enable us to achieve a resolution of up to four megapixel", says Andreas Brückner, project manager at the Fraunhofer Institute for Applied Optics and Precision Engineering IOF in Jena. "This is clearly a higher resolution compared to cameras in industrial applications – for example in robot technology or automobile production."
"It will be possible to place several smaller lenses next to each other in the smartphone camera. The combination of facet effect and proven injection molded lenses will enable resolutions of more than 10 megapixels in a camera requiring just a thickness of around three and a half millimeters," adds Brückner.
Update: As mentioned in comments, Fraunhofer has opened www.facetvision.de web site devoted to the project and its applications. LaserFocusWorld published an expanded article on the new camera and its future developments.
French Startup Presents Lightbulb with Camera
BBC reports that French startup Bell & Wyson presents a security camera integrated into a light bulb. Assuming the lightbulbs are single-use items having a limited life span, this can become a significant market for image sensors:
Monday, January 02, 2017
TSMC Proposes Pixel with No High Energy Implants
TSMC patent application US20160351604 "Implant damage free image sensor and method of the same" by Alexander Kalnitsky, Jhy-jyi Sze, Dun-nian Yaung, Chen-jong Wang, Yimin Huang, and Yuichiro Yamashita proposes pixel process flow with minimal usage of ion implantation, if at all. Instead, selective and non-selective epi growth, in-situ doping, and diffusion are used extensively:
"Conventional methods of addressing the dark current and cross-talk issues, such as deep trench isolation (DTI), require conducting an implant operation, which is likely to cause damage on the image sensor and is confined by an implant depth limitation. As such, additional defects and interference may be induced. It would therefore be desirable to be able to provide improved image sensor for capturing images.
Compared to an existing process, DTI without the use of high energy implants can reduce the chance of introducing defects to the pinned photodiode structure and allows for the formation of a much deeper P-N junction compared to the one formed by ion implantation. Very deep junction may be beneficial for the near infrared (NIR) sensor."
"Conventional methods of addressing the dark current and cross-talk issues, such as deep trench isolation (DTI), require conducting an implant operation, which is likely to cause damage on the image sensor and is confined by an implant depth limitation. As such, additional defects and interference may be induced. It would therefore be desirable to be able to provide improved image sensor for capturing images.
Compared to an existing process, DTI without the use of high energy implants can reduce the chance of introducing defects to the pinned photodiode structure and allows for the formation of a much deeper P-N junction compared to the one formed by ion implantation. Very deep junction may be beneficial for the near infrared (NIR) sensor."
Subscribe to:
Posts (Atom)