Every major city in the world faces the problem of illegal construction—owners building extra floors or people encroaching on reserved land. Not only does this cause problems for the people living in the house but can also cause major problems to the environment, public safety and are a huge fire hazard. As of today, most of the checking for illegal constructions happens manually through site visits and inspection. Many of these defaulters get away with their acts for decades before it is discovered that the house, they are living in is not approved/is illegally constructed. Also, it is a major loss of money for government authorities. This paper describes a method, using drones to capture aerial images and store them in a secure database. Once that is done, an algorithm is described that can comb through the millions of images that are present in the database and flag all those images that can be possible or positive illegal construction with the help of image processing using computer vision. The algorithm would flag the images which may have possible faults and present them to the supervisor who can then manually approve the flagged sites for further physical inspection or can deem them as having no faults. This paper only presents a method to reduce the time spent by the person who is manually checking each of the sites' images and automatically presents him/her only ones that have a high risk of faulting them. This will not only significantly reduce the time spent on checking but can also save the authorities lots of money and help them to tackle more such defaulters. This paper describes a method, using drones to capture aerial images and store them in a secure database. Once that is done, an algorithm is described that can comb through the millions of images that are present in the database and flag all those images that can be possible or positive illegal construction with the help of image processing using computer vision. The algorithm would flag the images which may have possible faults and present them to the supervisor who can then manually approve the flagged sites for further physical inspection or can deem them as having no faults. This paper only presents a method to reduce the time spent by the person who is manually checking each of the sites’ images and automatically presents him/her only ones that have a high risk of faulting them. This will not only significantly reduce the time spent on checking but can also save the authorities lots of money and help them to tackle more such defaulters.

Illegal construction Drones Image Processing Computer Vision